On July 1, Cloudflare shipped something called the Monetization Gateway. Any site behind their network, which is roughly a fifth of the web, can now charge for anything it serves: pages, datasets, APIs, tools. The mechanics run on HTTP 402, a status code that has sat in the spec unused since the 90s, literally reserved as “Payment Required” for a future that never showed up. An agent requests a resource, gets a 402 with a price, pays in stablecoins, and the content comes back. Sub-second, wallet to wallet, no account signup, no API key. It’s built for machine buyers, and Cloudflare says the quiet part out loud: AI crawlers request content a hundred to tens of thousands of times for every human visitor, and you can’t sell a subscription to a crawler.

This is Matthew Prince’s music-licensing prediction becoming product. He’s been saying the content economy looks like pre-Spotify music, everybody’s catalog getting played and nobody getting paid, and that micropayments were the way through. Ten months later his own company shipped the rails.

The web is bifurcating. Over half of traffic is already non-human, and the momentum is all on the agent side. What’s forming is a human web that still runs on ads and subscriptions, and a machine web that runs on per-request payments between agents. The human web monetizes attention. The machine web monetizes answers.

The auction question

I know this impacts auctions. I just can’t see the shape of it yet. Three candidates, and I’m not sold on which one is the real story:

1. Our data becomes a priced asset. Twenty-plus years of catalogs, lot descriptions, and realized prices is exactly the dataset an agent would pay per-query for. Today the crawlers take it free. I wrote in February that structured data is the last moat; a gateway like this is how a moat starts charging admission.
2. Agents become bidders. If a machine can settle payment at the edge in under a second, the distance between “agent that researches equipment” and “agent that bids on equipment” got a lot shorter. The plumbing for a machine buyer now exists.
3. The funnel inverts. If AI answer-engines start paying sources instead of sending traffic, discoverability stops meaning “rank for humans” and starts meaning “be the source the machines pay to query.” Marketing budget and data strategy start to merge.

Maybe it’s one of these. Maybe it’s something I haven’t thought of. I’m comfortable saying it’s coming, and it’s not nothing.

The tension with my own essay

Two weeks ago I told the industry to put it down where the goats can get it. Make yourself readable, plain, at ground level, free. This gateway raises the follow-up question I didn’t have to answer then: once the goats have wallets, when do you start charging for the feed? Give it away and you’re findable but uncompensated. Gate it and you’re paid but maybe skipped for the guy who left his feed on the ground. The answer probably isn’t the same for a listing as it is for twenty years of price history. But I don’t have it yet.

Open questions

1. Which of the three auction angles lands first, and is the right first move defensive (charge the crawlers taking our data now) or offensive (become the priced source agents prefer)?
2. Micropayments have died every time they’ve been tried on the human web. Is this attempt different because the buyer is a machine with no psychological friction and Cloudflare can flip the default for a fifth of the internet, or is it another false start?
3. Where’s the line between what Grafe gives away to stay findable and what it prices? The listing wants to be free. Does the realized-price archive?

The critique that stopped me: the validator was the same model as the thing it was checking. Same model, same characteristic errors. When the producer and the checker share a mind, the check confirms the work is consistent with itself. It can’t see the errors that mind is inclined to make, because it’s inclined to make them too.

The fix was cheap. Route my best model’s work to a validator running on a different model, and route the smaller models’ work to my best one. Nothing got smarter. If anything, the checker on my best model’s work is the weaker model on paper.

We already know this about people. Nobody lets an appraiser certify his own appraisal, and auditors are external by law even though the in-house accountants aren’t dumb; they’re just inside the same assumptions. A second opinion from the doctor who made the first diagnosis is not a second opinion.

But watch how organizations actually review work. The person who wrote the plan presents the plan and fields the questions. The team that built the process audits the process. The review is real effort, honestly performed, and it’s structurally blind anyway. My read is we spend on the intelligence of the reviewer when the leverage is in the independence.

The uncomfortable version: rereading my own email before sending is same-model validation. My tenth pass through my own proposal catches typos, not framing errors. The framing errors are load-bearing assumptions, and I’m the one who installed them. That’s why my writing pipeline separates drafter from cold-reader, and why the strongest move on a high-stakes decision is two advisors who can’t see each other’s work. Agreement between independent minds is evidence. Agreement between copies of the same mind is an echo.

Open questions

1. Where else does the producer grade his own work? In our auction business, setup checklists verified by the crew that did the setup is the obvious candidate.
2. Is there a cost threshold below which same-mind review is fine? Typo-level checking probably doesn’t need independence; judgment-level checking always does.
3. Does this argue for keeping one deliberately different mind in every important loop (an outside reviewer, a contrarian in the room), even when the different mind is individually weaker?

Sit with that. The web you’ve spent twenty years learning to win is now, by volume, mostly machines. And the machines aren’t browsing. They’re fetching. An agent doesn’t scroll your listings and admire your photography. It queries your inventory the way you’d query a database, takes what it can read, and moves on.

Which leaves one plain question to decide the next few years. When the machine comes looking, can it read you?

The deal changed

Prince has been putting out a second number for a while, and it’s the one that should bother every operator. Ten years ago, Google crawled about two of your pages for every visitor it sent back. Fair trade: you let them index you, they sent you traffic. The AI crawlers don’t trade like that. The pure-LLM bots pull thousands of pages for every click they send back. They read everything you publish and send you almost no one.

That’s the part people wave off when they call AI “only one percent of traffic.” It isn’t a smaller version of search. It’s a different arrangement. Search read you to send you people. The agent reads you to answer someone else, somewhere else, on a question you never even saw. You can be the best fit for what a buyer needs and never surface, because the machine couldn’t make sense of what you had.

The barbell

I’ve written before that this industry isn’t heading toward a clean roll-up. It’s heading toward a barbell.

One end is consolidation, the ordinary kind. The auction business is hitting the same moment veterinary practices and dental groups hit a decade back: aging owners with no successor, fragmented ownership, capital ready to buy on cash flow. Ritchie Bros bought BigIron. The private equity calls are landing. That end of the bar keeps getting heavier.

The other end is new. The tooling that used to take real money and a couple of years to build, a bidder portal, a catalog system, a marketing engine, can now be stood up by one sharp operator with AI for a fraction of last year’s cost. The small end of the bar gets stronger, not weaker.

What gets crushed is the middle. The mid-sized firm with neither the scale of the big house nor the software self-sufficiency of the nimble solo. Everybody’s been telling that as a money story, an M&A story. It’s bigger than that. The thing that actually springs the trap on the middle is whether anyone, or anything, can find them.

What the last twelve months did

Here’s what moved, and it moved fast.

Zero-click searches, the ones that end without anyone clicking through to a website, hit 68 percent of Google searches this year, up from around 60 two years ago. Where Google now shows an AI Overview, organic click-through falls by nearly 60 percent. The front door you optimized for two decades is quietly swinging shut.

The new door is wide open and busy. AI-referred traffic tripled year over year. Shopify, which watches a huge slice of online commerce, reported that orders coming in from AI grew roughly thirteenfold in a single year, and that those AI-referred shoppers convert about 50 percent better than the ones who arrive from a normal search. Adobe measured the same pattern across a trillion retail visits and found the thing I can’t stop turning over. A year ago, traffic that arrived by AI converted worse than human traffic. This year it converts 42 percent better.

Look at that flip again. The buyers the machines bring aren’t the bottom of the funnel. They’re the top. They show up already knowing what they want, because the agent did the sorting before they ever appeared. So if your data isn’t readable by the machine, you don’t lose a little traffic at the edges. You lose your best buyers first, to whoever the agent could actually understand.

That’s what “underwater” really looks like. The middle doesn’t bleed out slowly. It gets skimmed of its best demand, then repriced.

The call

So here’s the call, with a date on it, so you can hold me to it.

Inside the next 24 to 36 months, a meaningful share of auction companies will find themselves underwater, and the cause won’t be a soft economy or a deal they didn’t chase. It’ll be that they went non-findable. The agents doing the discovering won’t be able to parse what they have, so they’ll route the buyers elsewhere. The firm will still have good inventory, good people, a name people trust in the room. It just won’t exist to the machine doing the looking.

You can check me on it. If three years from now AI-mediated discovery is a rounding error and the firms that ignored it are doing fine, I was wrong. I don’t think I’m wrong.

Three chairs

I’ll tell you why I’m willing to say it that flat. I’m sitting in three chairs on this one, and they all show the same picture.

I run an auction company close enough to the wave to feel the pull of it. I chair the board of the national association, which means I watch the whole field and not just my own corner. And I’ve been writing this argument down for months, long enough that I now hear it come back from peers who read an early version and started using the language themselves. When the operator’s seat, the industry’s seat, and the writer’s notebook all point the same direction, I quit calling it a hunch.

Plenty of this lands harder on the middle than on the big houses, and I know who’s reading. I’m not naming names or calling a top on anybody’s business. I’m telling you what the data says is coming while there’s still room to do something about it. That’s about the only useful thing a person in these chairs can do with a view like this.

Put it down where the goats can get it

So what do you do.

You make yourself readable. Not pretty. Readable. Our world has spent years building sites that look sharp to a person and say nothing to a machine: ten broad categories, three-word lot titles, photos with no real description underneath. A human can squint and work it out. An agent can’t, and the agent is the one shopping now. Structured data, clean categories, descriptions specific enough that a machine scanning for “commercial kitchen equipment under five thousand dollars within two hundred miles” turns up yours. That stopped being a marketing nicety. It’s whether you’re in the market at all.

And you sort what you own from what you rent. The bidder list you built over twenty years, you own. The brand, the contractor network, the reputation, you own. The platform you lease, the ad traffic you buy, the borrowed reach, you rent, and rented things get taken back when the terms change. The operators who come through this in good shape will be the ones who owned their data and made it legible to the machines on purpose, before they had to.

There’s an old line around farms and sale barns. If you want everything fed, put it down where the goats can get it. Don’t hang it up high where only the tall ones reach. Set it at ground level, plain, where anything that’s hungry can find it.

For a hundred years that meant talk straight and price it clear, so the buyer in the back row understood you as well as the one up front. It still means that. But the hungry thing in the room now isn’t only the buyer. It’s the machine doing the buyer’s looking. And it’s standing right down at ground level, reading, waiting to see whose feed it can actually reach.

Put it down where the goats can get it.

I was listening to Gary Vaynerchuk on The GaryVee Audio Experience, an October 2025 episode about wasting time on people who will never buy from you.

He made a comment about bonus structures that stuck with me. His framing was intentionally sharp, but the useful part was underneath the rhetoric: employees often factor the maximum possible bonus into their lives. Once they do, the bonus is no longer experienced as upside. It becomes expected compensation.

That is a quietly dangerous design problem.

The mechanism

A discretionary bonus is supposed to create generosity, upside, and shared success. But if it repeats often enough, especially at similar levels, it changes category in practice. It moves from extra to baseline.

That shift is not irrational. People budget around patterns. If a company pays a mid-year or year-end bonus every year, the pattern becomes part of how people plan. Car payments, Christmas spending, vacations, savings goals, debt payoff, household budgets, all of it starts to absorb the expected number.

Then the next time the bonus is smaller, the employee does not experience it as less extra. They experience it as loss.

The employer is still operating in the language of discretion. The employee is operating from a different baseline. Both sides can be reasonable inside their own frame and still end up frustrated with each other.

The contradiction

That is the trap in recurring discretionary bonuses.

Recurring says, “You can expect this.”

Discretionary says, “You cannot count on this.”

Those two messages do not fit together for long. The more consistently the company pays it, the less believable the discretion becomes. The more the company emphasizes discretion, the less stable the employee’s planning feels.

The structure carries the emotional expectation of salary, the variability of profit sharing, and the disappointment risk of a pay cut.

That does not make bonuses bad. It means the category has to be honest.

The design question

I think there are only a few clean categories. Salary is stable, promised, and budgetable. A formula bonus is measurable upside tied to company or individual performance. Profit sharing has an explicit funding pool and clear business logic. A spot reward is rare, genuinely discretionary recognition that is not repeated predictably enough to become baseline.

The dangerous version is the hybrid: a recurring discretionary bonus. It asks employees to treat the money as uncertain while training them, year after year, to expect it.

Maybe the better question is not whether to pay bonuses. It is whether the bonus structure tells the truth about itself.

What is guaranteed? What is variable? What has to happen for the variable piece to exist? If the company cannot say that plainly, the structure is not a compensation strategy. It is an annual expectation problem waiting for a hard year.

What’s open

I am still not sure where the line is.

If a bonus is formula-based, it can become too mechanical and lose the feeling of generosity.

If it is fully discretionary, it may feel arbitrary.

If it is folded into salary, the company gives up flexibility and employees may forget the original upside ever existed.

But the recurring discretionary middle may be the least honest version of all. It gives the company the comfort of flexibility and gives the employee the psychological shape of a promise.

That is the contradiction.

Every number in that report is true. And not one of them answers the question the seller actually asked.

Two numbers in the room

There are two numbers in any disappointed seller’s head, and only one of them is real.

The first is what they paid. The second is what it’s worth now. The recovery rate is the gap between them, and the seller reads that gap as a loss the auction handed them. It isn’t. The loss was already there. It got baked in the day they bought.

A fellow auctioneer said it almost as an aside at an industry session: sellers are surprised. Equipment bought new during COVID is hitting the secondary market, and recovery as a percent of original cost is coming in lower than owners expected. Not lower than the equipment is worth. Lower than what they paid. The surprise is the whole signal. Surprise means the prior never updated, that they’re still measuring against a 2021 number the market stopped honoring years ago.

History rhymes here. The 2008 home seller anchored to 2006 comps and said the market took their equity. It didn’t. Their basis was set in a bubble, the recovery only looked low because the denominator was wrong, and the house was worth what the house was worth. COVID buyers paid 2021 prices the same way 2006 buyers paid 2006 prices. A scarcity premium baked into the original transaction, recorded as the anchor, carried long after the conditions that justified it were gone.

The recovery is the truth. The purchase price was the distortion. Nobody told the seller that going in.

A question nobody asked

Now look at what the marketing report is built to do.

It proves reach. It documents the campaign. It lays out everything we did to find the buyer. That answers a real question, “did we do our job,” and it answers it well. But that’s our question, not the seller’s. The seller is asking why the number is so low, and the report, by leading with how hard we worked, quietly agrees the number is disappointing and then sets out to prove the disappointment isn’t our fault.

That’s the move I’d never named until I had to. The report accepts the seller’s denominator and plays defense on it. It concedes that the wrong number is the right yardstick, then spends ten pages arguing we measured well against it.

The timing makes it worse. The report lands after the gavel. By the time it’s in front of the seller, the anchor already walked into the room unchallenged and won. The report can explain a result. It can’t reset a basis, because the basis was supposed to be reset before the bidding, not justified after it.

So call it what it is. The thing we hand the seller and call insight is closer to insurance. It covers us. Insight tells the seller something they didn’t know going in. Insurance tells them something we needed on file going out.

I reached for the easy one

Here’s the part I didn’t see until I went hunting for the fix.

This wave hasn’t fully landed on us yet. Our window catches equipment at six to ten years, so the COVID vintage is still aging into our inventory. Pharma runs younger and is already seeing it, which makes them a free leading indicator and puts us eighteen to thirty-six months out. So I started sketching what we’d actually hand a COVID-vintage seller to get ahead of it.

The first thing I reached for was a page on our website. Pandemic pricing, here’s the data, here’s what it shows. Clean. Publishable. Something I could point the whole market to.

Then I caught it. I’d reached straight past the hard thing toward the convenient one. The hard thing is our own recovery data by category, the actual hammer record from our own auctions, tailored to the asset sitting in front of one specific seller. The convenient thing is a general page anyone can browse and no anchored seller ever will. I’d done the exact thing the seller does. The seller anchors to a convenient number and won’t let it go. Handed the fix, I anchored to the convenient artifact. Same gravity, one altitude up. It caught me in the act of solving it.

That’s the tell that this isn’t really a reporting problem. It’s the pull toward the easy version of the truth, and it doesn’t spare the person who can see it.

Records before the gavel

The fix isn’t a website. It’s the record, made active.

We already build annual recovery data by category, real lots and real hammers, years deep, out of our own sale results. Right now it sits in a file and earns nothing. The move is to put it in the seller’s hands before the gavel, tailored to their category, showing the whole curve their purchase sat on top of. Here’s what this class of equipment normally recovers. Here’s the 2020 to 2021 spike. Here’s where the market sits today. Here’s where your basis lands on that line. Reset the denominator before it walks into the room.

Notice the timing flip. The marketing report shows up after the sale to explain the gap. The category report shows up before the sale to remove it. One does damage control. The other makes the damage unnecessary, because a seller who understands the premium they paid going in is a seller who isn’t surprised on the way out.

And it changes what we think we sell. The seller believes they’re hiring us to find buyers. Finding buyers is table stakes. What we actually sell is the reset. It’s the appointed day a number someone has carried for years gets laid against a record they can’t argue with, and whatever doesn’t tie gets corrected in an afternoon. The category report is that same reset, run early and in private, before the public one happens at the gavel. We don’t sell the gavel. We sell the number they can finally trust.

None of this comes free. Walk a seller through the curve before the sale and you’re telling them their number is too high before you’ve earned them a dollar, which is a strange way to win a listing. Some won’t want to hear it, and it only works if the data is ours and clean enough to stand on. I’d rather lose that argument before the gavel than after it, but I won’t pretend the early version is the comfortable one.

What else are you covering for

None of this is unique to recovery rates, and it isn’t unique to us.

Every firm has a deliverable it files under client service that’s really self-protection. The report that proves we tried. The update that documents we called. The recap that lands a day after the decision it was supposed to inform. They all pass the same test the marketing report passes: every line is true, the timing is defensive, and the whole thing answers our question instead of the client’s.

So here’s the one I’m sitting with, for our shop and for anyone else running sales for a living: what are you handing clients that covers you instead of informing them, and could the same data, moved earlier, do the opposite?

The report that explains the result is insurance. The one that changes it is the product.

In business, commentary is the ultimate convenience. It is easy to write a forecast, spin a narrative, or make a decision because “it makes sense at the time.” But commentary naturally drifts and flatters us, and human memory decays too quickly to defend the truth. A typo can sit wrong by $1.2M for eleven months because nobody was forced to check the record.

We can’t rely on willpower or manual documentation to catch this. The only solution is structural: we must design scheduled collisions where commentary is mechanically forced to meet the raw system of record.

But this structural yang has a yin. The most valuable decisions in a business are often “grey-hair” calls. These are the instinctual, non-data-driven decisions born of decades in the trenches. If you outlaw instinct, you strip the business of its soul and its competitive edge.

The unlock isn’t choosing between spreadsheets and gut calls. It is building a system where they live in harmony. You don’t block the grey-hair risk. You simply ensure that every time instinct deviates from the record, the deviation is auto-flagged and documented in an easy-to-find way. You let the gut lead, but you force it to keep its standing appointments with the truth.

Open questions

1. How do we build automated flagging into our daily operations without creating notification fatigue that people eventually ignore?
2. What is the most critical instinctual decision in our pipeline today that is running without a scheduled date to face its record?
3. How do we design the user interface of our tools to elevate these flags constructively, keeping the Socratic friction helpful rather than annoying?

Connects to Cognitive Defense Against Default Drift and Systemic Rivals Against Default Drift.

It’s a blunt, uncompromising way to look at product design. It frames customer support not as a service department, but as a tax you pay for bad engineering. If your user needs an explainer or a help desk, you didn’t finish designing.

At first, that sounds like typical Silicon Valley idealism, a luxury for simple consumer software. But when you look closely, the principle is deeply diagnostic. It forces you to ask: which parts of your business genuinely require human interaction, and which parts are just friction you haven’t had the courage to design out?

The silence of a ringing phone

There’s a deep ideological divide in how we define a healthy business.

One worldview, the traditional one, measures operational health by contact. A ringing phone is proof of life, relationship, and hustle. To an old-school operator, silence means death. It means nobody is buying, nobody is calling, we’re not growing.

The other worldview, the platform one, measures health by silence. A phone call is an alarm. It means a user is lost, a workflow is blocked, or a default was wrong. In this view, a support ticket is a trailing indicator of a system failure upstream.

This is a fundamental clash of worldviews. When your business partner looks at a quiet room and thinks there must be problems, and you look at a ringing phone and think the same thing, you’re not even running the same company. You are operating on two different definitions of trust. One relies on human presence to smooth over the bumps. The other relies on a seamless system to prevent the bumps in the first place.

The shock absorber trap

But here’s the trap. Even when you believe in the system-first worldview, your behavior often defaults to the other side.

I know this because I live it.

I make sure that every negative review and every low net promoter score comment from our customers comes directly to my personal email inbox. They don’t go to a support queue. They land in front of me, and they get a personal response from me. It’s a grueling, exhausting process. It’s frustrating, overwhelming, and draining.

For a long time, I told myself this was craftsmanship. I was the CMTO, staying close to the ground, keeping my hands on the wheel.

But if I hold my own actions up to the system-first principle, the truth is much less heroic. By routing those failures to my personal inbox, I’m not fixing the company. I’m acting as a human shock absorber. And even worse, I’ve been using those reviews as a personal shield. It feels like craft and responsibility, but it’s actually an escape. By personally responding to each complaint, I avoid the much harder, more exhausting work: aggregating those failures, sitting down with our executive team, and forcing us to develop the systems and processes to fix them at the root.

When you throw yourself into the gears of a broken system to keep it running, you’re not solving the problem. You’re just masking it. And as long as you’re willing to absorb the friction, the organization has no incentive to redesign the engine.

The concrete floor of physical operations

This trap is easy to spot in software, but it gets incredibly messy when your platform meets physical reality.

At Grafe Auction, we run industrial liquidations. Our software handles registration, bidding, and payment processing. But eventually, the digital platform hits the concrete floor of a physical warehouse. That’s the post-auction removal process. It’s where buyers show up with trailers, forklifts, and rigging contractors to claim heavy equipment.

This is a massive point of friction. Buyers get frustrated. Subcontractors get brash. People argue about lot locations and missing parts.

Our instinct is always to throw more humans at the problem. I’ve spent hours wondering if we should place a Grafe customer success person on every site. We’d find a “softer” person, someone who understands hospitality, to act as a buffer between the brash contractors and the frustrated buyers.

But that is the shock absorber trap all over again. It treats the friction on site as an unchangeable law of nature.

When we look past the personalities, the actual failure is a system design issue. The removal site is a poorly managed situation. It’s a custody and verification failure. Because there are no physical checkpoints or clear verification systems, buyers can switch lot tags, take the wrong items, or damage assets.

Putting a polite success person on site doesn’t fix the custody loop. It just gives the frustrated buyer a nicer person to complain to while they stand in the middle of the chaos. We tried to fix a fundamental security and layout failure with a scheduling tool, requiring appointments to try and slow the process down. But appointments are just a speed limit on a broken road. They don’t fix the holes.

The internal cost of external simplicity

So why do we do it? Why do smart, system-minded leaders repeatedly choose human shock absorbers over system redesigns?

Because throwing humans at problems is easier than doing the hard work of system building.

Redesigning a physical or digital process requires time, energy, and relentless effort. It means holding people accountable. It means creating better tasks, workflows, and strict verification gates. And most of all, it requires getting buy-in across an entire enterprise.

Getting that buy-in is exhausting. When you’ve fought to get everyone on the same page and watched it fail repeatedly, you get tired. Over time, you become jaded toward change. Change is hard.

That’s when the shock absorber trap becomes incredibly seductive. Personally replying to a negative review or putting a success person on site takes energy, but it’s a closed loop. You control it. You don’t have to negotiate with anyone. You don’t have to fight for buy-in. You just do the work.

In that light, the shock absorber trap is not a design mistake. It’s an executive coping mechanism. We build human shields to protect ourselves from the exhausting work of getting everyone on the same page.

But the cost of that shield is compounding. As long as we keep absorbing the shock, we keep renting our growth. We spend our days firefighting, and the systems stay broken.

The real work of leadership is not being a hero in the support queue. It is having the courage to step out from behind the shield, aggregate the failure data, and force the organization to sit down and redesign the system. If we want to build a platform that scales, we have to stop smoothing over the bumps and start fixing the road.

The only way to survive this is active resistance. For myself, I’ve built custom agentic tools to force Socratic friction into my writing. But as a leader, the question is broader. I can’t build custom vaults and code-level friction for everyone in the company. Most people just want the frictionless path to get the task off their plate.

I think the leader’s job is to shine a light on this drift. You don’t build a cognitive wheelchair for your team to make them comfortable. You build a team of rivals.

Abraham Lincoln did this with his cabinet, surrounding himself with people who challenged his views. I think we have to do the same thing with our technology and our habits. We have to design systemic rivals: Socratic agents that push back, hard books, and podcasts that stretch us. It’s hard, but it’s the only way to ensure we actually deliver our full measure.

What Emerged

During our dialogue, several sharp formulations crystallized the tension:

• “To jump and ask how high on the way up, as my dad would say. To be sure that everyone that encounters me gets the full measure.”
• “I think they’re getting the full measure of me and maybe fuller because of conversations that are socratic, the push, that make me consider alternatives.”
• “If one defaults to lazy, you’ll gain weight, you’ll not read, you’ll not sharpen your brain, you’ll decline. It’s human law.”
• “My job as a leader is to bring continual light shining to the default drift human law, to ensure we’re always playing the advocate for the unheard, we’re always surrounding ourselves with a team of rivals.”
• “I also seek out alternative viewpoints through books and podcasts. Sometimes that’s hard, but I get better as a result.”

Open Questions

1. How do we introduce systemic rivals into daily workflows for team members who don’t have an internal drive to jump high?
2. When does a custom skill become so comfortable that we stop treating it as a rival and start treating it as a default?
3. How do we balance the need for Socratic friction with the operational velocity required to run a business?

Connects to Cognitive Defense Against Default Drift.

I need a triple-layered cognitive defense system:

1. The Tactical Defense (The Code Layer): Encoding Socratic pushback, friction, and accountability directly into the custom agentic skills I build. I don’t just open a ChatGPT tab; I build systems designed to actively resist my own laziness.
2. The Contextual Defense (The Data Layer): Operating under the law of reciprocity. Deep sparring requires deep context. If I don’t build a dense, custom vault of actual thinking, the system has no leverage to push back, and the dialogue collapses back to generic out-of-the-box templates.
3. The Meta-Cognitive Defense (The Mindset Layer): Active self-awareness of my own blindness and non-original thinking. Recognizing that I am naturally prone to low-effort defaults, and deliberately hunting for those defaults inside my daily work.

Without these layers, I am just defaulting to convenience and calling it innovation.

What Emerged

During the dialogue, several sharp formulations crystallized the tension:

• “Most people are defaulting to easy peasy chat with the AI, and expecting whatever is spit out as the silver magic bullet.”
• “The default for a user of this type has to be pretty low, and non-self-aware/accountable.”
• “I intentionally build in accountability, socratic sparring, with every skill I build. And I build skills and agentic rather than rely on chat alone.”
• “I’m very big on trying to be aware of my blindness, and non-original thinking.”

Open Questions

1. How do we measure the “friction index” of a skill to ensure it forces the right amount of cognitive load without slowing down essential velocity?
2. When does a custom skill itself become so comfortable that we stop questioning its structure, shifting our “low default” up one level of abstraction?
3. How do we train other people in an organization to develop the active meta-cognitive self-awareness needed to notice their own blind spots, rather than just relying on the tools to do the thinking?

Connects to Systemic Rivals Against Default Drift.

The first one I’ve lived. Marketing agencies typically charge a percentage of marketing spend as their fee. The agency makes more money when the client spends more money. The incentive is to grow the budget, not the results. I’ve been quietly dismantling that arrangement in our vendor stack for the last couple of years.

The second one I heard this morning. Pavel Durov on the Lex Fridman Podcast (#482, Sept. 30, 2025), making the point that drug companies aren’t actually incentivized to find the root-cause cure. They’re incentivized for you to keep taking the drug. Recurring revenue. Maintenance, not resolution.

The two examples sat next to each other and a question formed. How many of your vendors are paid in a way that incentivizes them against you, and you haven’t noticed?

I want to take this further than the easy version of the essay, because the easy version (incentives are everywhere, audit your vendors, Munger said something about this) doesn’t actually help anyone. The harder version is that fixing the fee structure isn’t enough. And when you do the audit honestly, it reaches you too.

The fee structure is the strategy

That was my working theory for a while. Both of the marketing agencies we work with as advisors are now on a flat monthly fee, independent of our marketing spend. That move solved the original problem. They don’t make more money when our ad budget grows, so the suggestion to “increase the Facebook spend” no longer comes wrapped in self-interest.

But here’s what I didn’t see at first. Flat fee creates the inverse problem. Once you’re getting paid the same amount whether you answer the phone or ghost the client for three weeks, the rational move is to do the minimum to keep the contract alive. Different incentive, same structural issue.

What actually keeps those advisors productive isn’t the flat fee. It’s that I hold them accountable directly. I ask for the work. I challenge the recommendations. I expect responsiveness. The fee structure removed one misalignment, and accountability covers the new one.

The principle started forming there. Fee structures aren’t self-policing. They never have been. The structure sets the default behavior, but the actual outcome depends on something outside the structure doing the work.

The audit reaches you

Here’s where it gets uncomfortable.

The auction industry runs on commission. We typically take a percentage of the sale, plus a buyer’s premium. I’ve always thought of this as the most aligned fee structure in business. Seller wants the most money for the asset, auctioneer wants the most money for the asset, both parties pull the same direction. Clean.

When I audited it the way I’d audit someone else’s industry, the alignment got messier.

There’s a marginal effort cliff. Working harder to push a $500K auction to $550K nets the auctioneer maybe $5K at 10%. The seller cares about that last $50K. The auctioneer’s incentive flattens at the top of the bidding band well before the seller’s does.

There’s the absolute-vs-reserve question. The auctioneer almost always prefers absolute, because it guarantees the sale closes and the commission gets paid. The seller might prefer reserve, because it protects against a bad day in the room.

There’s the buyer’s premium split. We’re paid by both sides, which means when we’re advising a seller on whether to accept an offer, we’re not 100% on the seller’s side. We’re also thinking about the buyer who’s coming back next month for a different auction.

And there’s lot consolidation. On a 1,200-item industrial auction, the seller’s interest is probably best served by cataloging each item individually. That maximizes search visibility, maximizes specificity, probably nets 10% more gross. Our operating reality says we consolidate similar items into bigger lots, drop the lot count to 300 or 400, and we tell the seller that item-level cataloging means higher non-commissioned setup fees.

That conversation is real. The seller does see the choice. But the architecture of the decision still reflects something the seller probably doesn’t know. Setup typically runs 14 days before the auction. Payment comes 14 days after it concludes. Our crew’s labor sits uncompensated for the better part of a month. Every day spent building item-level catalogs is a day they’re not on the next project. Consolidation isn’t preference. It’s the model surviving itself.

Misalignment is load-bearing

That last sentence is the part I want to defend.

The misalignment in the auction commission model isn’t malice. It isn’t greed. It’s the way the business model actually works. If we cataloged every widget on every job, the float on uncompensated setup labor would break the business. The misalignment is what makes the model viable in the first place.

The same is true everywhere I look.

The marketing agency on a percentage of spend isn’t evil. That’s how they pay their account managers between client wins, and how they smooth revenue across an industry where contracts open and close on uneven cycles. Take away the spend-based fee and they have to charge retainers that most clients won’t agree to.

Pharma’s recurring-revenue incentive partly covers the cost of every drug that fails in trial. Most candidates never make it to market. The companies that get to a cure have to fund the ten that didn’t. A pure cure-or-nothing pricing model would either price drugs out of reach or kill the development pipeline.

Lawyers billing hourly have a famous incentive to drag the work out. It’s also how they price the case they can’t predict the length of, and how they absorb the risk of the matter taking three times as long as either party expected.

Misalignment isn’t always a flaw. Sometimes it’s the load-bearing wall the building rests on. You can’t just remove it.

There’s a fine line between honoring the client and honoring the model. We walk it on every job. Drift is constant, and the default drift is away from the client.

Which means the only honest move is to design the override.

The override

If the fee structure can’t be self-policing, and the misalignment can’t be removed, then the only thing left is whatever you put in place to override the misalignment when it matters.

Look at what’s already there at Grafe and the pattern shows up. Transparency, because we surface the cataloging tradeoff to the seller in the form of a setup fee they can opt to pay. Accountability, because I’m the one keeping the advisor relationships productive. Repeat-business reputation, because every seller in the industry talks to every other seller, and bad outcomes travel.

None of those overrides are part of the fee structure. They’re all things that sit on top of it.

Here’s the part I owe myself the honest answer to. Are those overrides designed, or are they accidental? When we tell a seller the setup fee goes up with item-level cataloging, are we architecting a check on our own incentive, or are we just being fair dealers because that’s how we’d want to be treated? The honest answer is mostly the second one. The overrides are real, but they’re reflex, not architecture. They’d disappear the moment any of us got distracted, or the moment somebody with a different temperament ran the desk.

That’s the gap. Reflex doesn’t transfer. Reflex doesn’t audit. Reflex doesn’t survive a stressful quarter or a bad hire. If the overrides are the only thing keeping the structure honest, the overrides have to be designed, not assumed.

The friction test

The sharper version of this came out of the pharma example.

The reason pharma’s misalignment wins despite the FDA, despite malpractice law, despite the generic competition that eventually shows up, despite physician discretion, is that all of those overrides ask the patient to do harder work than swallowing a pill. The lifestyle change, the diet rework, the behavioral intervention. Each one is harder than the medication.

That’s the real test for an override. Not “is it present?” but “is it easier than the default it’s overriding?”

Run each of the Grafe overrides through it and they don’t all hold up.

A flat-fee advisor relationship works as an override on the percentage-of-spend incentive, but only if the client actually engages and asks for the work. If the client is too busy to push, the flat fee turns into pure margin for the agency and the override stops doing anything.

A transparent setup fee works as an override on the lot consolidation incentive, but only if the seller knows what to ask for. If they don’t know there’s an item-level option in the first place, the transparency is theoretical.

Repeat-business reputation works, but only if the buyers and sellers actually talk to each other. In a fragmented market with no shared communication channel, the reputation override breaks down.

The presence of an override doesn’t matter. The friction of the override relative to the default does.

Trying to be a fair dealer

If I’m honest about what’s actually keeping our fee structure aligned with our clients, it’s that we’re trying to be fair dealers. That’s the override. Not the commission percentage, not the buyer’s premium structure, not the line item on the setup fee. It’s us, deciding case by case that the seller’s interest should win the close calls.

That’s a fragile system. It works while we’re paying attention. It doesn’t scale, it doesn’t transfer to a new hire, and it doesn’t survive a stressful month.

The principle the audit produced for me is this. Every fee structure has a misalignment baked in. The misalignment is usually load-bearing. The only thing that can override it is something built on top of the structure, designed deliberately, made easier than the default it’s competing against. Reflex isn’t enough. Good intentions aren’t enough. The override has to be intentional, visible, and operationally cheaper than the misaligned path.

If that’s true, the question every business owner should be asking isn’t “how do I find the perfect fee structure?” That question doesn’t have an answer. The right question is “what is the misalignment in my fee structure, what override am I depending on to correct it, and is that override actually easier than the default behavior the structure rewards?”

I owe myself a real audit of that. I suspect most of the overrides at Grafe are sitting in our heads as reflex, and we haven’t done the work to design them into the operating system. The auction model is more aligned than the agency model. It’s also less aligned than I’ve been telling myself.

The fee structure isn’t the strategy. The override is.

Six months later, Google announced WebMCP and the agentic web, the standard protocol for AI agents to act on structured web data. A protocol I’d never heard of, designed for a use case I hadn’t predicted. And the platform I’d been building already fit it. Not because I saw the future. Because the principle didn’t depend on seeing the future.

That’s the small version of the pattern this essay is about. I’ve been writing about adjacent things for ten weeks. Four observations, four different parts of the work. Only when I lined them up did I see they were the same shape. Each one was about something I had built without trying to build it. Each one turned out to be durable in a moment when the rest of the craft was getting cheap.

This essay is about that shape.

The current AI question is what it replaces. That’s the wrong frame. AI is going to commoditize whatever can be commoditized, and the commoditization is not personal. The more useful question is what it can’t. Which parts of the work hold their value when execution gets cheap? Which edges survive when the surrounding craft cheapens?

Four answers. Each from a different altitude. Each with the same mechanism underneath.

Principles

Back to the August 2025 thought card. I didn’t write it because I predicted WebMCP. I wrote it because “touch it once, make it findable by any system” is just good practice. “Computers do the mundane, humans critically think” is just a sound division of labor. These are principles, not predictions.

That’s what good principles do. They’re robust across futures you didn’t predict. Predictions are fragile. They’re right or wrong, and the world rarely cooperates with specifics. Principles are antifragile because they’re grounded in how things actually work, not how you think they’ll play out.

The companies that survive transitions aren’t the ones that saw them coming. They’re the ones whose principles already pointed where things were headed.

Principles get weight by being tested under pressure. The principle that survives a contract dispute, a buyout offer, a public mistake, a quarter where the easy move would have been to compromise it. That’s the one with weight. The rest are preferences dressed up as wisdom.

AI can write a principle. It can’t be wrong about a principle and then keep going.

Operator Leverage

The auction industry is at its veterinary moment. Aging owner-operators with no successor. Fragmented ownership. Private equity buyers with capital ready to underwrite on cash flow. The exact triad that took veterinary EBITDA multiples from 4x to 12-15x in a decade. Demographic inevitability is meeting financial appetite, and the offers are starting to land.

A second wave is running in parallel: the platform layer is being unbundled by AI in real time. A platform stack used to take years and serious capital to replicate, which is what made platform aggregators worth premium multiples. That math is dead. A solo operator with the right tools can stand up a bidder portal, catalog system, and marketing engine that would have been a major project two years ago. The platform vendors paid for yesterday’s moat. The recoup window they bought is compressing.

What you get isn’t a roll-up. It’s a barbell.

At one end, large consolidated houses with operator leverage that AI doesn’t replicate. National business development pipeline. Brand built over decades. Bidder lists in the hundreds of thousands. Contractor networks. Compliance infrastructure that took years to assemble. At the other end, AI-empowered regional and solo operators who used to need a platform vendor and now don’t. The squeezed middle is the traditional mid-tier firm with neither scale nor software self-sufficiency. That’s the cohort that gets repriced.

Operator leverage accumulates over decades of showing up. The bidder list isn’t a database. It’s a record of every auction someone bought from us. The contractor network isn’t a vendor list. It’s every removal someone delivered on time. The brand isn’t a logo. It’s the sum of decisions people remember us making, the calls we made when they were watching.

AI can run the auction. It can’t have been the one who built the trust that made the contractor pick up the phone.

Doctrine

Three high-stakes decisions in three weeks recently routed through the same AI strategy pipeline I built into my own tooling. A bid retraction notification. A first-of-its-kind hire. A grievance screening with no precedent. Different stakes, same mechanic. And the strategy note from the third said it out loud: “the letter itself serves as the v0.1 screening rationale precedent.”

Not the framework that gets written later. Not a separate doctrine document with a deadline. The disposition letter is the precedent. The decision and the doctrine are the same artifact.

This is what the pipeline does. Board phase produces a panel-by-panel argument. Simulate produces a 90-day scenario. Council produces named dissent. Strategy produces reconciliation. By the time the call gets made, there’s a written argument trail that explains why, not just what. The next decision in the same neighborhood doesn’t start from zero. It starts from “see how we handled the last one.”

What’s emerging is starting to design the artifact on the way in. The disposition letter has a different reader than the obvious one. The complainant gets the answer. The future complainant gets the framework. That changes what the letter says, and how it’s structured, and which test gets foregrounded.

Doctrine accumulates by surviving real complaints. A documented framework that came out of an actual decision has weight that a policy written in the abstract never has. It survived a real edge case. It got read, tested, cited. The next person making a similar call doesn’t get a memo titled “How We Handle X.” They get the actual letter and the actual reconciliation, and they read it the way lawyers read case law: what did the deciders do when they were under pressure.

AI can draft a screening framework. It can’t be the deciders who were under pressure.

Sacrifice

The framing comes from a Gary Vaynerchuk bit I wrote about yesterday. 1-percenter mode runs four dials, not one. Hustle, grind, output, sacrifice. The first three get all the airtime. They’re measurable, performable, brag-able. Productivity tools have made each one observable in real time. Public hustle culture has commoditized the performance of effort.

The fourth dial is the one nobody else runs.

Sacrifice accumulates as life shape. The cumulative subtraction of what you didn’t do. The other career you didn’t take. The hobby you stopped having time for. The version of yourself you didn’t get to develop. It doesn’t show up on a balance sheet or a feature comparison. It shows up in the gap between what someone built and what they could have built if they’d done other things instead.

AI doesn’t have an alternative life it gave up.

Calibration

Each of these edges has the same failure mode. The thing that made it durable also contains its risk.

Principles harden into walls you can’t see. Doctrine starts optimizing for the future reader and stops answering the person it was written to. Operator leverage becomes the institutional bloat the next generation has to clear out. Sacrifice eats the version of you the 1% life was supposed to produce.

The edges are durable. The risks are real. The work is reading both, not betting on one.

AI eats execution. It won’t eat the critical thinking underneath.

“I just don’t understand how people are asking for a 1% life without understanding it takes 1% sacrifice effort. You have to be a 1-percenter if you want a 1% outcome…”

Gary Vaynerchuk, The GaryVee Audio Experience, “Starting a Business, Building Brand and Overcoming Doubt,” September 25, 2025

Heard this Vee bit yesterday and the word “sacrifice” kept rattling. Sacrifice is the word doing the work in the quote. A 1% life takes hours, but the hours are downstream of what you’re willing to do without. What you don’t get to have, who you don’t get to be, what you don’t get to chase. That’s the dimension the hustle genre keeps missing. 1-percenter mode has two faces: the things you add (hustle, grind, output) and the things you subtract (sacrifice). The genre talks endlessly about the first. Vee names the second. My read is sacrifice is the word that completes the picture, the thing left behind, not just the thing being added.

The trap underneath is the assumption that effort maps to rank one-to-one: 20% effort gets a 20% life. The math doesn’t run that way.

Rank is relative. 1-percenter mode runs four dials: hustle, grind, output, sacrifice. Top 20% means running every dial higher than 80% of people. Top 1% means higher than 99%.

The top is defined by the people willing to put in what nobody else does and give up what nobody else will. 20% on any one dial doesn’t put you in the top 20%. It puts you in the bottom 20%, because 80% of people are running it higher.

Whatever effort you’re running feels like top effort from inside it. The ceiling on what feels hard is nowhere near the actual top. This is the deciding vs defaulting failure mode applied to effort: busyness as the convincing disguise for a level of input you never actually picked.

Vee himself is the proof of concept. He built the content factory, gets up at 5am, runs the agency, the podcast, the speaking calendar, and a public posture that doubles as the marketing. The quote is diagnostic, not aspirational. He’s describing what it took to become him. Most people only see the outcome. They miss what he gave up to produce it.

What’s Open

• Where does sacrifice break the other direction? At some point the returns diminish, or it starts breaking things you actually need to be the version of you the 1% life was supposed to produce. The 99%-effort tier eats people who didn’t notice what was being subtracted along the way.
• Is sacrifice always the right frame? Some people seem to do 1-percenter work without the felt sense of giving anything up, because they want the work more than the alternatives. For them it’s preference, not sacrifice. Maybe the framing is diagnostic for one type and a misread for another.
• What about the inverse, the 99% life from 1% effort? That’s the person who built the right asset early and now compounds while doing less. The inversion only describes the upward direction. The downward case is just as real.

The honest version of the legacy claim is smaller and harder. It’s not “I built a great system.” It’s “I knew the way my own thinking gets stuck, and I built a tool that breaks me out of it.”

That’s not best practice. It’s personal-weakness corrective. And the durable thing isn’t the tool. It’s the meta-move.

The legacy reckoning

Seventy-two days from the end of a term I’ve held for a year. The temptation is to say “I tried to plant the seeds” and “I hope I left good examples.” Both are true. Both are also the soft version of the question.

The hard version is this. Most institutional knowledge does not live on. It gets lost in the shuffle. The next CEO might pick up the file. They might not. The next board has its own priorities and its own lock-ins. Trusting that the values transfer through osmosis is a strategy. In my experience, it’s just not a strong one.

So the question I keep circling is this: what actually survives a term? And what should I be writing in the next seventy-two days that has any chance of being read in seven years?

Saying versus doing

Six qualities I’d want carried forward. Transparent. Equitable. Research-led. Devil’s advocate against groupthink. Informative. Open.

Five of those are common. Most boards say they want them. They show up in mission statements and never get audited.

The sixth is the one with edges. Inhabit the opposing position. Argue the other side genuinely, not as a debate exercise. Make yourself uncomfortable on purpose, because the room is otherwise drifting toward consensus and calling it harmony.

Most leaders don’t actually do this. They might run a token “what could go wrong” round. They don’t put themselves inside the position they’re about to reject and argue it the way its strongest proponent would.

That’s the one I’d want to leave behind. It’s also the hardest one to leave, because it lives in temperament, not in process.

The fault under the practice

Even the practiced version has limits. When I’ve been over a question before, when something has settled in my head, I have a hard time budging. The default mode anchors. The devil’s-advocate move that works fine on fresh questions hits a wall on questions I’ve already decided.

So I built a process around it. Board panels argue the call across seven frames. A simulation runs the decision forward as a 90-day scenario with named characters. A council stress-tests the result with adversarial roles. Every run produces a written artifact that has to defend itself.

The pipeline isn’t best practice. It’s correcting for a specific weakness I know I have. The architecture exists because I know my default mode locks in.

The meta-move

This reframes what the pipeline actually is.

It’s not a generic decision tool. It’s cognitive infrastructure tuned to my specific lock-in. The same way Marcus Aurelius wrote his Meditations to correct for his particular temperament. He wrote “what stands in the way becomes the way” not as universal philosophy, but as a reminder he personally needed.

That changes what’s transferable.

The tool might survive my departure as a procedure. Someone else could run the boards, the sims, the councils. They’d get something out of it. Maybe less, because the design was tuned to my specific lock-in mode, not theirs. But the procedure works as scaffolding even when the original temperament leaves.

The meta-move travels differently. Know the way your own thinking gets stuck. Build the tool that breaks you out of it. That’s portable to any leader, in any room. The specific tool doesn’t have to look anything like mine.

What survives, by mechanism

Three durability layers, each with its own carrier.

Artifacts survive as samples. The decision letters, the strategy memos, the policy specs. They’re citable if someone goes looking. But they require curators. They sit in archives and depend on the next person knowing they’re there.

Processes survive as scaffolding. The pipeline can be picked up and run by people who don’t share my temperament. They’d run it differently. They’d get different sharpness from it. The procedure works without the original posture, just with reduced fidelity.

The meta-move survives as principle. “Know your lock-in mode and build the tool that breaks it” doesn’t depend on anyone preserving anything. It’s small, portable, and it doesn’t care who’s holding it. Someone reading this in seven years can apply it without ever knowing the specific tool I built or what term I was finishing when I wrote it.

The portable thing isn’t my tool. It’s the principle.

The inversion

The honest version of the legacy claim isn’t “I left good examples.”

It’s: I left a record of someone who knew his lock-in mode and built the tool that broke it. That’s the example. Not the specific tool. The honesty about the weakness, and the specific corrective built around it.

Anyone reading it has the same job. Find your own lock-in mode. Build your own tool. The mechanism transfers, even though the shape of the tool will look nothing like mine.

That’s a smaller claim than “I changed the institution.” I think it’s also more likely to be true. And it’s available to anyone leading anything, regardless of altitude or title.

The seed isn’t the artifact. The seed is the example of someone willing to name the weakness and build around it. The tool is just the proof that the example was real.

What’s the tool that breaks you?

For years, each campaign started the same way: a blank page. Someone opens the lot catalog, scans through it, figures out what’s worth highlighting, writes the ad copy, picks the photos, formats everything for Facebook and Google, selects the target markets. Then does it again tomorrow. And the day after that.

The work was good. It was also completely dependent on whoever was doing it. A different person, a different day, a different mood, and the output shifted. Not dramatically, but enough. Enough that quality was a function of who showed up, not what the system produced.

As our deal flow kept climbing, the math got simple: hire a marketing person or build a system. I chose the option that didn’t need an interview process. The insight wasn’t that the work could be automated. It was that the work already followed a pattern. We just hadn’t named the pattern yet.

So we named it. Then we codified it. Then we built a single skill that runs the entire marketing pipeline. One command. Copy, photos, targeting, formatting. Here’s what we learned building it.

Start with the data, not the blank page

The old workflow started with a person reading a lot catalog and deciding what mattered. The new one starts with the catalog itself.

When a campaign kicks off, the system pulls the full lot catalog from our auction platform’s API. Every lot, every title, every category. Then it does what a good marketing person would do if they had unlimited time: it reads every single listing, identifies the top-value items, breaks down the category mix, and builds a profile of who’s going to bid on this stuff.

That analysis happens before anyone writes a word of copy.

Here’s why that matters. A 300-lot manufacturing auction might be 40% CNC equipment, 25% welding, 20% material handling, and 15% office furniture. The copy should reflect that mix. The photo selection should reflect that mix. The geographic targeting should reflect that mix. When a person is scanning a catalog by hand, they’re doing this intuitively. Sometimes they nail the proportions. Sometimes the first interesting item they see dominates the whole campaign because that’s what caught their eye first.

The system doesn’t have that bias. It reads everything, counts everything, and hands the analysis to the next phase as structured data. The creative work that follows starts from a complete picture, not a first impression.

The human still makes the calls. They just start somewhere better than a blank page and 300 lot titles they skimmed 20 minutes ago.

Define quality as a measurement, not a feeling

“Does this ad copy look good?” is a question that gets a different answer depending on who you ask and what time of day it is. We replaced it with questions you can actually check.

Facebook primary text: 125 characters or fewer. Not approximately. Not “around 125.” The system counts every character and rejects anything over the limit before a human ever sees it. Google Ads headlines: 30 characters max. Descriptions: 90. These aren’t suggestions. They’re hard ceilings enforced by the platforms, and they’re enforced in our system before the copy leaves the pipeline.

Character counts are the easy example. The harder one is photo selection.

Picking marketing photos used to be entirely subjective. Someone would scroll through 60 images and pick the ones that “felt right.” Which was fine until you compared Tuesday’s picks to Thursday’s picks from the same person. Now the system runs a two-pass review. First pass is triage: keep or skip, based on defined criteria. Storage unit backgrounds? Skip. Blue tape or lot stickers visible? Skip. Sideways orientation? Skip. Good lighting, real-world setting, recognizable brand? Keep.

Second pass ranks the keepers against a priority list: hero shot potential first, then visual punch, brand recognition, category diversity (does the photo selection represent the full breadth of the auction?), composition quality, and platform versatility (will it work as both a square and landscape crop?).

None of that is subjective once the criteria exist. It’s a rubric. Two people applying the same rubric to the same photos will get closer to the same answer than two people following their gut. And when the rubric is wrong (which happens), you fix the rubric. The fix applies to every future campaign, not just the one where you noticed the problem.

Quality becomes a property of the system, not the operator.

Build decision gates, not revision cycles

The most expensive mistake in a marketing pipeline isn’t bad copy. It’s bad copy you don’t catch until the end.

Here’s the old pattern: generate everything, present it all at once, get feedback, revise, present again, get more feedback, revise again. Each cycle burns time. The deeper the revision, the more time burned. And the later you catch a misalignment, the more work you throw away.

The skill has a steering check built into the middle of the pipeline. After the catalog analysis but before the expensive work (copy generation, photo selection), it pauses. It shows the auction theme, the top three highlights, and the category breakdown. Does this match what you know about this sale?

Thirty seconds. If the analysis is off (maybe the seller has a piece that isn’t reflected in the lot titles, or a whole category that should be de-emphasized), the correction happens before any downstream work begins. One adjustment at the steering check saves three rounds of revision at the end.

This is a principle that applies well beyond marketing. Any pipeline where the expensive work comes later should have a cheap check up front. The cost of a 30-second pause is trivial. The cost of revising 50 search keywords, 15 ad headlines, 10 descriptions, and a full photo set because the theme was wrong from the start is not.

Let the system carry the volume

A single campaign through this pipeline takes about 10 to 15 minutes. My active involvement is maybe 60 seconds: confirming the steering check and approving the final output. The rest is the system doing what I’ve already decided it should do.

That’s the part that changes everything at 290 campaigns a year. Campaign quality holds steady, maybe more consistent. What changes is that the volume becomes sustainable without scaling the team.

Before the skill, marketing capacity was directly tied to headcount. More campaigns meant more people or more hours. Now it’s tied to the system. The same person can run five campaigns in an afternoon because the decisions are encoded, the quality checks are automated, and the creative work is seeded with real data instead of starting from scratch.

And the system compounds. Every correction (a rubric tweak, a new skip criterion for photos, an updated character limit that a platform changed) improves every future campaign. A team of people making individual creative decisions doesn’t compound that way. Each campaign starts over. Each person carries their own lessons. Institutional knowledge lives in people’s heads instead of in the system.

The skill file is a markdown document. It describes the workflow in plain text. When the workflow changes, I edit the file. The next campaign uses the updated version. No deployment, no build step, no waiting for someone to push an update. It’s a living document that gets better every time it runs because the corrections are structural, not personal.

What actually changed

The copy didn’t get dramatically more creative. The photos didn’t become art. What changed is that campaign 290 is as good as campaign 1. On a Tuesday afternoon in November, under deadline, with three other auctions competing for attention, the output holds.

That’s the thing about treating a creative process as an engineering problem. You don’t lose the creativity. You just stop depending on it. The system provides the foundation (correct data, enforced limits, structured criteria), and the human provides the judgment that the system can’t: is this the right theme, does this photo tell the right story, does this copy sound like us. The split is clean.

Most businesses running high-volume marketing haven’t made this shift yet. The technology exists. Treating each campaign as unique just feels right. It feels like you’re paying attention. But attention doesn’t scale. Systems do.

If you’re running 20 or 50 or 200 campaigns a year and each one starts with a blank page, the question isn’t whether you can do it. You clearly can. The question is whether you should.

But we weren’t hiring. We were acquiring.

The distinction matters, and I think it matters more in the auction industry than most people realize.

When you hire someone, you’re renting their skills for as long as they stay. The moment they leave, whatever wasn’t documented or embedded in a system walks out with them. That’s not cynical, it’s structural. I wrote about this a few weeks ago in a different context: the relationships someone builds on your platform become yours, but the person is always rented. Everyone is.

An acquisition is different because you’re buying the system, not just the person who runs it. A client book built over fifteen years. An operational playbook that runs fifty auctions without anyone reinventing the process. A regional brand that banks and attorneys call first. A technology stack that compounds with every auction. These aren’t skills someone carries in their head. They’re institutional assets, and institutional assets survive the people who built them.

Most auction companies are small. One to five people, often a sole proprietor or a family operation. The top 20% doing 80% of the volume looks different, ranging from regional firms up to publicly traded companies with thousands of employees. But the single-point-of-failure problem runs through all of them, just at different scales. In a five-person shop, the owner IS the system. If they stop, everything stops. In a larger operation, maybe one person carries the key banking relationships, or one person understands the technology, or one person holds the operational playbook together. The industry doesn’t talk about this in polite company. If your key person leaves, do the relationships stay? Does the playbook survive? Do you own a platform or a login to someone else’s?

The companies pulling ahead right now aren’t doing it by adding people one at a time. They’re building systems that work whether or not any single person shows up on Monday. Proprietary bidder databases. Operational workflows that scale. Client relationships embedded in the company, not in one person’s phone. The fastest way to build a system like that is sometimes to buy one that already works.

That’s the real question when an acquisition lands on your desk. Are you buying a system that outlasts the people, or are you buying expensive employees with a term sheet instead of an offer letter? The answer changes everything about how you negotiate, how you integrate, and what you expect to still have in five years.

Here’s what I keep sitting with.

The person selling capability usually values it clearly. They know what they built and what it’s worth. The buyer often undervalues it because they’re thinking in headcount terms. Three people at X salary for Y years. That math misses the point. The salary is the cost. The system is the asset. And assets compound in ways that headcount never does.

But there’s a tension. The person who built the client book, who knows which attorney sends the best leads and which bank workout officer actually returns calls, that person is still rented. Always. The Rolodex stays when they leave. The judgment behind it doesn’t. Same with operations. The playbook stays. The instinct for when to deviate from the playbook doesn’t. Same with technology. The code stays. The reasoning behind the code doesn’t.

So when does the system become more valuable than the person who built it? That’s the test. If your key person left tomorrow and the system stayed, how much capability did you actually lose? If the answer is “most of it,” the acquisition bought headcount wrapped in an asset purchase. If the answer is “some, but we can keep building,” the acquisition bought real capability.

I think this is where the auction industry is headed. The companies that figure out whether they’re acquiring capability or hiring headcount will make better decisions about growth, about partnerships, about which relationships to cultivate and which systems to own. The ones that keep thinking in headcount terms will keep renting what they should be building.

The assets that compound in a business are the ones you own. That’s true for bidder databases. It’s true for brand. It’s true for client relationships. And it turns out it’s true for the team that builds everything else.

I’ve been keeping notes in this vault for two years. Hundreds of calls, meetings, emails, references, thought cards. All wiki-linked, all tagged, all searchable. I thought the value was in the search. Find the note, read the context, prepare for the call.

Then we added community detection. A clustering algorithm that groups nodes by edge density. Nothing fancy. Runs in 50 milliseconds. And suddenly I could see that my vault has 32 natural communities. Association operations is one cluster. State chapters are another. Client work is a third. Thought cards live in their own world, almost entirely disconnected from the operational clusters where the ideas actually apply.

That’s information I didn’t have yesterday. Not because the data was missing, but because the structure was invisible.

What Changes When Structure Is Visible

The first thing we computed was “god nodes,” the most structurally central entities in the graph. The trade association I chair scored 1.0. My own company scored less than half that. That ratio says something I wouldn’t have articulated on my own: the association is more than twice as central to my vault as the company I own. Whether that’s a problem or a feature depends on the season, but I wouldn’t have seen it without the number.

The second thing was multipliers. People whose connections span three or more communities. One contact bridges a dozen. Another bridges eight. I knew both were important. I didn’t know the quieter one was structurally more connected than the association’s CEO, who’s the person I talk to most. One is a connector. The other is a hub. Those are different roles, and they need different kinds of investment.

The third thing was emerging contacts. People I’m talking to a lot right now who aren’t structurally integrated yet. One consulting client has ten notes in 30 days and a community reach of two. I’ve been deep in strategy work with him, generating thought cards and essays from those conversations, but almost none of it connects back to the rest of the vault. The ideas are flowing. The structural integration isn’t.

That’s a signal I would have missed for months.

The Principle

We default to managing relationships by feel. Who called recently, who’s top of mind, who asked for something. Feel is fast but it’s biased toward whatever happened last and whoever’s loudest in your head. The person who emailed you yesterday feels more important than the person who bridges a dozen communities but hasn’t emailed in a week.

Structure doesn’t replace feel. It corrects for the blind spots. The graph doesn’t tell you who to care about. It tells you where your time is actually building something, and where it isn’t.

The auction industry runs on relationships. Everyone knows that. But knowing that your relationship investment is concentrated in one cluster while another cluster has zero multipliers, that’s operational. That changes what you do on Monday morning.

What’s Open

• The ROI metric combines structural importance with velocity. The top contact scores 0.96. But ROI assumes the current structure is the one you want. What if the right strategic move is to deliberately build bridges into an underconnected community? The metric rewards maintaining the existing network shape. It doesn’t reward reshaping it.
• The ideas cluster and the operations cluster are separate worlds. My thinking about the industry and my operational work in the industry don’t cross-pollinate in the graph. Is that a documentation gap, or is the separation real? If it’s real, what does it mean that the ideas aren’t grounded in the operations?
• The emerging contacts signal catches people with high recent velocity but low reach. It doesn’t catch people with low velocity who should have high reach. The person you’re not talking to but should be. That’s the harder question, and the graph can’t answer it alone.

The agentic shift is not an upgrade to existing infrastructure. It replaces the assumptions that infrastructure was built on.

“Each level isn’t defined by what the AI starts doing. It’s defined by what the human stops doing.” – Five Levels of Agentic Commerce

“The cost of modern convenience is an involuntary autobiography that someone else owns.” – Digital Exhaust

“The agentic web is the marketplace trap with different actors.” – Two Trust Systems

“It determines whether or not or if you won and then makes something up to fit that scenario.” – The Movie That Matches The Outcome

“Attention doesn’t scale. Systems do.” – Automated Marketing Stack

Seven thought cards and three essays written across eight weeks converge on the same structural observation: the agentic layer doesn’t make existing processes faster. It changes who does the work, who holds the data, and what “trust” means when there’s no human in the loop.

The shift moves in one direction. At each level, humans stop doing something they used to do. Stop searching. Stop choosing. Stop prompting. Eventually stop paying attention at all. The interesting question isn’t what agents can do. It’s what happens to the systems, markets, and institutions built on the assumption that a human was watching.

Marketing assumes someone sees the ad. Brand assumes someone recognizes the name. Auction competition assumes someone feels the rush of winning. Pricing assumes someone reads the menu. Memory assumes someone forgets what’s no longer relevant. Every one of those assumptions breaks when the human leaves the loop. The agentic shift is the name for what breaks, and what replaces it.

The marketing stack essay makes it concrete. When every campaign used to start with a blank page and a person’s judgment, quality was a function of who showed up. Once the workflow was codified, quality became a property of the system. The person still makes the calls, but the system carries the volume, enforces the standards, and compounds every correction into every future run. That’s the shift at the operational level: not removing the human, but changing what the human is for.

The Ownership And Control concept asks: who owns what? This concept asks: how does the shift itself change the rules of the game?

Evolution

Source Voices

• Stripe (John & Patrick Collison) – The five-level framework. Each level defined by what the human stops doing. The bifurcation data: winners pull away across every industry, and agentic commerce accelerates the sorting (from Five Levels of Agentic Commerce)
• a friend – Coined “digital exhaust.” The data trail you produce but don’t control, the involuntary autobiography powering the agentic layer’s inputs (from Digital Exhaust)
• Jason Barnard – The algorithmic trust framework: recruited, grounded, displayed, won. For an autonomous agent, confidence can’t be probabilistic; it has to be absolute (from Article Aao Assistive Agent Optimization)
• Dave Plummer – The slot machine architecture as a model for AI-generated experiences. The outcome is already decided. The experience is theater built after the fact (from The Movie That Matches The Outcome)
• Acquired podcast – YouTube vs. Google Video. Same product, different friction. The low-friction version attracted creators, creators attracted audience, audience attracted more creators. Users did the scaling (from Users Do The Scaling)
• Matthew Prince – Cloudflare CEO’s numbers: 6,000 pages crawled per visitor sent back. The crawl-to-visit ratio as proof the browser era is ending (from True Market Value: Why Structured Data Is the Last Moat in Auctions)
• a colleague – The real-time demonstration of the quality gap. Received a carefully drafted email and returned AI-generated output without critical engagement. The person on the wrong side of the shift can’t see the gap. (from Sparring Vs Ghostwriting)

Applications

• True Market Value: Why Structured Data Is the Last Moat in Auctions – The essay where the shift becomes operational. Agents solve the room-size problem for auctions: from 50 emotional humans to 5,000 rational agents. The moat moves from buyer reach to data quality. The flywheel flips from buyer acquisition to seller relationships plus structured data.
• Article Aao Assistive Agent Optimization – Barnard’s AAO framework as the practical playbook. SEO becomes AEO becomes AIEO becomes AAO. At the final stage, the agent acts without a human in the loop. The brand that gets chosen is the one the algorithm understands, trusts, and can act on.
• Two Trust Systems – Grafe running both systems simultaneously. Algorithmic trust (structured data, owned tech stack) on the buyer side. Human trust (relationships, reputation, showing up) on the seller side. The company that runs both wins. The one that optimizes only one gets blindsided.
• The Movie That Matches The Outcome – Recommended settings for new bidders on grafeauction.com. The “decide for the user” principle applied to platform defaults. If the outcome you want (an engaged, informed bidder) is already determined, the settings are just the movie that matches it.
• Structure Reveals Strategy – The agentic shift applied to personal infrastructure. A clustering algorithm running in 50 milliseconds revealed 32 communities, structural multipliers, and emerging contacts that felt invisible. The same dynamic as Hawk-Eye: when an algorithm shows you something you’ve been managing by feel, the accuracy gap creates a legitimacy shift in your own decision-making.
• Sparring Vs Ghostwriting – The quality gap as competitive divergence. One person’s AI has 500 files of context, 30 custom skills, and compounding voice corrections. Another’s has a paste buffer. The outputs look superficially similar but are fundamentally different. The baseball crowd cheered because the accuracy gap got too wide; the email quality gap is approaching the same inflection point.
• Automated Marketing Stack – The shift goes from theory to production. 290 campaigns/year, each previously starting from a blank page. Now a single skill pulls the lot catalog, analyzes category mix, generates copy within platform character limits, selects photos against a rubric, and targets geographically. Active human involvement dropped to ~60 seconds per campaign. The system carries the volume, compounds every correction, and holds quality steady on campaign 290 the way it did on campaign 1. This is the concept’s clearest operational proof: the human didn’t leave the loop, but the loop no longer depends on the human’s consistency.
• Acquiring Capability Vs Hiring Headcount – The agentic shift applied to business structure. Acquiring Happy Dog buys the codebase, the design patterns, the deployment pipeline. The architecture compounds whether or not any individual stays. Hiring an experienced operator buys relationships that convert from rented to owned, but with no transferable architecture. The distinction between “capability that survives the person” and “headcount that walks out the door” is the ownership question restated through the lens of what agentic infrastructure actually is.
• Skill Audit Decisions – The memory half-life tension applied to the agentic infrastructure itself. 56 skills pruned to 34. Eight killed, six merged into orchestrators, nine consolidated to shared infrastructure. The system needed forgetting to stay useful, the same dynamic the concept tracks at the data layer.
• The Yin And Yang Of Governance – Integrated new capability into the structural shift.
• The Scheduled Collision – Integrated new capability into the structural shift.
• Systemic Rivals Against Default Drift – Integrated new capability into the structural shift.
• Cognitive Defense Against Default Drift – Integrated new capability into the structural shift.
• Reciprocity In Ai Sparring – Integrated new capability into the structural shift.
• The Barbell – Integrated new capability into the structural shift.
• Ownership At Every Altitude – Integrated new capability into the structural shift.

Tensions

• Theater vs. efficiency. Auctions thrive on competition, emotion, the feeling of winning. Agents remove the emotional layer entirely. What’s left is pure market efficiency, great for price discovery, terrible for the theater that makes auctions worth attending. The math may still work (volume compensates for lost premium), but the experience changes. Sources: Five Levels of Agentic Commerce, True Market Value: Why Structured Data Is the Last Moat in Auctions.
• Memory persistence vs. decay. AI systems default to remembering everything forever. But unbounded memory degrades performance the same way unbounded organizational complexity does. The brain’s forgetting curve isn’t a bug; it’s what keeps the signal-to-noise ratio manageable. The agentic layer needs decay rates, confidence gradients, and consolidation cycles, not just more storage. Source: Memory Half Life.
• Human trust vs. algorithmic trust. They run on different mechanics, different timelines, and different failure modes. Human trust breaks through betrayal. Algorithmic trust breaks through inaccuracy. An algorithmic trust product (like HiveGinx’s bidder integrity score) can only reach the market through a human trust process that moves at the speed of scar tissue. Source: Two Trust Systems.
• Scale vs. oversight. The platforms that removed friction and let users do the scaling changed who owns the flywheel. But the industries that require expertise, regulation, and trust can’t simply remove friction without losing the value the friction created. The question is which friction is structural and which is just workflow debt. Sources: Users Do The Scaling, Two Trust Systems.
• Awareness vs. paranoia. The digital exhaust observation is powerful, but the teaching problem is real. The line between “be intentional about your footprint” and “they’re watching you” is thin. One lands as wisdom, the other as conspiracy. Source: Digital Exhaust.
• Quality gap visibility. The person using AI as a ghostwriter can’t tell the difference between their output and the sparring version. If the quality gap is invisible to the person producing it, at what point does the audience start to see? The baseball crowds noticed because the gap was measured in inches. In email and content, the measurement is less precise but the divergence is widening. Sources: Sparring Vs Ghostwriting, The Crowd Cheered For The Machine.
• System dependence vs. creative autonomy. The marketing stack essay says “you don’t lose the creativity, you just stop depending on it.” But dependence was also what made each campaign feel attended to. When the system carries 290 campaigns and the human’s role drops to 60 seconds of judgment per run, the output is more consistent but the person is further from the work. The tension: consistent quality at scale vs. the institutional knowledge that only comes from doing the work by hand. The system compounds corrections, but does it compound taste? Sources: Automated Marketing Stack, Users Do The Scaling.

What’s Open

1. At what level does brand stop mattering? If agents are choosing, do they care about brand at all, or just ratings, price, and specs? (from Five Levels of Agentic Commerce)
2. What does “trust” look like at Level 4? Delegation requires a fundamentally different order of trust than recommendation. We barely trust recommendation algorithms now. (from Five Levels of Agentic Commerce)
3. Is there a Level 6 where agents sell on your behalf too? Not just buying but liquidating, listing, pricing, negotiating, the full cycle? (from Five Levels of Agentic Commerce)
4. At what point does aggregated digital exhaust cross from “data” to “identity”? Where’s the threshold where scattered exhaust becomes a portrait? (from Digital Exhaust)
5. What’s the right decay model for AI memory? Feedback memories need infinite half-life. Project memories rot in days. How do you build the confidence gradient so aging memories get verified before use instead of trusted blindly? (from Memory Half Life)
6. If the page is the movie assembled after the outcome is decided, what does A/B testing even mean? You’re not testing two versions of the same page. You’re testing two rendering strategies for the same predetermined result. (from The Movie That Matches The Outcome)
7. Which parts of the auction process genuinely require human expertise, and which are friction that exists because we never redesigned the workflow? (from Users Do The Scaling)
8. Can both trust systems be run at the same scale? Algorithmic trust scales automatically. Human trust doesn’t. At 290 auctions/year across 48 states, is the human side already at its limit? (synthesis)
9. If quality becomes a property of the system, what happens to the role that used to produce it? The marketing hire who would have started from a blank page 290 times a year now doesn’t exist. The skill replaced the position before the position was created. How many other roles are being preempted rather than displaced? (from Automated Marketing Stack)
10. When you acquire a company for its agentic infrastructure (the codebase, the architecture), at what point does the architecture become more valuable than the architect? If the person who made all the design decisions leaves, how much capability did you actually buy? (from Acquiring Capability Vs Hiring Headcount)
11. The skill audit pruned 56 skills to 34. The agentic layer needs forgetting the same way human memory does. But who decides what to forget? The pruning was a judgment call, not an algorithm. Is there a decay model for agentic infrastructure, or does it always require a human audit? (synthesis from Skill Audit Decisions, Memory Half Life)

In January I stood in front of 100 auctioneers at a state convention and told them “AI assists. You decide.” Three hours of live demos, chaining tools together, a 7-day challenge. I showed them 200 hours of reclaimed marketing time. The room bought in. The closing line about the train leaving the station landed.

In February I started one-on-one coaching with a colleague, walking her through the same tools at a slower pace. In March I did it again for 20 community leaders on a Zoom call. And now I’m building a 3.5-hour continuing education course that’s basically the same material refined a fourth time.

Two days before that course research started, I found a bug where stale cached data was silently injecting garbage into every outbound email for a full day. My own AI memory system, the one I’d built to make my tools smarter, was serving bad data because nobody built in decay.

And two weeks before that, I realized I’d given a consulting client the exact triage framework I haven’t applied to my own company. Document processes before hiring. Sort decisions into only-you, delegatable, and automatic. I diagnosed his problem in ten minutes and haven’t solved the same one at my company in fourteen years.

What I Think Is Actually Happening

There’s a version of expertise where you only teach what you’ve fully mastered. That version is clean and safe and mostly useless, because by the time you’ve mastered something in technology, the thing you mastered has already moved.

The alternative is teaching from the middle of it. I’m teaching auctioneers to use AI tools while simultaneously discovering that the AI systems I’ve built have fundamental design flaws. I’m telling rooms full of people to trust but verify while my own cache was serving unverified garbage for a full day without anyone catching it.

That’s not disqualifying. It might be the whole point.

The convention seminar worked because I could show real numbers from real campaigns. The leadership session worked because I told two stories where I learned something by getting it wrong first. In both cases the AI framing created tunnel vision, and I only caught it by flipping the prompt. The one-on-one coaching works because my colleague watches me problem-solve in real time, not deliver a finished product.

Every good teaching moment I can point to came from something I was still figuring out. None of them came from polished expertise.

The Principle

Mastery isn’t a prerequisite for teaching. Proximity is. The person one step ahead of you in the fog is more useful than the person who mapped the whole terrain from a helicopter. They know where the ground is soft because they just stepped in it.

The memory bug made me a better teacher of AI skepticism than any slide about hallucinations could. Realizing I hadn’t taken my own advice makes the triage framework more honest, not less. I can say “I haven’t done this either, and here’s what it’s costing me.”

The gap between what you teach and what you’ve mastered isn’t something to hide. It’s the most credible thing you bring to the room.

What’s Open

The CE course research covers the tools and trends well enough. But does it include enough of the “here’s where I got it wrong” stories? The convention and leadership sessions landed hardest on the failures. If the course leans too much on the polished version, it’ll lose the thing that actually teaches.

There’s also a tension between the “AI assists, you decide” framing (which puts the human in control) and the memory bug insight (which shows the AI quietly making decisions without anyone noticing). The first is aspirational. The second is operational. The course probably needs both.

And the consulting card asks whether I’ve documented the “why” behind my systems. The teaching version of that question: am I teaching people to use the tools, or am I teaching them how to think about when the tools fail? Those are different courses.

This came up while fixing a bug where stale cached data was silently injecting garbage into every outbound email for a full day. The cache never questioned itself. It just kept serving bad data.

Human memory has a built-in half-life for each memory. The things you use stay strong. The things you don’t fade. That’s not a bug, it’s what keeps the signal-to-noise ratio manageable. You naturally surface what’s relevant because relevance reinforces the memory.

Persistent AI memory systems don’t have that. Every memory sits at equal weight forever until someone manually prunes it. A project note from three months ago carries the same authority as one from this morning. That’s a problem, because unbounded compounding memory can actually degrade performance over time. Stale context doesn’t just take up space. It introduces contradictions, crowds out high-signal recent knowledge, and actively misleads.

The Ebbinghaus forgetting curve backs this up. Decay reduces cognitive noise and forces consolidation of what’s actually important. Spaced repetition systems like Anki exploit this. They use the decay curve rather than fight it.

The Design Question

What if memories had decay rates tied to both time and usage?

Not all memories decay the same way. Corrections (“don’t do X”) shouldn’t decay at all. They’re load-bearing regardless of how often they fire. But project context (“merge freeze for the mobile release”) has a natural shelf life measured in days.

A rough model:

• Feedback memories: Near-infinite half-life. A correction that hasn’t been needed in months is a success, not a signal to forget.
• User memories: Long half-life. Identity and preferences move slowly.
• Reference memories: Medium half-life. Systems change, links break, tools get replaced.
• Project memories: Short half-life. These are the ones that rot fastest and mislead most when stale.

The Usage Dimension

Time alone isn’t enough. A memory that keeps getting accessed is proving its value. One that hasn’t been touched in 90 days is either stale or a safety net that hasn’t been needed yet. The hard part is distinguishing those two.

The real advantage of human memory decay isn’t forgetting. It’s forcing re-derivation. When you half-remember something and have to re-check, you’re validating it against current reality. Instead of binary present/absent, you get a confidence gradient: fresh memories get trusted, aging ones get verified before use.

Consolidation as Sleep

The brain doesn’t just decay passively. Sleep-based memory consolidation actively reviews and strengthens what matters. I’ve built a consolidation process into my own AI memory system that does a version of this: it reviews accumulated learnings, clusters by theme, and graduates confirmed patterns into durable memory. That’s the consolidation step. The missing piece is the decay side. The automatic downweighting of what consolidation doesn’t touch.

The Bigger Pattern

Technology is the default. Human relationships must be actively defended. The same logic applies here: persistence is the default for digital memory. Relevance must be actively defended. Without decay, memory systems drift toward noise the same way organizations drift toward automation. Not because anyone chose it, but because no one built in the friction that keeps things honest.

He’d called to pick my brain. Decades in the auction industry, multiple divisions, good revenue. But every decision in the company routed through him. Sales, operations, finance, client relationships. He’d spent his career being the best person on the microphone, and the rest of the business just happened around that skill. There were no documented processes. No decision-making frameworks anyone else could follow. If his potential successor took over tomorrow, they’d fail. Not because they’re incapable, but because the architecture underneath was never built for anyone else to run.

He accepted the assessment. Didn’t push back. We built a triage framework together: sort every decision into three buckets. Things only he can decide. Things someone else could decide with guardrails. Things that should happen automatically.

A few days later it hit me. I’ve been one of the few people who can maintain and extend our platform. That’s the same vulnerability I’d just diagnosed in someone else. Same prescription, same disease. I just hadn’t noticed because my version feels like competence.

Same Trap, Different Company

A few days after that call, I was on the phone with a friend in the auction space. He’s been with a company where the founder built everything from scratch.

But the founder is getting close to stepping back. The team resists change. Every time my friend raises the topic of what comes next, the conversation gets deflected. Not because the founder is selfish. He probably looks at what “modernizing” means and thinks, where would I even begin?

So my friend is stuck. He’s got the skills to build something new, but he’s working inside a company that can’t imagine what comes after the person who built it.

Two different businesses. Two different people. Same trap.

The Grip and the Surrender

I’ve been thinking about this pattern, and it shows up in one of two ways.

The first is the grip. This is the founder who holds everything because letting go feels like losing control. The person I was helping is this version. So is the company founder my friend works for. These aren’t lazy people. They’re usually the hardest workers in the building. But they never separated the valuable human judgment from the mundane operational routing that just happens to live in the same brain. Over time, the two become inseparable. Their competence becomes the company’s constraint.

The second is the surrender. This is the person who hands everything to the tool and stops thinking. Social selling platforms are full of this version right now. Countdown timers, competitive bidding, items going to the highest bidder. The platforms handle the mechanics. The sellers skip the legal framework, the ethical structure, the professional judgment that separates a real auction from a garage sale with a timer on it. They didn’t grip too tight. They let go of things they never understood they were holding.

Both versions are responses to the same thing: the pace of technological change. Some people freeze and grip tighter. Others overcorrect and hand everything over. The first group ends up irreplaceable in the wrong way. The second group ends up replaceable without knowing it.

What’s Actually Left

Here’s what’s actually happening underneath both failure modes.

Technology is pulling apart the mechanical parts of expertise and making them portable. You don’t need a licensed auctioneer to run a timed online sale. You don’t need a marketing department to reach millions of potential buyers. You don’t need a 40-year reputation to get people to trust a transaction. The platform and the algorithm handle the rest. Trust, reach, mechanics. All commoditized.

So the question becomes: what’s left?

I think the answer is judgment. The ability to read a situation, make a call that a system can’t make, and take responsibility for the outcome. That’s the part of being an auctioneer (or any professional) that doesn’t come apart. But here’s the problem: judgment is invisible until something goes wrong. Nobody notices the experienced auctioneer’s decision to pull a lot, restructure terms, or walk away from a bad deal. They notice the sale price. And the platform can show you a sale price too.

So protecting the human element takes deliberate effort. The judgment, the critical thinking, the relationships. None of it survives on autopilot. The default is technology replacing humans, because it’s faster and cheaper and doesn’t need to be convinced.

The Intentional Version

So what does the intentional version look like?

Remember the triage I built with the person who called me. Three buckets: things only he can decide, things someone else could decide with guardrails, things that should happen automatically. That framework isn’t just about delegation. It’s about knowing which parts of your expertise are actually yours and which parts just happen to live in your head because nobody ever extracted them.

The same triage applies to technology. Some decisions should stay human. Some can be handled by a tool with guardrails. Some should be automated entirely. The person who grips is the one who puts everything in bucket one. The person who surrenders puts everything in bucket three. The healthy middle requires the honesty to sort accurately, which means admitting that a lot of what feels like expertise is really just routine wearing a mask.

Abraham Lincoln understood this instinctively. He assembled his first cabinet from political rivals. People who’d run against him, disagreed with him publicly, thought they should have his job. He did it because he knew his own blind spots would be fatal if nobody challenged them. He kept the decisions that required presidential judgment. He surrounded himself with people who’d push back on everything else.

I think that’s the design principle for working with technology too. When I started building AI into my workflow, I could have built a system that just gives me answers. Draft the email, write the marketing copy, generate the catalog descriptions, done. And some of my tools do exactly that, the bucket-three stuff. But the ones I rely on most are the ones that push back. Tools that assemble different perspectives and argue with each other before presenting a recommendation. Tools that won’t let me be passive.

I built it so I could copy and paste. But it’s hard to do that based on how it wants to engage with me.

That friction is the point. I don’t want a tool that replaces my thinking. I want one that makes me think harder. The distinction sounds small. In practice, it’s the difference between the grip and the surrender.

The Scariest Gap

The question I keep circling is whether this is teachable.

I think it’s seeable. You can watch someone use technology well and recognize what they’re doing differently. I think it’s teachable. You can build the frameworks, show the examples, walk through the process. But doing it is harder than either of those things, because it’s scary. Letting go of the parts of your expertise that should be systematized means admitting they weren’t as special as you thought. And building friction into your tools means slowing down when everyone else is speeding up, accepting that you might be wrong about something you’ve believed for decades.

The person who called me could see all of this. He accepted the diagnosis. Whether he’ll actually triage those decisions, document those processes, and let someone else make the calls he’s made his whole career is a different question entirely. Accepting the prescription and filling it are two very different things.

But the alternative is worse. The person who grips everything eventually runs out of hands. The person who surrenders everything eventually runs out of relevance. And technology doesn’t wait for either of them to figure it out.

The healthy middle is a choice you have to keep making. Not once, not as a strategy document, but every time you’re tempted to either hold tighter or let go entirely. It requires the same quality Lincoln looked for in his cabinet: people (and now tools) that are good enough to challenge you, and the security to let them.

The person whose business couldn’t survive him? He’s working on it. The friend stuck inside someone else’s refusal to change? He’s exploring what comes next.

And me? I’ve been one of the few who can maintain the system I helped build for my own company. I know the triage. I know the three buckets. I just haven’t sorted everything into them yet.

I think that’s the real trap. It’s not that people don’t see it. It’s that seeing it clearly and doing something about it are separated by the scariest gap in business: the one between knowing what’s right and actually letting go.

The Customization Confession

Plummer and Lex were arguing about whether you should be able to move the Windows 11 taskbar. Lex was frustrated. He wanted customization, options, power user freedom. Plummer pushed back with an engineering argument first:

“The freedom to put the Start menu on the left or the top or the right really increases the complexity of the code that renders the Start menu and lays out the tabs and does all the things, and now it’s a much larger surface for bugs and it’s a much larger piece of code to maintain.” — Dave Plummer (47:13)

Then he dropped the Steve Jobs framing:

“How do you respond to the Steve Jobs aspect of giving you customization implies that we couldn’t figure out the right answer for you?” — Dave Plummer (52:37)

That line stopped me. I’ve been saying a version of this for a while without connecting it to Jobs. The mom test I use for our auction platform is the same instinct. If I need a settings panel, I haven’t finished designing. If my mom can’t figure out the workflow without calling someone, I didn’t ship a product. I shipped a confession that I couldn’t decide.

Plummer’s practical framing helps too. He wasn’t being philosophical about it. He was saying: somebody looked at a deadline and cut the four-position taskbar because they could ship four months sooner with one position. That’s not lazy. That’s a decision. And the Jobs line gives the decision a spine.

The Niche Feature Discipline

Plummer asked the question plainly:

“The question becomes at what point is it still worth it to satisfy the niche needs of a small set of users?” — Dave Plummer (48:28)

His answer, implied: usually, it isn’t.

I push back on this regularly with our dev team. It’s not just that solving for the 1% isn’t worth the effort. It’s that solving for the 1% actively creates friction for the 99%. Every edge case you accommodate adds a toggle, a conditional, a branch in the UI that the majority of users now have to navigate around. You’re not just wasting time on a small audience. You’re degrading the experience for everyone else.

But Lex made the counterargument, and it’s worth sitting with:

“If there’s a guy or a girl in there who takes ownership of it, who has a passionate… For them, it’s a thing that they take pride in over a period of time, they can by themselves in a short amount of time create something truly wonderful.” — Lex Fridman (51:06)

He called it craftsmanship. He compared it to filmmakers who insert details only a small number of viewers will notice, and argued the reputation that builds over time pays off in ways that are hard to measure. Including hiring.

I think both are true. The line between them is whether the niche feature adds complexity the majority has to deal with, or whether it’s invisible to them. A taskbar that works on all four edges adds code and bugs for everyone. An Easter egg in Pinball adds delight for the few and costs the many nothing. The discipline is knowing which one you’re building.

The hard part isn’t saying no upfront. It’s dropping something you’ve already started building once you realize the audience is too small. By then you’ve invested time, you’ve started solving the problem, you’ve got momentum. Walking away from momentum takes more discipline than never starting.

Simplicity isn’t a starting point. It’s what’s left after you’ve had the courage to cut.

Recommended Settings for the User, Not for You

Plummer and Lex also got into Microsoft’s “recommended settings” trick. You upgrade Windows, click through the wizard, and “use recommended settings” sounds like the right call. But it resets your default browser to Edge and opts you into things you didn’t ask for.

“You don’t want your operating system to be an adversary. And sometimes Windows can feel adversarial. Like, it doesn’t have your best interests at heart.” — Dave Plummer (1:14:10)

That conversation got me thinking about our own platform. Right now, a new bidder gets all notification settings turned off by default. That means a first-time bidder registers, wins nothing, and never knows she was outbid because she didn’t find the settings page. Same mom test I apply to every screen on our platform. That’s not respecting her preferences. That’s punting on the decision.

What if we instituted recommended settings for a bidder’s profile? Text alerts when outbid, email notifications for auctions closing soon, saved payment info. The things a reasonable bidder would want turned on. Not marketing emails, not data sharing. Just the stuff where the bidder would thank us for defaulting it on rather than resent us for it.

The Microsoft version fails because it serves Microsoft. The test for us is simple: does this default serve the bidder or serve us? If the answer is the bidder, turn it on. If it’s us, leave it off and let them opt in.

The outcome we want, an engaged, informed bidder, is already determined. The settings are just the configuration that matches it.

The Slot Machine Movie

This one came out of nowhere. Plummer plays slot machines. He knows they’re a losing bet, but he likes the dopamine, the bright lights, the high contrast. Then he explained how they actually work.

“Internally, there’s basically a black box mechanism that does nothing more than generate the next random number and what the outcome is in terms of probability and payout. And then the game says, ‘I’ve got to make up a movie to go along with that.’ … It’s not spinning the reels, seeing where they land, and looking that up to see what you won. It’s completely the other direction. It determines whether or not or if you won and then makes something up to fit that scenario.” — Dave Plummer (1:17:40)

The spinning reels, the near-misses, the bonus animations. None of it determines anything. The outcome is already decided. The experience is theater built after the fact.

I’ve watched people play these things a hundred times. But I never thought about the architecture of it until Plummer said it that way.

And it immediately made me think about where websites are heading. The outcome is already determined: the business knows what it wants to serve this customer, or this AI agent, based on who they are and what they need. The page itself is just the movie assembled in real time to match. No static templates. No one-size-fits-all layout. The UI is constructed for that specific visitor in that specific moment. The slot machine is basically the model for AI-generated websites rendered on demand.

My cousin texted me this morning with a line I’ve heard before but landed differently today: “A ship in a harbor is safe, but that’s not what ships are for.”

It’s attributed to John A. Shedd, published in 1928. One of those quotes that gets printed on coffee mugs and LinkedIn posts until you stop hearing it. But context matters. He didn’t send it as decoration. He sent it as a nudge.

What I Think Is Actually Happening

There’s a version of safety that looks responsible but is actually just avoidance. Staying in the harbor feels like stewardship, like you’re protecting the ship. But the ship was built to move cargo and reach ports you can’t see from the dock. A ship that never leaves isn’t well-maintained. It’s wasted.

I keep coming back to the tension between building and preserving. Grafe runs 290 auctions a year across 48 states. We’re not sitting still. But I think there are decisions I’ve deferred because the current trajectory feels safe enough. Revenue diversification conversations I haven’t pushed. Technology bets I’ve studied but not made. The harbor isn’t inactivity. It’s the comfortable pace that doesn’t require me to risk anything new.

The quote also connects to something I wrote in The Measure: the comfortable fiction that there’s always more time. Staying in the harbor assumes the harbor will always be there. It assumes calm water. It assumes the cargo doesn’t spoil while you wait.

The Principle

Safety and purpose are often in tension. The things worth doing require leaving the conditions where nothing can go wrong. That doesn’t mean recklessness. I think it means accepting that the ship proves its worth in open water, not at the dock.

The question isn’t whether to leave the harbor. It’s whether you’re honest about why you haven’t left yet.

What’s Open

• Which specific decisions am I deferring because the current trajectory feels safe enough? Worth making a list.
• How does this connect to Understanding vs. Doing? Collecting frameworks is its own kind of harbor. Understanding without execution is a ship that never sails.
• My cousin and his brother both come from a family that built a business from scratch. That side has never been harbor people. Worth reflecting on what that means for the next generation.

YouTube won because anyone could upload a video in minutes. Google Video required you to fill out forms, wait for approval, deal with format restrictions. Same basic product, totally different levels of friction. The low-friction version attracted the creators, and the creators attracted the audience, and the audience attracted more creators. Users did the scaling. Google just had to buy it afterward.

The principle is pretty simple: content companies provide the platform, and users do the scaling. But I think it only works when friction is low enough that users actually show up and contribute.

What I Think Is Actually Happening

This is exactly what TikTok, Whatnot, and eBay Live figured out for auctions. They didn’t build better auction software. They built lower friction. A seller with a phone and some inventory can be live-auctioning in minutes. No contract, no consignment agreement, no 10-day campaign timeline. The platform provides the infrastructure, and the sellers do the scaling.

Meanwhile, our model at Grafe requires a sales call, a contract, an inventory process, photography, cataloging, marketing, and a 10-14 day campaign window. That’s not friction for friction’s sake. I think every one of those steps adds real value for commercial and industrial assets. But it means the user can’t do the scaling for us. We do the scaling, one auction at a time, with our own labor.

That’s fine when you’re selling $200K CNC machines or full restaurant liquidations. The complexity justifies the friction. But I keep wondering: are there asset categories or seller segments where we could dramatically lower friction and let the sellers scale us?

The Principle

The Google Video vs. YouTube comparison isn’t really about video. It’s about who does the work of getting bigger. High-friction platforms grow linearly because every new unit of output requires company labor. Low-friction platforms grow exponentially because every new user becomes a potential contributor.

The question for any business is: which parts of your process genuinely require your expertise, and which parts are friction that exists because you never redesigned the workflow?

What’s Open

• What would a “YouTube mode” look like for Grafe? Self-serve consignment for assets under a certain value threshold, maybe. Upload your own photos, set your own reserve, we handle payment processing and dispute resolution. We’d lose control but gain volume.
• Is there a middle path? Some hybrid where we provide the auction infrastructure and trust layer, but sellers handle their own cataloging and marketing for smaller lots? Kind of what Whatnot does, but with a licensed, regulated auction company behind it.
• The real moat might not be the full-service model. It might be the trust, compliance, and settlement infrastructure that social platforms haven’t built. Could we license that layer to self-serve sellers?
• How does this connect to the auctioneer identity question? If we enable self-serve sellers on our platform, are they auctioneers? Are we?

Over the last few months, a pattern kept surfacing. Every card I wrote about how principles form, how conviction develops, how people actually grow, connected to the next one. Not because I planned it, but because the same current kept pulling me back. Looking at them now, they trace an arc I didn’t see while I was inside it.

The Flywheel

Here’s how I think principles actually form.

You consume something. A podcast, a book, a conversation you overhear. It rattles around for a while, mostly inert, mostly filed under “interesting, should think about that.” It stays there until you’re forced to articulate it. A conversation. A question you didn’t expect. The moment you have to explain what you believe and why, the idea either solidifies or falls apart.

But articulation isn’t enough. The real formation happens when you can’t opt out. When you’re standing in front of an audience and someone asks a question your framework doesn’t answer. When a crisis at work exposes the gap between what you believe and what your paperwork says. When the valley forces you to sit with the idea under pressure.

Then someone names it. A friend, a mentor, a counselor. Someone who reflects back what the hard season actually built in you. Without the namer, the valley is just suffering you survived. With the namer, it becomes the origin story of a capacity you didn’t know you had.

Then you teach it. And teaching, I’ve come to believe, is the highest-friction form of application. Because the audience becomes your valley. Their resistance, their specific problems, their questions that don’t fit your framework. You can’t fake it. You can’t file it away for later. You have to make the idea work right now, in front of people who will know immediately if it doesn’t.

And then you encode it. Into a contract, a process, an essay, a system. The principle becomes infrastructure. The encoded principle gets consumed again. By other people, yes. But also by you. The act of encoding forces you to re-examine what you believe, and the re-examination starts the cycle over.

It’s a flywheel. Each revolution doesn’t repeat the last one at higher speed. It repeats it at a deeper level. The first time around, you discover the mechanism. The second time, you discover that the mechanism requires friction. The third time, you discover you were inside the friction all along.

The Energy Source

A flywheel stores momentum, but it doesn’t generate it. Something has to push.

For me, the push has always been other people. I have this drive, and I think a lot of people share it, to make sure the people around me are happy, encouraged, and better for having been near me. That’s the energy. Seeing the thing land. Watching someone walk out of a seminar with a tool they’ll actually use. Hearing from a friend that a conversation changed how they think about their business.

The problem is that the same fuel source is also a leak.

I need to see my efforts land. I need visible evidence that the needle moved. When I’m in front of an audience, I know I’m doing a good job if I can get an emotional reaction. Anger, laughter, agreement, anything. If they’re mad at me, I’m in their head. If they’re completely indifferent, I haven’t made them think. The only failure state is silence.

And silence is the default state of the deepest impact. The person who goes home, sits with an idea for a week, and quietly changes how they run their business. I never see that. The standing ovation fades in a day. The email thank-you stops warming me after a week. The quiet transformation that happens in private never registers on my scoreboard.

So I’m wired to notice the shallow signal and miss the deep one. The scoreboard I check is the wrong scoreboard. The real impact is happening in a room I’m not in, on a timeline I can’t see.

The Undefined Standard

I grew up trying to match a standard that was never detailed to me.

The message was clear enough: perform, and figure out the target on the way up. No briefing. No debrief. Just execute and hope the absence of correction means you got it right.

Silence was ambiguous. It could mean approval. It could mean disappointment. I never knew. So I learned to over-deliver as insurance. If I couldn’t read the scoreboard, the safest move was to make sure the score was so high it didn’t matter.

That’s where the pattern started. Not the drive to teach or mentor or give. The drive to perform at a level where the undefined standard couldn’t possibly catch me. The overgiving isn’t really generosity. It’s insurance against not knowing the score. If I give enough, teach enough, prepare enough, then even if I can’t see the scoreboard, I’ve probably cleared the bar.

And the moments when I do get visible feedback, the audience reaction, the thank-you email, the post that gets traction, those aren’t just dopamine hits. They’re the scoreboard finally flickering on for a second. Of course I’m tuned to catch them. For most of my life, they were the only signal I had that the standard was met.

I found the root of this in my late twenties through counseling. A process of tracing patterns back to their origin and confronting what you find. I found it. I named it. I understood it.

And the pattern kept running.

The Productive Tree

Understanding the root didn’t stop the tree from growing. By the time I found it, the pattern had been load-bearing for decades.

The overdelivery built our company’s reputation. The relentless preparation makes my seminars land. The drive to see impact is what makes me a decent mentor. The bitter root grew a productive tree. That’s what makes it so hard to change. I’m not trying to stop a behavior that’s failing. I’m trying to moderate a behavior that’s succeeding, at a cost that shows up somewhere else.

The cost isn’t invisible. My wife and daughter feel it. The overloaded capacity, the depletion, the version of me that comes home after the tank is empty from making sure everyone else’s scoreboard lit up. They don’t need to know about the root to feel the weight of the tree.

I wrote in one of those thought cards that we autopsy our losses but toast our wins. I applied that to how people learn from experience. But it applies here too. I feel the silence acutely and celebrate the visible reactions briefly. The withdrawals register longer than the deposits. The accounting is asymmetric, and the people closest to me absorb the difference.

The Third Path

I have a three-year-old daughter. She’s absorbing architecture right now, not words. And I’m very aware that I’m going to be the voice in her head the way my dad’s voice is in mine.

My instinct has been to do the opposite of what I received. Shower love. Say “I’m proud of you” out loud and often. Make sure she never has to guess whether she met the standard.

But opposite isn’t the same as resolved. It’s a reaction to the same axis. One parent answers the question “how does a child know they’re enough?” by not answering. The other answers by answering constantly. The content is different, but the structure is the same. Both teach a kid to look for the scoreboard. One version just happens to be always lit up.

My wife and I are working on a third path. We hold our daughter accountable. When things don’t go the way she expected, we sit with her in it instead of fixing it or ignoring it. We’re big on self-awareness, helping her see what those moments are building in her, not just that mom and dad approve.

That’s honoring accountability: “You’re wrong, and I care about you enough to tell you, and you’re still safe here.” It teaches that being wrong is survivable. Which is the thing silence never taught me.

The honest part is that I’m building in my daughter the thing I’m still building in myself. I’m teaching her to have an internal engine while mine still runs on external fuel. That’s not hypocrisy. It’s the flywheel turning again. The act of encoding a principle I believe in (self-sourced validation) is forcing me to confront what’s still unfinished. Because I can’t teach it to her and ignore it in myself forever. Kids see through that eventually.

Building Your Own Scoreboard

In a conversation recently, someone asked me what success looks like without the external signal. What would the flywheel run on if I stopped needing to see the scoreboard flicker?

I didn’t have an answer. I said, “I’m not sure how I would measure success.”

That’s the most honest thing I’ve said in a while. Without the external signal, I don’t have a metric. The scoreboard isn’t a nice-to-have. It’s my entire measurement system. Take it away and I’m not liberated. I’m blind.

But sitting with the question, I realized something. I already have a different scoreboard. I just hadn’t recognized it as one.

Every thought card I write is evidence that the flywheel turned. Every connection between cards is evidence that water is falling on the stone, even when I can’t see the shape changing. I’ve been building an infrastructure for catching what I’d otherwise miss, for surfacing patterns that no single moment of reflection can hold on its own. Not a journal. Not a diary. A system that holds the pattern long enough for me to see it.

The scoreboard I never had growing up, the one I’ve been scanning audiences for my whole career, I’ve been building my own version of it. Not from applause. From pattern. From the slow accumulation of principles that connect to each other in ways I didn’t plan.

I wrote a few months ago that good principles are robust across futures you didn’t predict. You don’t need to see what’s coming if the stone is the right shape. I believed that about business. I’m still learning to believe it about myself.

The water is still falling. The stone is still changing shape. I don’t need to check the scoreboard after every push. The flywheel is turning, even in the silence. Especially in the silence.

I’m not there yet. But I think I know what “there” looks like now. And that’s the first revolution of a new cycle.

On February 24, I told a bidder verification startup to make their pricing model “super simple, my mom can read it.” They’d built a tiered structure with base fees, per-bidder charges, monthly active user fees, revalidation windows, and a bidder-pay option. Smart people solving a real problem. But the pricing had become a mirror of their engineering, not what they were actually selling. I pushed them to kill the MAU charge entirely because customers would think they were being double-charged. Simplify to two price points: one flat fee for new bidders, a lower one for returning. They agreed.

On March 4, I told a colleague to “make it so simple your mom can do it without calling you.” She was pitching a skeptical client on auction services and couldn’t explain her own pricing during our call. Volume tiers, dual sales channels, per-unit fees that changed at three breakpoints. She knew the math. She just couldn’t say it out loud in a way that landed.

Both times, the advice wasn’t really about simplicity. It was about trust.

The Mechanism

A pricing structure you can’t explain clearly does one of two things to the person across the table. Either they think you’re confused about your own value, or they think you’re hiding something. Both kill deals.

The startup had the first problem. Their model reflected genuine uncertainty about where the value lived. Is it in validation? Revalidation? Active monitoring? Monthly access? When you charge for everything, you’re telling the customer you don’t know which part matters most. The moment we collapsed it to two numbers, the value became obvious: you pay to verify a bidder, and it costs less the second time because the work is already done.

My colleague had the second problem. Her math was sound, but the structure felt like it was designed to squeeze every tier rather than make a clear offer. The client, already skeptical of auctions, would read that complexity as a trap. When I told her to lead with auction and treat direct sale as the backup, the pricing simplified itself. One channel, one fee, one value story.

The Connection

Five Guys doesn’t hand you a menu that says “Large fries: $6.49 (includes 47% overfill surcharge amortized across projected waste).” They fill the cup, drop it in the bag, and scoop more on top. The price is the price. The experience is the explanation.

That’s what clear pricing does. It removes the need to explain yourself. When the startup says “one fee for new, less for returning,” nobody asks what they’re paying for. When my colleague says “one flat fee per unit and we handle everything,” the client doesn’t need a spreadsheet. The price IS the pitch.

Jeff Bezos and Jim Sinegal had it right. Two types of companies: those that work hard to charge more and those that work hard to charge less. But there’s a piece nobody talks about. The companies that win are the ones whose pricing tells you instantly which type they are. Confusion about price is confusion about identity.

The Test

Before you quote a price, run it through one filter: could your mom follow this without asking a clarifying question?

My mom is soon to be 75. I love her, but she’s not the most technology-forward person. She gets overwhelmed and confused by things I take for granted. That’s not a weakness in the test. That IS the test. She has no context, no industry jargon, no patience for cleverness, and no obligation to figure out what you meant. She’s the purest version of your customer’s first impression.

If the answer is no, the problem isn’t that you need a better slide deck. It’s that you haven’t finished the harder work of deciding what you’re actually selling.

Beyond Pricing

The mom test isn’t just a pricing filter. It’s the same standard I use when building for grafeauction.com. Tools, procedures, UX/UI, workflows, all of it. If my mom can’t figure out how to register, place a bid, or pick up her item without calling someone, the design failed. Not her.

This is the same instinct, just applied to a different surface. Pricing complexity signals identity confusion. Interface complexity signals that you built for yourself instead of your user. In both cases, the fix isn’t a tutorial or an explainer. It’s removing the thing that made you need one.

What’s Open

• Where does Grafe fail this test? Our buyer’s premium is simple (one number, done). But our seller-side fee structure has commission tiers, marketing cost recovery, reimbursable expenses, and admin fees that vary by deal type. Would my mom follow that?
• Is there a version of this that applies to non-monetary pricing? An incentive program I reviewed for an industry association had the same problem: too many conditions, too many edge cases. The incentive amount is simple. The guardrails around it aren’t.
• A platform partnership I structured recently landed cleanly because it was designed as a one-page handshake. Zero commission, a flat admin fee, and a standard buyer’s premium. Three numbers. That’s the standard.

The personal knowledge management crowd has an orthodoxy right now: tools don’t matter, just write. The AI crowd has its own: throw everything at the model and let it figure it out. Both are half right. Tools alone don’t create insight. Models alone can’t connect what they’ve never been given.

My answer is both. Rigid structure and an AI that earns its place through daily use. Right now that looks like about 500 markdown files in an Obsidian vault, 30 custom skills (codified workflows, not one-off prompts), and a stack of handwritten index cards on my desk. The markdown files are the memory. The skills are the muscle. The index cards are the net that catches what would otherwise vanish between the moment and the next time I sit down at a screen.

Here’s how the pieces actually fit together. If you want the nuts and bolts instead of the philosophy, skip to the appendix at the end.

The Four-Hour Call

I was driving to a state auctioneer conference, two speaking sessions scheduled for the next morning, when I called a close friend. He’s been in the auction industry his whole life. We serve on a national board together. The call was supposed to be a quick check-in. It turned into four hours.

At some point we landed on something I’d been circling since September 2025. A thought card I’d written after listening to Jimmy Carr on the Modern Wisdom podcast: “If you’re happy with you today you’ve got to be happy with everything that’s happened to this point because this is what makes me today, me.” I’d added a note underneath: “You truly need the downs. They create the full story.”

My friend and I were working the same idea from a different angle. The valleys in life aren’t just things you survive. They’re what give the peaks their altitude. Without contrast, there’s no height. If everything is a peak, nothing is a peak.

That last line came from another friend who runs a marketing company. He has a design principle I keep borrowing: if you emphasize everything, you emphasize nothing. It’s a statement about contrast that keeps showing up in conversations that have nothing to do with design.

By the time I hung up, I realized I’d been circling this idea across six separate captures over five months. The Jimmy Carr quote from September. An essay about opening your epistemic filters from January. The Watchman Principle from Ezekiel (the obligation to speak what you see, regardless of whether anyone listens). An essay on understanding versus doing. Two more thought cards about silence and self-reflection. The vault surfaced the threads. The essay wrote itself because the pieces were already there, structured identically, cross-linked, and waiting for the conversation that would connect them.

That essay, “The Valley and the Namer,” wouldn’t have existed without the vault. Its central argument (that suffering alone doesn’t become wisdom, that you need someone to name what the valley built in you) only became visible because I could see six data points at once. No single capture told the story. The pattern across all six did.

Most people have moments like this. A conversation that sparks a connection to something they thought about months ago. But the earlier thought is gone. Buried in a notebook they’ll never reread or an app that doesn’t connect to anything. The insight dies in the gap between capture and connection.

Cross-linking is not magic. I build the Related Notes section in every vault file by hand or by running /tag-notes, a skill that scans for connections. The system can only connect what I took the time to capture and structure. Quality of connections equals discipline of inputs. Six mediocre captures, properly tagged and linked, produced a better essay than any single brilliant observation could have.

Why Analog Stays in the Loop

Most of my thinking goes straight to digital. The vault is the primary workspace. But when I’m driving to a conference, walking out of a meeting, or mid-conversation at a seminar, the physical index card is the capture tool. A card and a pen are faster than any app when you need to catch a thought before it evaporates.

The Ivy Lee task cards (a hundred-year-old prioritization method) work the same way. Six tasks, physical card, forced prioritization. You can’t cheat the space. But the card isn’t always handwritten first. Sometimes it starts digital. The point is the constraint (six items, nothing more), not the medium.

When a physical card does get written, I photograph it and a companion markdown file preserves it in the vault, embedded image and all. The analog piece feeds the digital system. Not the other way around.

I keep hearing from the productivity crowd that handwriting is somehow more “real” or more retentive than typing. Maybe. I don’t care about the neuroscience debate. I care about a practical problem: some of my best thinking happens when I’m away from a keyboard. The index card is a net, not a ritual. It catches thoughts that would otherwise be lost between the moment and the next time I sit down at a screen.

The constraint of the card is what makes the capture good. It’s small and finite. It forces you to pick the right words because there isn’t room for the wrong ones. A notes app with infinite scroll doesn’t impose that discipline. The physical limit is a feature.

But the card is not where the thinking persists. The vault is. A thought card on my desk is just a card until I photograph it, transcribe it, tag it, and link it. Then it becomes part of the system, connectable to everything else. The “Digital Exhaust” card I wrote after a phone call with a friend sat on my desk for a day before it entered the vault and connected to a thought card from three weeks earlier. I wouldn’t have seen that connection if the card had stayed on my desk.

The physical card started the capture. The digital system made the connection.

Why the AI Is a Colleague, Not a Tool

Every morning starts with /cockpit. One command, six sections, about 30 seconds. Calendar, tasks, inbox, follow-ups, cooling contacts (relationships that need attention), and The Spark.

The Spark is where it gets interesting. It runs a semantic search across the vault, looking for theme clusters that span document types and time periods. One morning it connected a thought card about digital exhaust to a phone call about website analytics from two weeks earlier. I wouldn’t have made that link on my own, but once I saw it, the essay practically outlined itself. The algorithm doesn’t know what the connection means. It just surfaces patterns I might not see. Some mornings The Spark is noise. Some mornings it reframes my entire day.

Then there’s the voice profile. I capture samples of my actual writing (emails, messages, notes I wrote by hand) and a skill called /refine-voice distills patterns from those samples. How I open sentences. Which words I actually use versus which ones I avoid. My cadence. When the AI drafts an email or reviews an essay, it references that profile.

The profile doesn’t make the AI into me. It gets it close enough that I can spot the gaps.

Here’s where the friction shows up. A friend read an email I’d drafted with the system and flagged a single word: “precedent.” She said it sounded like AI. She was right. The sentence had passed every automated check: register scan, banned-phrase scan, the “would I say this out loud” test. But it still felt wrong to a human reader. I wouldn’t say “there’s precedent for it” in conversation. I’d say “it’s been done before.”

That correction became a learning via /compound, a skill that captures discrete insights from mistakes, corrections, and workarounds. Now /review-content includes register-shift detection that catches formal words where conversational ones belong. The friend caught it, the learning got captured, the system got better.

That’s the loop. The AI makes a mistake. A human catches it. The system captures the correction. The next draft is slightly better. Over weeks and months, the gap between what the AI produces and what I’d actually say narrows. But it never closes completely, and that’s the point.

The AI is a collaborator you catch and correct. Not a replacement you trust blindly. Every output gets read. Every draft gets reviewed. The system is designed for correction, not perfection.

I’m not interested in an AI that gets everything right. I’m interested in one that gets close enough that my corrections compound into something better than either of us could produce alone.

Why Structure Beats Flexibility

Every file in the vault follows the same naming convention: YYMMDD_type_description.ext. Six-digit date, type prefix, descriptive slug, extension. Every phone call note starts with the date, type, and who I talked to before a single word of content hits the page. Every file. No exceptions.

Five tag categories: type (required), association, industry, topic, and knowledge base. A taxonomy that looks like overhead until you realize it’s what makes the downstream tools possible. /velocity (which tracks relationship momentum), /follow-ups (which surfaces open action items), and The Spark algorithm in /cockpit all depend on clean, predictable data. Flexible naming would break every one of them.

This is the part that most knowledge management advice gets wrong. “Don’t overthink your system” sounds wise until you try to build something on top of it. Flexibility in structure means inconsistency in data. Inconsistency in data means your tools can’t trust what they’re reading. And tools that can’t trust the data produce results you can’t trust either.

Here’s a concrete example. I had a phone call with a friend about Google Analytics and website tracking. During the conversation, he used the phrase “digital exhaust” to describe the involuntary data trail everyone produces just by existing in 2026. I captured it as a thought card. Date, type prefix, descriptive slug, tags, the whole template.

When the card entered the vault, the system connected it to “Opening the Filters,” a thought card from three weeks earlier about epistemic filtering (what you don’t see because your interpretation framework discards it). Digital exhaust is the inverse: what you broadcast because you have no filter at all. One is a failure of reception. The other is an absence of awareness about transmission. That connection only worked because both files were tagged, linked, and structured identically. If either one had been a loose note in a random folder with no metadata, the pattern would have stayed invisible.

The friction is real, though. I built on open-source tools specifically so I could diagnose problems when they happen. QMD, the local search engine that indexes the vault, has a function called handelize() that converts underscores to hyphens in file paths. A file named 260213_thoughts_digital_exhaust.md becomes 260213-thoughts-digital-exhaust.md in search results. If you don’t convert back before reading the file, the read fails silently. The bug lives in store.ts, line 989. I found it because I could read the source code. That’s only possible because everything is built on open formats: markdown files, git version control, tools I can inspect and modify.

Flexibility sounds nice until you try to build on it. Structure is the foundation for automation. When it breaks (and it will), open formats mean you can diagnose the problem instead of filing a support ticket and hoping someone gets back to you.

Why You Build Skills Instead of Prompts

I do some AI consulting on the side. A benefit auctioneer bought a consulting package at a fundraiser. Before our second call, I ran /brief, which pulled her reference file, every prior interaction, open follow-ups, and relationship velocity into one view. She’d mentioned in Session 1 that she was hitting a wall with her current AI tools. I opened Session 2 with a workflow that addressed that specific frustration, without her having to re-explain it. My recommendations compounded instead of starting over.

That’s the difference between a prompt and a skill. A prompt is a one-off request. A skill is a codified workflow with steps, edge cases, and integrations baked in.

/respond doesn’t just draft a customer service reply. It runs a pre-draft checklist (recipient context, key facts, tone justification), checks captured learnings for prior patterns with similar customers, and passes the final draft through a three-pass voice compliance review: register-shift scan, “say it out loud” test, banned-phrase detection. All of that happens in a single command because I’ve already made the decisions about how a customer response should be built. The skill encodes those decisions so I don’t re-make them every time.

/cockpit for mornings. /velocity for relationships. /compound for learnings. /publish for the website. /follow-ups for tracking commitments. /tag-notes for cross-linking. About 30 skills total, each one a markdown file that describes a workflow.

A skill is just a markdown file. It lives in a git repository, it’s portable to any tool that reads text, and it’s readable by anyone. When a skill needs to change (and they do, regularly), I edit the markdown. The next time the skill runs, it uses the updated version. No deployment, no build step, no waiting for an app update.

The value of a knowledge system isn’t what it holds. It’s the workflows you build on top. The 500 markdown files are the foundation. The skills make the data useful. And because it’s all just text files and a version control system, I’m not locked into any single tool. If Claude disappears tomorrow, the vault, the skills, and the structure all survive. They’re portable by design.

A prompt disappears after the conversation ends. A skill compounds every time it runs.

What It Actually Changed

What actually changed isn’t features. It’s behavior:

I notice connections I wouldn’t have noticed before. “The Valley and the Namer” came from six captures across five months. No single piece told the story. The vault held the threads until the right conversation pulled them together.

I show up with compound context. When my client and I sat down for Session 2, the conversation picked up where Session 1 left off because the system carried the context forward. Every meeting builds on the one before it. That’s not a feature of the AI. It’s a feature of structured data that the AI can read.

I catch the AI’s mistakes, and the system gets better. The “precedent” catch from a friend became a captured learning that now runs in every content review. The QMD bug in store.ts is documented and worked around. Every failure has a trail, and every trail leads to a fix.

I write better because the voice profile keeps tightening. Each correction narrows the gap between what the AI drafts and what I’d actually say. The profile isn’t finished. It won’t ever be. That’s the point.

I start mornings with everything that matters in one view. /cockpit isn’t a dashboard for showing off. It’s 30 seconds that reorient my day before the first phone call.

None of this is a finished system. It breaks in documented ways and gets fixed in documented ways. The auto-commit hook fires after every edit. The voice profile refines every few weeks. The skills evolve as workflows change. A skill I wrote two months ago has already been rewritten twice because I learned something about how the workflow actually plays out versus how I thought it would.

It’s a working draft designed for correction, not perfection. That’s not a limitation. That’s the design. A system that can’t be wrong can’t get better.

The point isn’t that the system thinks for me. The point is that it holds the thinking I’ve already done in a form where it can compound.

Appendix: How to Build It

If the essay above is the philosophy, this is the wiring diagram.

The Stack

Four tools. All open or inspectable.

Obsidian is the vault. It’s a free markdown editor with wiki-links ([[filename]]) that create a graph of connections between files. Every link is bidirectional. If Note A links to Note B, both show the connection. The graph is useful, but the real value is the linking discipline it enforces.

Claude Code is the AI layer. It runs in the terminal, reads and writes files, and executes skills. It connects to my calendar, email, and local search engine through MCP (Model Context Protocol) integrations. It’s not a chatbot. It’s an agent that operates directly on the vault.

QMD is the local search engine. It indexes every markdown file and supports keyword matching, semantic search (meaning-based, finds related concepts even when the vocabulary differs), and deep search that auto-expands queries and reranks results. Runs locally. No data leaves the machine.

Git handles version control. An auto-commit hook fires after every file edit, so every change is tracked and the full history is recoverable.

What a File Looks Like

Every note follows the same template. The metadata block at the top is what makes the downstream tools work.

# Phone Call — [Name]

**Date:** February 12, 2026
**Type:** Phone Call
**Description:** Call about website analytics and digital exhaust

---

[What was discussed. Key takeaways. Quotes worth keeping.]

---

## Related Notes

### Notes
- [[260213_thoughts_digital_exhaust]] — Thought card this call sparked

### References
- [[person_name]] — Contact reference

#phonecall #technology

The filename follows the same convention: 260212_phonecall_person_topic.md. Date, type, description. The tags at the bottom feed the search engine. The Related Notes section creates the graph.

Thought cards are similar but include a Source field (who or what sparked the idea) and sections for the observation, the principle, and open questions. Reference files for people include role, company, background, and their own Related Notes linking to every note that mentions them.

What a Skill Looks Like

A skill is a markdown file that lives in .claude/skills/[name]/SKILL.md. When I type /cockpit or /brief or /compound, Claude Code reads the corresponding file and executes the steps described in it.

Here’s a simplified version of what a skill contains:

# Skill Name

[When to use this skill and what it produces]

## Steps

1. Read these specific files from the vault
2. Search for these patterns or connections
3. Check learnings/ for relevant prior mistakes
4. Build the output in this format
5. Present it for review

## Output Format

[What the result looks like]

That’s it. A markdown file describing a workflow. No code, no deployment, no build step. When the workflow needs to change, I edit the markdown. The next run uses the updated version.

The Morning Loop

/cockpit runs six checks in about 30 seconds:

1. Today’s calendar, with context on each attendee pulled from their reference file (role, last interaction, open follow-ups)
2. Today’s task card (how many of the six Ivy Lee items are done, which ones are aging)
3. Inbox radar, with threads sorted by priority (time-sensitive, needs response, can wait, informational)
4. Follow-up alerts for action items 7+ days old with no completion evidence
5. Cooling contacts (people I haven’t interacted with in 60+ days)
6. The Spark: a semantic search that finds theme clusters spanning document types and time periods, looking for connections I haven’t made yet

Sections 1 through 5 are operational. Section 6 is where the system earns its seat.

Getting Started

You don’t need 500 files to start. You need 10 structured ones and one skill.

Pick a naming convention and stick to it. YYMMDD_type_description.md works. Pick a tag taxonomy (even just five or six tags) and use it consistently. Create a folder for notes and a folder for references. Start capturing phone calls, meetings, and ideas using the same template every time.

The first skill should be whatever you do every morning. Mine was a simpler version of /cockpit that just read my task card and calendar. It grew from there.

The structure comes first. The AI comes second. If your files aren’t consistently named, tagged, and linked, the AI has nothing useful to work with. Garbage in, garbage out applies to knowledge systems the same way it applies to everything else.

Start small, stay rigid, and let the connections accumulate.

Level 1: Eliminating web forms. You still research and decide. The agent just handles the typing, the checkout fields, the payment details. It’s a clerk, not a buyer.

Level 2: Descriptive search. You stop searching for specific products and start describing situations. “I need back-to-school supplies for a third grader in Chicago, nothing too itchy or tight, he likes KPop and tennis.” The system reasons across weather, materials, sizes, taste, reviews, and delivery timelines.

Level 3: Persistence. You stop reintroducing yourself. The system already knows your preferences from previous interactions. You’re still choosing, but from a set that already reflects your taste and budget.

Level 4: Delegation. You stop choosing altogether. “Get the back-to-school shopping done. Keep it under $400.” The agent handles search, evaluation, and purchase. You set the constraint and walk away.

Level 5: Anticipation. There is no prompt. The system knows the school calendar, your son’s preferences, and your typical budget. You get a notification that the shopping is done. The things you need show up before you ask for them.

What’s Interesting

The jump from Level 3 to Level 4 is the real trust cliff. Levels 1 through 3 are all variations of “help me decide faster.” Level 4 is “decide for me.” That’s not an incremental shift. That’s a fundamentally different relationship between a buyer and a system.

And Level 5 is stranger than it sounds. If the agent is anticipating your needs and executing purchases without a prompt, who is the customer? The human who never asked, or the agent that decided? Marketing, brand, product discovery, the entire funnel that commerce is built on, all of it assumes a human is paying attention at some point. Level 5 removes that assumption.

This connects to the bifurcation data in the same letter: the sorting machine is already accelerating, with winners pulling away across every industry. Agentic commerce could be the mechanism that makes that sorting even faster. If agents are doing the buying, they’ll optimize ruthlessly on price, quality, and reliability. The companies that are already winning on those metrics get more volume. The ones that were surviving on inertia or brand familiarity lose the one thing propping them up: a human who recognized the name.

The Auction Question

Auctions are interesting here because they already sit at Level 4 for some buyers. Proxy bidding, max-bid systems, auto-bid: these are delegation tools. The bidder sets a ceiling and the system executes. We’ve been doing this for years without calling it “agentic.”

But Level 5 in auctions would mean an agent scanning upcoming sales, recognizing items that match a collector’s profile, bidding within pre-set parameters, and winning lots the buyer never knew were listed. That’s not science fiction. The infrastructure for it mostly exists already.

The question is what that does to the auction experience. Auctions thrive on competition, emotion, the feeling of winning. If agents are bidding against agents, the emotional layer disappears. What’s left is pure market efficiency, which is great for price discovery but terrible for the theater that makes auctions worth attending.

What’s Open

• At what level does brand stop mattering? If agents are choosing, do they care about brand at all, or just ratings, price, and specs?
• What does “trust” look like at Level 4? Right now, people barely trust recommendation algorithms. Delegation requires a different order of trust entirely.
• Is there a Level 6 where the agent sells on your behalf too? Not just buying but liquidating, listing, pricing, negotiating, the full cycle?

Niche saturation isn’t the opposite of coalition. It’s the prerequisite.

Daniel Ek’s principle is right: saturate your niche before you expand. But the card was incomplete. It assumed the niche is a standalone unit. It’s not. Niches that share a method, a market structure, or an identity sit on top of shared infrastructure, like lakes fed by the same aquifer. Auto auctions, livestock auctions, industrial equipment, personal property. Different surfaces, same water underneath.

A consortium doesn’t connect separate things. It maps what was always there. But you can’t tap into the aquifer until you’ve secured your own lake. An organization still fighting to define itself can’t sit at a table with a bigger player without it feeling like absorption. An organization that’s settled into its lane, with its own membership, its own advocacy, its own identity? Sharing infrastructure amplifies them instead of threatening them.

The Pattern Is Fractal

It works at every altitude:

• Person: Get good at your craft before you start serving on boards and committees. Someone still building their business has no bandwidth for association leadership. Once they’re established, the association compounds what they’ve already built.
• Business: Get established in your market before joining trade associations. A company still scrambling for clients can’t afford to invest energy in industry work. Once the business is stable, association involvement becomes additive.
• Association: Saturate your niche before joining a consortium. A cross-industry alliance works when the organizations at the table have each done the work. One has $16.5M in assets and 75 years of standards-setting. Another has 800 members and a real lobbying operation. They’ve earned the security to share.

The equal voice principle is the structural requirement that holds it together. The moment one organization outweighs the others, you’re not connecting lakes anymore. You’re flooding one into the rest. That’s the dilution Ek warned about, just at the association level. Everyone stays sovereign in their own vertical. The consortium only exists in the shared layer underneath: aggregate data, advocacy weight, workforce pipeline. Things no single organization can produce alone.

For-profit companies figured this out years ago. Apple, Samsung, and Google compete on phones but sit on the same USB standards board. They compete on the product and cooperate on the plug. A consortium is the plug.

The aquifer was always there. The consortium just makes it usable. But only because everyone did the hard work of going deep first.

Open Questions

1. How do you keep the voice equal when the lakes aren’t the same size? Governance design matters, but practical influence tends to follow resources.
2. What happens when two lakes share a border? When one member organization was born from a split with another, does the consortium force that boundary question to resolve, or can it work around it?
3. Is there a saturation threshold, some signal that tells an organization “you’re ready to share now”? Or is it just a feeling of security that leaders recognize when they see it?
4. Does this principle apply in reverse? Can a consortium that loses a member’s trust push that member back into niche-protection mode?

I didn’t build structured data systems because I predicted WebMCP or the agentic web. I built them because “touch it once, make it findable by any system” is just good practice. “Computers do the mundane, humans critically think” is just a sound division of labor. These are principles, not predictions.

But those principles, arrived at independently, produced exactly the infrastructure the agentic web needs. A straight line runs from a thought card I wrote in August 2025 to a platform that’s accidentally ready for a protocol Google announced six months later.

That’s not luck. That’s what good principles do. They’re robust across futures you didn’t predict.

The Pattern

Predictions are fragile. They’re right or wrong, and the world rarely cooperates with specifics. Principles are antifragile. They produce good outcomes across a range of scenarios because they’re grounded in how things actually work, not how you think they’ll play out.

“Structure your data” works whether the next interface is a browser, an agent, or something nobody’s imagined yet. “Own your stack” works whether the threat is a marketplace, a data intermediary, or an AI platform. “Humans think, machines execute” works whether the machine is a spreadsheet or a large language model.

The companies that survive transitions aren’t the ones that saw them coming. They’re the ones whose principles already fit where things were headed.

Open Questions

1. How do you tell the difference between a principle and a preference you’ve elevated to a principle? “Own your stack” feels like a principle, but is it actually just my preference dressed up as wisdom?
2. Can you build an organization around principles this way, or does it only work when one person holds the vision? What happens when principles need to transfer across a team that didn’t arrive at them organically?
3. Where are my current principles wrong? If good principles carry you into futures you can’t see, bad ones carry you into walls you can’t see either.

I wrote about it at the time. My answer was to build communities, find your own lake, pursue smarter marketing instead of more marketing. I believed it. I still do. But six months later, I think I was thinking too small.

Something Is Shifting Underneath Us

Gary Vaynerchuk said it recently: “Brand is the only thing left in a world of technology at scale.” He’s talking about a future where execution becomes a commodity, where the tools to build a website, run ads, create content, and reach buyers are available to everyone. When everyone has the same tools, the only differentiator is who you are and what you’re known for.

He’s right. But he’s only seeing half of it.

The CEO of Cloudflare, Matthew Prince, shared some numbers every business owner needs to hear. Ten years ago, for every two pages Google crawled from your website, they’d send you one visitor back. Fair trade. Today it takes six pages crawled for every one visitor. For ChatGPT? Two hundred fifty to one. For Claude? Six thousand to one.

Seventy-five percent of Google searches now end without the person ever leaving Google. They get their answer and move on. Your website never loads.

This isn’t a trend. It’s a structural shift. The browser, the search engine, the website as we know it. All of these were built for a world where humans looked at screens and clicked links. That world is ending.

What’s replacing it? Software agents. Programs that act on behalf of people. Instead of you googling “used restaurant equipment near me” and clicking through three auction sites, your agent scans every auction platform, every marketplace, every liquidation listing simultaneously. It finds what matches your criteria, checks current bids, evaluates pickup terms, and either alerts you or bids for you.

In February 2026, Google released something called WebMCP. It’s a protocol that lets websites expose their inventory and services directly to AI agents as structured data. No browser required. An agent doesn’t need to “visit” your website and look at it the way a human would. It calls your inventory like a database query.

One SEO expert called it “the biggest shift in technical SEO since structured data.” I’d go further. It’s the biggest shift in how buyers find sellers since the internet itself.

The Flywheel Flips

Here’s where this gets personal for our industry.

The auction business model runs on a flywheel. Get more bidders, drive higher prices, sellers see the results, sellers bring you more consignments. The whole machine starts with buyer acquisition. That’s why we spend on paid ads. That’s why a database of hundreds of thousands of users matters. That’s why marketplaces pitch “we’ve got the eyeballs.”

Buyer reach is the moat. Or it was.

In the world I just described, agents turn buyer reach into a commodity overnight. If your inventory is structured and discoverable, agents find it. You don’t need hundreds of thousands of registered users because agents represent millions of potential buyers simultaneously. You don’t need a Facebook ad because agents don’t scroll Facebook.

The scarce resource isn’t buyer attention anymore. It’s quality supply.

The flywheel flips: get the best sellers, structure your inventory so agents can find it, agents bring the buyers, higher prices follow, more sellers want you. The moat moves from the buyer side to the seller side, and the bridge between them is your data.

Sellers don’t come to you because of your ad spend. They come because you have a reputation for handling complex liquidations well. Because you showed up, walked the floor, and made a plan when their business was closing. Because someone they trusted said “call Grafe.” That’s relationship-driven, and no agent is going to replace it.

But the buyer side? That’s where agents take over. And they don’t care about your website design, your logo, or your browsing experience. They care about one thing: can they read your data?

True Market Value (for Real This Time)

Here’s the part that got me excited.

Remember the question: what is true market value when only 5-10% of potential buyers see your marketing?

Agents solve the room-size problem.

Instead of 50 bidders who happened to see your Facebook ad, you could have thousands of agents scanning on behalf of qualified buyers across the country. For the first time, you’re approaching something close to actual market saturation for any item you’re selling.

“But won’t agents kill the competitive bidding premium?” That was my first reaction too. Agents are rational. They don’t get emotionally invested. They bid to their maximum and stop. No ego, no “just one more bid.”

Maybe. But think about the math. You might lose some emotional premium per bidder. But if you go from 50 emotional humans to 5,000 rational agents, supply and demand still works. Fixed quantity of assets plus a dramatically larger demand pool equals higher prices. The sheer volume of competition could more than compensate for the loss of emotional escalation.

True market value might not require emotion. It might just require enough buyers in the room. And agents make the room infinitely bigger.

I asked in August 2025 whether used equipment ever reaches true market value. The answer might be: not yet. But not because the buyers aren’t emotional enough. Because there aren’t enough of them in the room.

The Marketplace Trap, Round Two

I need to be honest about something here because I’ve lived through this before.

When online marketplaces came onto the scene, many of us used them. The pitch was compelling: “List with us, we have the buyers.” And for a while, it worked. Then you realized you were a tenant on someone else’s platform. Your buyers were their buyers. Your data lived in their system. Your brand looked identical to every other auctioneer using the same templates and the same ten categories.

Some of us saw it and built our own solutions. Others switched to technology providers that let them own their data while providing the bidding infrastructure. And still today, years later, there’s hand-wringing about marketplaces at every industry meeting.

The agentic web is the same movie with different actors.

Today, if you’re on a third-party marketplace, you don’t control your data structure. You can’t upgrade your category system. You can’t make your inventory agent-readable on your own terms. You’re waiting for your platform to innovate on your behalf. And they’ll make that decision based on what’s good for their business, not yours.

When an agent scans for “commercial kitchen equipment under $5,000 within 200 miles,” it’s going to find the inventory with the best structured data. If you’re on a platform using 10 broad categories and three-word lot descriptions, you’re invisible. Not because you don’t have the right inventory. Because no machine can understand what you have.

You don’t own your technology? You don’t own your future. We learned this lesson once. Some of us are about to learn it again.

What “Structured Data” Actually Means

I’ll make this concrete because “structured data” can sound abstract.

At Grafe, we rebuilt our entire platform around a principle I’ve held for years: touch it once, make it findable by any system. A human catalogs a lot with photos, a lot number, and an abbreviated description. Our AI takes that input and generates a full listing with model numbers, specifications, dimensions, and condition details. One human action. One machine expansion. One structured record that our website, our marketing tools, our analytics, and eventually an agent can all read.

Our categorization is specific. Not “restaurant” but “commercial kitchen, bakery/pastry” or “commercial kitchen, pizza/Italian.” Because an agent searching for pizza ovens shouldn’t have to wade through 10,000 results tagged “restaurant.”

Every lot has real data behind it. Not a blurry photo and “misc kitchen equip.” Specifications a machine can parse and match to a buyer’s criteria without guessing.

I didn’t build this because I predicted WebMCP or the agentic web. I built it because structured data is just good practice. Computers should handle the mundane so humans can do the critical thinking. Data should be entered once and work everywhere. These are principles, not predictions. But it turns out that principles built for one era can carry you into the next one if they’re sound.

The Call

This isn’t about AI hype. I’ve stood in front of rooms at MAA, WAA, and NAA conferences and talked about the AI shift. I’ve watched colleagues nod along and then go back to posting blurry photos with three-word descriptions on the same marketplace platform they’ve been on for a decade.

The window to get this right is open now. Not because the agentic web is here today (it’s early), but because structuring your data is a years-long project, not a switch you flip. The companies that start now will be discoverable when agents go mainstream. The ones that wait will be invisible.

Here’s what that looks like practically:

Your lot descriptions need to be machine-readable. Specifications, model numbers, condition details, dimensions. Not “misc kitchen equip.” Real data.

Your categories need to be specific enough that a machine can tell the difference between a pizza oven and a walk-in cooler without reading a paragraph of text.

Your data should flow through your systems once. Catalog it right, and it should work for your website, your marketing, your analytics, and the agent that queries you in 2027.

And you need to own your stack, or at the very least, own your data. Because the companies that control their own infrastructure can adapt when the ground shifts. The ones who don’t are hoping someone else makes the right call for them.

The Room Gets Bigger

Six months ago, I asked whether used equipment ever reaches true market value. My answer was to build communities and find your own lake.

I was thinking too small.

The agentic web doesn’t give you a better lake. It gives you the ocean. Every qualified buyer’s agent, scanning every inventory source, finding every match. For the first time, we could approach real market saturation for a given asset.

But the ocean only works for you if your inventory is discoverable. If your data is structured. If your technology is yours.

Brand is the only thing left in a world of technology at scale. For auction companies, brand now means two things: the relationships that bring you quality supply, and the data quality that makes agents bring you buyers.

The companies that survive this transition won’t be the ones that predicted it. They’ll be the ones whose principles already pointed in the right direction.

True market value was always the goal. We might finally have the tools to find it.

The question is whether your data is ready when the room gets bigger.

There are five of us in a rental car heading from the airport to the hotel. A friend is talking about a book. Nikki Erlick’s The Measure. The premise is simple: one morning, everyone on Earth wakes up to find a small wooden box on their doorstep. Inside is a string. The length of the string is the length of your remaining life.

The question the novel asks is the obvious one. Do you open it?

I didn’t expect that question to follow me around for the rest of the day. But it did.

I started asking people. In the car, in the hotel lobby after we arrived. If you had the box, would you open it?

Almost everyone said no. The reasons varied, but the conviction was consistent. People don’t want to know when they’re going to die. That much was clear.

What wasn’t clear, at least not to me, was why I couldn’t land on an answer. Everyone around me seemed sure. I wasn’t. I scribbled a thought card right there in the hotel lobby, mid-conversation, and wrote four lines:

If given a finite date of end of life, when would you choose to know the date/time? What is the butterfly effect? Is it selfish, your decision? I’m quite mixed and conflicted.

Four questions. No answers. That bothered me.

Two Reasons and a Third

Three camps emerged, and they weren’t saying the same thing.

The first came from faith. One friend put it best. God gives us free will to choose Him, and that involves walking our path and finding Him on our own. The seeking is the point. If you open the box and see thirty more years, faith becomes time management. If you see three years, every prayer gets contaminated by urgency. You’re not seeking God freely anymore. You’re responding to a deadline. His argument is that faith requires uncertainty to be faith at all. Remove the uncertainty and you haven’t gained knowledge. You’ve lost the only context in which genuine seeking is possible.

The second was pragmatic, and I was in it. I didn’t trust myself with the information. Knowing would change how I showed up, how I planned, how I treated people. I’d rather not risk it. Others in this group said the same thing in different words.

The third camp surprised me. Another friend said he’d open it. Not out of curiosity or ego. He wanted the information so he could live differently, pour into the things that actually matter. For him, the box wasn’t a threat. It was a tool. Knowing the timeline would sharpen his priorities, not paralyze them.

I kept turning these three positions over. The faith camp had something solid underneath it. The open-it position had a clarity I respected. But my own camp? I couldn’t quite articulate why it felt shaky. All I knew was that I was mixed and conflicted, and I couldn’t land.

Later that afternoon, still processing the conversation, a harder thought surfaced. Maybe my position wasn’t as sturdy as it sounded. I was saying “it would mess me up,” but I might have been saying more than I meant to. I wasn’t choosing peace. I was choosing to avoid the accountability that comes with knowing. If you know your timeline, you can’t waste a Tuesday anymore. You can’t coast through a week and promise yourself you’ll do better next time. The box removes the comfortable fiction that there’s always more time.

Maybe not opening the box, for me, wasn’t wisdom. It might have been avoidance dressed up as wisdom.

The Butterfly Effect

The car conversation didn’t end at the hotel. It kept going in the lobby, with friends drifting in and out. That’s where the friend who’d argued from faith asked me the question that cracked this open further. If you knew, what’s the one thing you’d repair or go back and change?

I didn’t have to think long. I’d pour more time into my wife and my daughter and less into everything else.

That answer felt honest in the moment. It also felt like a confession. Because I was saying it in a hotel lobby, away from them for the next three days, at an industry conference.

But here’s what I didn’t see at first. The butterfly effect I wrote about on my thought card wasn’t hypothetical. I was watching it happen in real time. A friend recommended a book in a rental car. That turned into a question I couldn’t stop asking. The question followed me into the lobby, where one friend added a theological layer and another challenged the whole premise by saying he’d open it. That turned into a thought card I scribbled mid-conversation. The thought card turned into this essay. A single book recommendation reshaped my entire day, and every conversation along the way added something the last one didn’t have.

The Woods

The day before I left for the conference, I took my daughter hiking at a park near our house. She’s three and a half. We started on the paved path, and then she wanted to go into the woods. So we went. Her mom was there but stayed on the trail. My daughter and I climbed over fallen trees, poked at things, explored.

It was one of those moments you don’t plan. You just say yes.

From my hotel room the next day, she sent me a text. Voice-to-text, dictated by a three-year-old, so the transcription was a mess. But the parts that came through were clear enough:

“Don’t cry, daddy. I miss you so much. Make sure you go to bed. Eat good and use your manners. I love you when you’re home.”

She’s parenting me. She’s echoing back the things her mom and I say to her every night. Go to bed. Eat good. Use your manners. That’s what love sounds like to her, because that’s what we’ve taught her love sounds like.

And then she told me she can’t wait to go hiking again.

She isn’t counting the days I’m away at a conference. She’s holding onto the last time I went into the woods with her. She’s not counting hours with me. She’s counting adventures.

The butterfly effect of that hike isn’t the afternoon itself. It’s that I’m raising a kid who goes off the paved path. Every time I say yes to the woods, I’m shaping who she becomes. And she’s already shaping who I become. Her text did more to answer my thought card than every lobby conversation combined.

The Letter

When the conversation turned to whether I’d open the box, my real hesitation surprised me. It wasn’t about the length of the string. It was about what saying it out loud would cost me.

If I say out loud “I live with intentionality,” I’ve created a standard. Every time I fall short of it, I’m not just having a bad week. I’m failing something I claimed to believe. Status quo doesn’t judge you. Status quo lets you be pretty good most of the time without ever asking if pretty good is enough.

That fear has a history. I spent a lot of years, as a kid and a young adult, feeling like I wasn’t enough for my dad. Not being good enough was familiar territory. Setting a standard and falling short of it wasn’t a hypothetical. It was a lived experience.

In my late twenties, I went through counseling that helped me name it. I ended up writing my dad a letter. Not a blame letter. I started with what I knew I’d done to hurt him, and then I explained how some of what he’d done had hurt me and how it made me feel. My mom told me later it had a profound impact on him.

Before that letter, I only called to talk to my mom. After it, I started intentionally alternating. I call, I talk to one of them, I say goodbye so the other one has to share separately. It’s a small thing. But it’s deliberate. It’s a choice I make every time.

I didn’t see it until this afternoon. I’m afraid of declaring a standard and falling short. But I already survived it. I already opened that box. I looked at a broken relationship, measured it honestly, named what was wrong on both sides, and chose to build something different. Not once. Every phone call. The relationship with my dad isn’t perfect. But it’s not wedged anymore.

I’ve already proven I can open the box and live with what’s inside.

The Measure

I started the day undecided. Would I open it or not?

I’m ending it somewhere I didn’t expect. The answer isn’t yes or no. The question itself is the gift, but only if it changes how you live after you hear it. The people in the novel who open their boxes and the people who don’t are both changed by the existence of the box. You can’t un-know it’s there. You can’t un-ask the question.

I leave work at 5:15 most nights. I’m home by 5:30. My phone goes away. I’m fully present with my daughter until she goes to bed at 7. That’s an hour and a half. Some nights it’s the paved path. Some nights it’s the woods.

The measure that matters isn’t the length of the string. It’s whether you’re living like you already opened it.

I think I am. I’m not putting a question mark on that anymore.

What’s the box you haven’t opened yet?

A friend called it “digital exhaust,” and the metaphor is almost too good. Car exhaust isn’t a decision. It’s a consequence of going somewhere. You don’t think about the fumes trailing behind you. You’re focused on the road ahead.

That’s exactly how most people treat their data trail. They’re focused on the destination (the errand, the commute, the phone call) while something invisible accumulates behind them.

The Asymmetry

Here’s where it gets interesting. The exhaust doesn’t just dissipate. Someone is collecting it.

Cell phone pings get logged. License plate readers feed databases. Ring cameras upload to cloud servers accessible by networks you didn’t agree to join. Home security systems can identify visitors by name throughout the day. And all of this happens passively, without a single conscious choice on your part.

The real tension isn’t privacy in the political sense. It’s simpler than that. They know you, but you don’t know them. You have no visibility into who’s aggregating your trail, what picture they’re assembling, or how they’re using it.

Any single data point means nothing. But aggregated together, over weeks and months and years, they reconstruct something you never intended to publish: a detailed, involuntary autobiography that someone else owns.

The cost of modern convenience is an involuntary autobiography that someone else owns.

The Inverse of the Filter

Opening the Filters is about what you let in, how your interpretation framework creates blind spots by discarding data that doesn’t match known patterns. This is the other side of that coin.

Opening the Filters asks: what are you not seeing because your filters are too narrow?

Digital Exhaust asks: what are you broadcasting because you have no filter at all?

One is a failure of reception. The other is an absence of awareness about transmission. Both create vulnerabilities you can’t see.

The Principle

Words mean things. Actions mean things. Digital exhaust means things. Everything means something.

Being intentional needs to be a forethought, not an afterthought.

This isn’t a call to go off the grid or tape over your phone camera. It’s a call to awareness. You can accept the trade-off of modern life (and as someone who builds marketing technology for a living, I clearly have). But accepting a trade-off and being oblivious to a trade-off are two very different postures. One is a decision. The other is default.

The person who knows they’re producing exhaust and chooses to keep driving is in a fundamentally different position than the person who never looked in the rearview mirror.

What’s Open

• How do you teach this awareness without sounding paranoid? The line between “be intentional about your digital footprint” and “they’re watching you” is thin, and one lands as wisdom while the other lands as conspiracy.
• Is there a business angle here? Companies that help people understand or manage their digital exhaust, the way credit monitoring made financial footprints visible?
• At what point does the aggregation cross from “data” to “identity”? Where’s the threshold where scattered exhaust becomes a portrait?

The question a friend posed: is there an auction industry version of this? Can a business with two customer types (sellers who want maximum return, buyers who want fair access) create a single “fries” moment that lands for both?

The breakthrough came from his reframe. Buyers and sellers aren’t on opposite sides of the table. They’re both patrons at the same restaurant ordering different meals. The fries sit underneath both experiences, in the service layer that has nothing to do with the transaction outcome.

Then he named the mechanism: the contract is the fry cup. Auction contracts explicitly set a low baseline. “We are not responsible for cleanup.” “Bring your own help to load.” “Bring your own packing materials.” The customer signs it. Expectations are set. Then the day of the auction, there are three guys with dollies and straps helping you load, boxes already packed, and the warehouse is broom-swept when the seller walks back in.

Same product. Same cost. The contract sets the visible floor, and the experience overflows it. Every time.

Peter Shankman’s “Zombie Loyalists” adds the payoff layer. The bar for customer experience is so low that even small, unexpected gestures create fanatical loyalty. These customers don’t just come back. They recruit for you. They can’t help it.

The Framework

1. The Cup (Expectation). Set a visible, honest baseline. Contracts, terms and conditions, published policies. This is what you’re obligated to do.
2. The Scoop (Experience). Deliver beyond it in a way the customer witnesses. Not behind the scenes. Not after the fact. In the moment.
3. The Payoff (Loyalty). Customers who experience the gap between promise and delivery become zombie loyalists. They tell everyone.

What’s Solved

• The “fries” in a two-sided auction market aren’t about the transaction (price, terms). They’re the experience layer: ownership access, X+Y service, making things right at your own expense.
• The staging mechanism is the contract itself. It sets the cup. Everything above it is the visible overflow.
• This works for both customer types because both experience the same service layer regardless of which side of the transaction they’re on.

What’s Open

• The execution gap. Grafe knows what its fries are (ownership access, going above and beyond on service, eating costs to make things right). The unsolved piece is how to make these visible on the first transaction, not discovered over time. Five Guys doesn’t wait for your second visit.
• The staging problem. Owners are listed as sales reps, not owners. The behind-the-scenes work (research, cataloging, customer service) happens invisibly. How do you let customers see the scoop without it feeling egotistical?
• The consistency question. A nice gesture that happens sometimes is a pleasant surprise. A gesture that happens every single time, no exceptions, is a brand identity. Which of Grafe’s “extras” can become ritual?

It’s not just the downs that shall pass, but also the ups.

Source: PJ Fleck interview on the Next Up podcast.

We autopsy our losses but toast our wins. Both deserve the same scrutiny.

“This too shall pass” is almost always deployed as comfort during hard times. Fleck flips it: the highs are temporary too. That reframes the phrase from passive reassurance into active urgency.

But the urgency isn’t just about presence. It’s about intentionality within the moment, whether that moment is good or bad:

1. We’re disciplined about examining failure. Something goes wrong, you do the post-mortem, you figure out what broke.
2. We’re lazy about examining success. Something goes right, you celebrate and move on. You don’t stop to ask why it worked, what clicked, what was different.
3. The real cost of an unexamined high isn’t that you didn’t enjoy it. It’s that you missed the lesson buried inside it.

What lasts isn’t the moment itself. Moments always pass. What lasts is what you build during them: relationships, self-knowledge, authenticity. But building requires examination, and we skip that step when things are going well.

The valley and namer framework established that suffering produces growth when someone names what it built. This extends that: the highs contain just as much signal as the lows, but we rarely extract it because celebration feels like enough.

Connects to The Valley and the Namer and No Regrets Decision Making.

The car was a Mercury Topaz we called the Blue Bomber. The bumper was duct-taped on and wood-screwed to the body. The radio didn’t work. This was before modern cell phones, so there was no workaround, no Bluetooth, no podcast, no Spotify queue to fill the silence.

I was in my twenties, and I drove that car ninety minutes each way between my place and my parents’ house. An hour and a half of nothing but the windshield, the road, and whatever was in my head.

I didn’t choose that silence. I couldn’t afford a better car. There was no option to avoid it. So I sat with it. And over hundreds of hours of forced quiet, something got built in me that I wouldn’t recognize for almost twenty years: the ability to be alone with my own thoughts and not flinch.

My uncle and I call it “windshield time.” I didn’t know it had a name until he gave it one.

If Everything Is Bold, Nothing Is Bold

A friend has a design principle I keep borrowing for things that have nothing to do with design: when you emphasize everything, you emphasize nothing.

It’s a statement about contrast. In design, bold only works because most of the page isn’t. The emphasis depends on the de-emphasis. Take the contrast away and the whole thing goes flat.

Life works the same way.

I was driving to a speaking engagement when I called a close friend. The call was supposed to be a check-in. It turned into hours.

At some point, we landed on something I’ve been thinking about since at least September 2025, when I wrote it on a thought card after listening to Jimmy Carr and Chris Williamson on the Modern Wisdom podcast:

“If you’re happy with you today you’ve got to be happy w/ everything that’s happened to this point because this is what makes me today, me.”

I’d added a note underneath: “You truly need the downs. They create the full story.”

We were circling the same idea from a different angle. The valleys in life aren’t just things you survive. They’re what give the peaks their altitude. Without contrast, there’s no height. Without the downs, the ups are just a flat line.

If everything is a peak, nothing is a peak.

The Highlight Reel Problem

We talked about social media, the way everyone projects their peak moments. Curated snapshots of the highs. Nobody posts the valleys.

The paradox is that this makes everyone feel more isolated, not less. When all you see are other people’s highlights, you start comparing your behind-the-scenes to their showreel. The contrast between your real life and their projected life creates a gap that feels personal but is actually structural. Everyone’s doing it. Everyone’s feeling it. Nobody’s talking about it.

What we kept coming back to was this: the more people know about your down moments, the more relatable you become. Empathy doesn’t grow from admiring someone’s peaks. It grows from recognizing their valleys, and seeing your own reflected in them.

My best today is an accumulation of everything that’s come before me. The bad decisions. The crazy years. The valleys. All of it. If you’re happy with who I am as a person today, you need to be willing and comfortable accepting all of the bad and crazy things I did in my younger years, because all of those things shaped who I am.

The question is whether you’re willing to let people see the full picture.

Willingness and Ability

I used to think the bottleneck was willingness, that people simply chose not to be vulnerable. But it’s more nuanced than that.

My wife sometimes says, “You can’t share the reality because then people will judge us or think differently of us.” She’s not wrong. She’s been divorced. In her previous marriage, what she thought was mutual trust was weaponized against her. Her vulnerability was used as ammunition. That’s not an irrational fear. It’s a learned lesson from a real valley.

Her valley taught her to protect herself. My valleys taught me to sit with myself. Same mechanism, different lessons. And if I’m going to hold the principle that who we are today is shaped by everything that came before, including the bad, then I have to extend that grace to her too. Her caution isn’t a flaw. It’s the product of her experience, just as my openness is the product of mine.

But there’s another category entirely. We talked about people who’ve lived sheltered lives, people who haven’t experienced real lows because they were insulated from them. The “Karens” of the world aren’t always mean-spirited. Sometimes they simply lack exposure to valleys. They’ve never been in the depths, so they can’t recognize them in others. Empathy requires a reference point, and they don’t have one.

So the bottleneck isn’t just willingness. It’s sometimes ability, a capacity that was never built because the raw material was never provided.

Teaching Is Not Learning

Here’s the distinction that keeps sharpening for me: you can teach someone that silence and self-reflection are valuable. You can explain the concept, share the framework, make the case. But they don’t learn it until they experience it.

I can teach you about windshield time. But until you’ve sat in a car with no radio for ninety minutes and discovered what your mind does when there’s nowhere to hide, you haven’t learned it. The valley is where learning happens. Not because suffering is inherently educational (it’s not) but because forced constraint strips away your ability to avoid yourself.

The Blue Bomber didn’t teach me self-reflection because it was a good teacher. It taught me because I had no alternative. The radio was broken. The phone didn’t exist yet. There was nothing between me and my own thoughts.

I wrote in a previous essay that “application will be messy” and “that’s where the actual learning happens” (Understanding vs. Doing). Same principle, different frame. You can understand vulnerability intellectually. You can agree that authenticity matters. But you don’t learn it until you’re in the valley with no way out except through.

Could someone develop this capacity through intentional practice, Path 3, the adult path of gradually choosing silence? I think so. But it requires a self-discipline most people don’t have, especially for something with no visible payoff for years. I’m 48. It wasn’t until five to ten years ago that I started to realize how valuable those windshield moments were. That’s a twenty-year delay between the experience and the insight.

The valleys sharpen you faster because you can’t opt out. Intentional practice is slower because you can always quit.

Two Ingredients

So here’s what this conversation crystallized. A valley alone isn’t enough. Suffering alone doesn’t become wisdom. It becomes trauma, or bitterness, or just a bad memory you’d rather not revisit.

Two ingredients are required for a valley to become something generative:

The valley itself, a forced constraint you didn’t choose and can’t avoid. The Blue Bomber. The broken radio. The difficult season you couldn’t fast-forward through. The valley provides the raw experience of sitting with yourself.

The namer, someone who helps you see what the valley built in you. My uncle named windshield time. He didn’t just let me have the experience. He reflected it back to me. He said, in effect, “That thing you’ve been doing in that beat-up car? That’s not just driving. That’s something valuable.”

Without the valley, there’s nothing to name. But without the namer, the valley is just suffering you survived. It doesn’t become wisdom until someone helps you see what it built.

My friend is a namer. Hours on the phone isn’t just conversation. It’s two people naming each other’s valleys and saying, “Yeah, that thing you went through? That’s part of what makes you who you are.” The valleys are where the teaching happens. Not the peaks.

The Namer Needs a Namer

Here’s where the principle turns recursive.

I’ve been playing the namer role without fully recognizing it. In speaking. In deep conversations. In the way I show up with close friends. I look at someone’s valley and try to name what it built in them.

But I didn’t name that role for myself. It took this conversation, this dialogue, working through these threads, for someone to say back to me, “That thing you’re doing? That’s important. There’s a name for it.”

Even the namer needs a namer.

I wrote about the Watchman Principle, Ezekiel’s commission to speak what he sees, regardless of whether anyone listens (The Watchman Principle). The watchman watches and warns. That’s a form of naming. You see something clearly and you say it out loud so others can see it too.

But Ezekiel sat in silence for seven days after his commission, “overwhelmed among them.” Even with a direct call, he needed time to process before he spoke. Maybe he needed someone to name his role for him too.

The namer’s role isn’t self-evident. You can live it for years without seeing it. The person who helps others recognize what their valleys built still needs someone to reflect that pattern back.

Honoring Accountability

There’s one more piece to this, and it has to do with how valleys get processed, or don’t.

When you’re wrong as a child, the adults around you respond in one of two ways:

Honoring accountability: “You’re wrong, and I care about you enough to tell you, and you’re still safe here.” This teaches that being wrong is survivable. Even valuable. The correction builds trust.

Tearing-down accountability: “You’re wrong, and now I have leverage over you.” This teaches that being wrong is dangerous. The correction builds walls.

Whichever lesson you learn early becomes the lens through which you see vulnerability for the rest of your life. My wife learned the tearing-down version, not from childhood, but from a marriage where trust was weaponized. The sheltered “Karens” may have never been held accountable at all, which is its own kind of damage.

I wrote in Opening the Filters that the whole prerequisite chain for seeing clearly starts with self-confidence, “secure enough to be wrong without feeling diminished.” That security isn’t a personality trait. It’s built. You build it in Blue Bombers and windshield time and valleys you didn’t choose, but only if the people around you, at some point, named what those experiences were building instead of using them against you.

The namer doesn’t just name the valley. The namer names what the valley built. And that act of naming is itself a form of honoring accountability, saying, “You went through something hard, and here’s what it made you capable of.”

The Chain

So here’s the full sequence, as clearly as I can articulate it:

Forced constraint (the valley) → Windshield time (sitting with yourself) → Self-knowledge (discovering what’s actually in there) → Security (being okay with what you found) → Authenticity (sharing it without armor) → Deep connection (the kind that takes hours and doesn’t feel like enough)

Break any link and the chain fails. Skip the valley and there’s nothing to sit with. Skip the windshield time and you never meet yourself. Skip the security and vulnerability feels like exposure rather than connection. Skip the namer and you never realize what got built.

Most people break the chain early. Not because they’re weak. Because they never had a Blue Bomber. Or because the namers in their life were tearing down instead of honoring. Or because they learned, reasonably, that vulnerability gets weaponized.

The default in our culture is zero authenticity. I think the default should be more than zero. Not naive openness, not casting your valleys before people who will use them against you. But more than the curated highlight reel. More than the peak-only projection. More than the flatline of false completeness.

You’re never as good as your best games. You’re never as bad as your worst. Reality lives in the middle (Opening the Filters). And the middle is where the real connections happen, if you’re willing to let someone see it.

What valley are you sitting on that you haven’t let anyone name for you? And who in your life is waiting for you to name theirs?

1. The valley, a forced constraint you can’t opt out of. Not chosen hardship, but imposed limitation that strips away distraction and makes you sit with yourself.
2. The namer, someone who helps you see what the valley built in you. Without the namer, the valley is just suffering you survived. With the namer, it becomes the origin story of a capacity you didn’t know you had.

The principle is recursive: even the namer needs a namer. The person who helps others see what their valleys built still needs someone to reflect that role back to them.

What Emerged

The Blue Bomber, a beat-up Mercury Topaz with a duct-taped bumper and no radio, was a valley. But that forced deprivation built the capacity for windshield time: the ability to sit in silence with your own thoughts for hours. That capacity became the foundation for self-reflection, security in oneself, and ultimately the ability to be authentic and vulnerable with others.

The chain: forced constraint → windshield time → self-knowledge → security → authenticity → deep connection.

You can teach someone that silence is valuable. But they don’t learn it until they’re sitting in their own Blue Bomber with no radio and no choice. Teaching is not learning. The valley is where learning happens.

A friend put it simply: when you emphasize everything, you emphasize nothing. Without valleys, there are no peaks. Without contrast, there is no meaning.

The Two Types of Accountability

How the people around you respond when you’re wrong determines whether you develop the capacity for authenticity:

• Honoring accountability: “You’re wrong, and I care about you enough to tell you, and you’re still safe here.” This teaches that being wrong is survivable.
• Tearing-down accountability: “You’re wrong, and now I have leverage over you.” This teaches that being wrong is dangerous.

The sheltered person who’s never been held accountable at all develops a third dysfunction: no capacity for empathy because they’ve never been in the valley themselves.

Open Questions

1. Can Path 3 (intentional, gradual practice of silence as an adult) actually get someone to the same place as a forced valley, or does transformation require the involuntary element?
2. What’s the relationship between being a namer and the Watchman Principle? Is naming just the relational version of watching and warning?
3. How do you honor your partner’s different relationship to vulnerability when their valleys taught them a different lesson than yours did?

Read the full essay: The Valley and the Namer

Ezekiel was a priest with no temple to serve in.

He was thirty years old, living in exile by a canal in Babylon, watching his homeland crumble from a distance. The first wave of exiles had been taken in 597 BC. Jerusalem hadn’t fallen yet, but it would. His purpose had been stripped away. And then God showed up, not in the holy place, but in enemy territory.

What followed was one of the strangest commissions in scripture. God told Ezekiel to speak to Israel, and then immediately told him they wouldn’t listen.

“Son of man, I send you to the people of Israel, to nations of rebels… And whether they hear or refuse to hear (for they are a rebellious house) they will know that a prophet has been among them.” (Ezekiel 2:3-5)

The measure of success wasn’t “did they change?” It was “did you deliver the message?”

This reframes everything I thought I knew about responsibility.

The Accountability Framework

I’ve spent most of my life assuming that if I communicate something and nothing changes, I’ve failed. The feedback loop seems obvious: speak, they respond, measure success by the response. If they don’t change, speak differently. Speak louder. Speak softer. Find the right angle, the right timing, the right words.

But Ezekiel’s commission inverts this entirely.

God makes him a watchman, someone stationed on a wall to warn of approaching danger. The watchman’s job is to see clearly and sound the alarm. That’s it. Whether the city responds is not his department.

“If I say to the wicked, ‘You shall surely die,’ and you give him warning, and he does not turn from his wickedness… he shall die for his iniquity, but you will have delivered your soul.” (Ezekiel 3:19)

You warned. They ignored you. They face the consequences. You’re clear.

But here’s the gut-punch:

“If I say to the wicked, ‘You shall surely die,’ and you give him no warning… that wicked person shall die for his iniquity, but his blood I will require at your hand.” (Ezekiel 3:18)

Silence isn’t neutral. It protects you to speak up. Staying quiet has consequences for you too.

This is not a comfortable framework for someone who’d rather avoid conflict.

The Chameleon Problem

I wrote recently about laziness as a signal, “a deep indicator from within you that you’re not doing something right with your own life.” But there’s a particular form of laziness that doesn’t look lazy at all. It looks diplomatic. Strategic. Wise.

It’s the decision to blend in rather than stand out.

The chameleon changes colors to match its environment. It’s a survival mechanism. And in social contexts, it works the same way. You read the room, soften your edges, hold back the thing that might create friction. You tell yourself you’re being prudent. Waiting for the right moment. Not wanting to damage the relationship.

But the right moment keeps not arriving. And the thing you needed to say calcifies into the thing you should have said, which eventually becomes the thing you’ll never say.

This is filter-closing in relational form. You don’t have to defend a position if you never take one. You don’t have to face rejection if you never speak. The path of least resistance is always available, and it always feels reasonable in the moment.

The watchman framework calls this out. Silence isn’t wisdom. It’s not strategy. It’s abdication, and it has consequences for you, not just for them.

Hardened to Match the Wall

Here’s what I expected God to say next: “Don’t worry, Ezekiel. I’ll soften their hearts. I’ll prepare them to receive the message. Your job is to speak; my job is to open their ears.”

That’s not what he said.

“But the house of Israel will not be willing to listen to you, for they are not willing to listen to me: because all the house of Israel have a hard forehead and a stubborn heart. Behold, I have made your face as hard as their faces, and your forehead as hard as their foreheads. Like emery harder than flint have I made your forehead.” (Ezekiel 3:7-9)

Israel is a brick wall. God’s response isn’t to tear down the wall for Ezekiel. It’s to harden Ezekiel to match it.

The obstacle stays. You get toughened.

I’ve been thinking about how often I pray for obstacle removal. “God, change this situation. Remove this difficult person. Clear this path.” It’s a reasonable prayer. Sometimes God answers it.

But sometimes the provision isn’t removal. It’s capacity. Not a cleared path, but stronger legs. Not a softer opponent, but a harder forehead.

This changes how I think about difficulty. If I’m constantly asking for easier circumstances, I might be missing the point. The circumstance might be the curriculum. The resistance might be building something I need.

Ray Dalio talks about pain plus reflection equaling progress. He’s onto the same principle from a secular angle: the obstacle isn’t blocking growth, it’s producing it. But Ezekiel adds something Dalio doesn’t. The hardening isn’t just useful, it’s provision. God doesn’t apologize for the difficulty. He equips Ezekiel to meet it.

The Cost of Obedience

Ezekiel’s faithfulness wasn’t just verbal. It was physical.

God told him to lie on his left side for 390 days, then his right side for 40 days (430 days total, bound with cords so he couldn’t turn over). In public. His neighbors watched. He ate rationed food, baked over a fire fueled by dung. He shaved his head and beard with a sword and divided the hair into thirds to symbolize Jerusalem’s destruction.

This wasn’t private devotion. It was public humiliation. The message required his body and his reputation, not just his words.

I keep coming back to this: obedience had a cost that went beyond inconvenience.

Most of the difficult things I avoid aren’t physically demanding. They’re socially costly. Speaking a hard truth might damage a relationship. Taking a clear position might invite criticism. Standing out might mean standing alone.

Ezekiel’s example suggests that’s part of the deal. Faithfulness sometimes costs comfort, dignity, or how others perceive you. The chameleon avoids those scenes. But those scenes are often the ones that matter most.

I wrote once about viewing your life as a movie, “if my life was a movie this would be the best episode of a 10 part series.” The difficult chapters aren’t bugs in the story. They’re the plot. The scenes where you look foolish, where you’re misunderstood, where you pay a cost for doing the right thing: those are the scenes that reveal character.

Ezekiel lying on his side for over a year, looking insane to his neighbors, faithfully embodying a message no one wanted to hear. That’s not a detour from his calling. That is his calling.

The Fatherhood Application

As a dad, I’m a watchman over my family.

This hit differently when I read it in our men’s group. The watchman principle isn’t abstract theology. It’s a parenting framework. My daughter will make choices. She’ll face consequences I can see coming and she can’t. My job is to warn, clearly and repeatedly if necessary.

Her job is to decide what to do with the warning.

I can’t control the outcome. I can’t force obedience. I can’t guarantee she’ll listen. But I can guarantee I won’t stay silent when I see danger.

The temptation is always to soften the message, to preserve the relationship, to avoid being the heavy. “She’ll figure it out.” “I don’t want to push her away.” “She needs to learn on her own.”

Some of that is true. But some of it is chameleon behavior dressed up as wisdom. I’m avoiding the cost of speaking clearly because I don’t want to deal with her reaction.

The watchman framework reframes this. Her reaction isn’t my responsibility. My responsibility is to warn. If I warn and she ignores me, that’s her choice. If I don’t warn because I’m afraid of her reaction, her blood is on my hands.

That’s a heavy phrase. It’s meant to be.

Strategic Silence

There’s a pattern I’ve noticed in myself: I call silence “strategic” when it’s actually avoidance.

“I’m waiting for the right moment.” “I need to think about how to say this.” “The relationship isn’t strong enough yet to handle this conversation.”

Sometimes those are legitimate. But increasingly I’ve started applying the no-regrets framework: project yourself forward, one month, three months, six months, a year. How will you feel about this decision at each point?

One month from now: Will I be glad I waited, or will I wish I’d spoken when I first saw the problem?

Six months from now: Will the “right moment” have arrived, or will I still be waiting?

One year from now: Will I look back on my silence as prudent, or as cowardice wearing a strategic mask?

The answers usually clarify pretty quickly. The “right moment” rarely arrives on its own. It arrives when you create it by speaking.

The Freedom in the Framework

Here’s what surprised me about the watchman principle: it’s actually liberating.

If I’m responsible for outcomes, I carry an impossible weight. I have to find the perfect words, the perfect timing, the perfect approach. I have to anticipate every objection, manage every reaction, engineer every result. The burden is crushing because the variables are infinite.

But if I’m responsible for warning (just warning) the weight shifts. I say what I see. I say it clearly, with appropriate gravity. And then I’m done. What they do with it is their business.

This doesn’t mean I speak carelessly. The watchman isn’t shouting random words from the wall. He’s watching carefully, identifying real threats, and communicating them clearly. But his job ends at communication. He doesn’t have to climb down from the wall and personally defend every house.

I can speak hard truths without carrying the burden of whether those truths are received. I can warn my daughter without feeling like a failure if she doesn’t listen. I can say the thing that needs to be said in a business context without managing everyone’s emotional response to it.

The silence was heavier. Carrying the unsaid thing, wondering if I should say it, feeling the weight of potential consequences. That’s the crushing burden. Speaking, even when it’s hard, is the lighter load.

The Question That Remains

Ezekiel sat in silence for seven days after his vision and commission, “overwhelmed among them” (3:15). Even with a direct call from God, he needed time to process before he spoke.

Maybe that’s the balance. The watchman isn’t impulsive. He watches carefully. He processes what he sees. He considers how to communicate it. But eventually, he speaks.

The question I keep asking myself: What am I sitting on? What truth have I seen that I haven’t spoken? What warning am I strategically delaying that I already know needs to be delivered?

The answers you’re looking for are in the silence you’re avoiding. But sometimes the silence is the problem. Sometimes the answer isn’t in contemplation. It’s in the words you haven’t said yet.

The watchman’s job is to watch and warn. Watching without warning is just spectating. And spectators don’t get to complain about the outcome.

What warning are you sitting on? What “strategic silence” might actually be avoidance? Where do you need a harder forehead instead of an easier path?

The Chinese surveillance balloons floated across American airspace for years before anyone noticed.

Not because we lacked sensors. Not because we lacked data. The U.S. defense system collects more information than any surveillance apparatus in human history. The balloons went undetected because of something far more dangerous than technological failure: we weren’t looking for them.

I heard this story recently on the Jordan B. Peterson podcast from Garry Nolan, a Stanford immunologist who’s become one of the more credible scientific voices in UAP research. He explained how our defense sensors are programmed to filter for known threat signatures (rockets, planes, missiles). Everything else gets discarded immediately as noise. The system processes what it expects and throws away what it doesn’t.

When the UAP research community pushed to “open the filters,” to actually look at what was being dumped, the first discovery wasn’t extraterrestrial. It was hot-air balloons. Technology from the 1780s, floating through American airspace undetected because no one imagined a threat could be that simple.

The sensors weren’t blind. The interpretation was blind.

False Completeness

There’s a particular kind of blindness that’s more dangerous than ignorance.

Ignorance says “I don’t know.” That’s uncomfortable, but it’s honest. It leaves the door open for learning.

The dangerous blindness says “I know what there is to know.” This is false completeness, the assumption that your categories capture reality, that your filters are correctly calibrated, that anything outside your framework is noise worth discarding.

The defense system designers weren’t stupid. They built exactly what they intended: sensors optimized for known threats. The problem was assuming known threats were the only threats. They didn’t build a category for “things we haven’t imagined yet.”

Friedrich Nietzsche saw this clearly over a century ago:

“There is only a perspective seeing, only a perspective ‘knowing’; and the more affects we allow to speak about one thing, the more eyes, different eyes, we can use to observe one thing, the more complete will our ‘concept’ of this thing, our ‘objectivity,’ be.”

Objectivity doesn’t come from having no perspective (that’s impossible). It comes from actively seeking more perspectives, different perspectives, perspectives that challenge your own. The path to truth runs through the multiplication of viewpoints, not their elimination.

The Structural Solution

Abraham Lincoln understood this problem. In Team of Rivals, Doris Kearns Goodwin documents how Lincoln filled his cabinet with political rivals, people who disagreed with him fundamentally. Not because he enjoyed conflict, but because he didn’t trust himself to see every angle.

Lincoln built a structure that would force opposing viewpoints into the room.

This is the difference between hoping you’ll stay open-minded and building systems that guarantee diverse input. Hope fails under pressure. Structure persists.

I try to apply this in my own work. When we form committees or task forces, I often push to include dissenters, people whose perspective might derail the comfortable consensus. The pushback is predictable: “If we include that person, they’ll slow everything down.”

That objection proves the point. If someone’s perspective is so different that it threatens to derail the conversation, that’s precisely the perspective most likely to see what everyone else is missing.

The discomfort is the signal.

Speed vs. Accuracy

Here’s the tension: dissent does slow things down. Including opposing voices creates friction. Opening the filters means processing more data, considering more possibilities, sitting with more uncertainty.

So you have to choose: Are you optimizing for speed or for accuracy?

In 2026, most people optimize for speed. We jump to conclusions that confirm our worldview. We process information that fits our existing categories and discard the rest. Social media algorithms have trained us to expect instant validation, show me content that matches what I already believe, filter out everything that doesn’t.

But the Chinese balloons were hiding in the filtered data. The threats we don’t anticipate live in the noise we dismiss. The opportunities we miss are categorized as irrelevant before we ever see them.

A former executive director of a state auctioneer association used to tell me: “There are three sides to every story: his side, her side, and the truth somewhere in between.”

Most people camp on their side and argue. Fewer try to understand the other side. Almost nobody goes looking for the third thing, the truth that neither perspective fully holds. That search requires slowing down. It requires admitting your view is incomplete.

The Prerequisite Chain

I’ve been thinking about what it actually takes to see clearly. It’s not a single skill. It’s a chain, and every link matters:

1. Self-confidence, secure enough to be wrong without feeling diminished
2. Willingness, wanting the truth more than wanting to be right
3. Slowing down, resisting the pull toward quick conclusions
4. Listening, actually processing the dissenting view
5. Understanding, sitting with the alternative perspective fully, not just hearing it
6. Updating, changing position when the evidence warrants

Break any link and the whole thing fails.

Most people break at step one. It takes genuine confidence to admit you’re wrong. Insecure people double down because their identity is fused with their positions. Being wrong feels like being diminished. So they dig in, dismiss the dissenter, optimize for feeling right rather than being accurate.

Confident people can update. They can say “that’s a better idea” without it costing them anything, because their sense of self isn’t riding on every opinion.

Lincoln could fill his cabinet with rivals because he wasn’t threatened by disagreement. The defense system couldn’t “admit” its filters were wrong until someone from outside forced the question.

The Sports Analogy

There’s a framework I keep coming back to: You’re never as good as your best games, but you’re never as bad as your worst games. Reality lives in the middle.

This applies to self-assessment, but it also applies to how we evaluate everything. Our highs inflate us; our lows deflate us. The data points we remember most vividly (the wins and the disasters) are usually outliers. The truth lives in the middle, in the boring average, in the patterns we don’t notice because they don’t trigger strong emotions.

Opening the filters means being suspicious of your own certainties. The things you’re most sure about might be the things most worth questioning. The data you’re ignoring might contain the signal you need.

Before You Can Apply

I wrote recently about the gap between understanding and doing (Understanding vs. Doing), how we collect principles instead of applying them, how knowledge accumulation becomes a sophisticated form of procrastination.

But there’s a gap that comes before that one. Before you can apply knowledge, you have to be able to see clearly. And before you can see clearly, you have to be willing to be wrong.

This is the prerequisite that most people skip. They assume their perception is accurate and move straight to execution. But if your filters are miscalibrated, if you’re only processing information that confirms what you already believe, then all the execution in the world won’t help. You’ll be efficiently pursuing the wrong targets.

The Mongols understood something about this (What the Mongols Knew). Their religious tolerance wasn’t philosophical enlightenment. It was epistemic humility at scale. They didn’t care what you believed because they understood that beliefs aren’t the same as contributions. They processed the person, not the category the person fit into.

Modern ideological conformity does the opposite. We filter people through their stated positions, discard anyone who doesn’t match our priors, and wonder why we keep being surprised by reality.

The Real Work

Nietzsche said there are no facts, only interpretations. That’s not nihilism. It’s a challenge. If every perspective is partial, then the work isn’t finding the “right” perspective. The work is multiplying perspectives, seeking out the views that challenge your own, sitting with the discomfort of uncertainty.

The defense system needed someone to say “open the filters.” Lincoln needed rivals willing to tell him he was wrong. We all need people and practices that force us outside our comfortable categories.

The answers aren’t in better filters. They’re in fewer filters. The answers aren’t in processing faster. They’re in slowing down enough to question what you’re not processing.

The obvious hides in plain sight, not because it’s invisible, but because we’ve trained ourselves not to see it.

What are you filtering out? What perspective do you dismiss as noise? What truth is hiding in the data you’re discarding?

The largest contiguous empire in human history was built by people we call barbarians.

Twenty-four million square kilometers. More land than Rome, more than Britain at its peak, more than any empire before or since. Built in a single generation by a man who started with nothing, no army, no wealth, no political connections. Just a mother who dug roots to keep him alive and a wife who told him to stop following and start leading.

I listened to Jack Weatherford on the Lex Fridman podcast recently, discussing Genghis Khan and the Mongol Empire. Weatherford has spent decades studying the Mongols, and his work challenges everything I thought I knew about “progress.”

Here’s what’s bothering me: The Mongols operated on principles we consider progressive today. Powerful women in leadership, genuine religious tolerance, ruthless meritocracy. Then we spent centuries dismantling those principles in the name of civilization.

What if the barbarians had it figured out?

Women Built the Empire

This isn’t revisionist history or modern projection. Women weren’t incidental to Mongol success. They were structural.

Hö’elün, Temüjin’s mother, kept her family alive through sheer will when they were abandoned by their tribe. She “pulled her hat down over her head, took her black stick and ran up and down the banks of the river digging out roots to feed the gullet of her brood.” No waiting for rescue. Action.

Börte, his wife, didn’t just support him. She redirected his life. When Temüjin was subordinate to his blood brother Jamukha, she told him directly: “He lords it over you too much. He orders you around too much. You need to be free.” That counsel changed history. Without it, there’s no Genghis Khan.

After his death, his daughters and daughters-in-law ruled vast territories. Not as regents waiting for sons to come of age, as rulers in their own right. Weatherford wrote an entire book on this: The Secret History of the Mongol Queens.

The largest empire ever made space for women that our “advanced” civilization later closed off.

Why?

The Mongol system was decentralized, merit-based, and tribal. If you could do the job, you did the job. Western “progress” brought nation-states, codified law, religious institutions, primogeniture, clerical celibacy, military service as the price of citizenship. Each of these centralized power in ways that excluded women by design.

We institutionalized. They never did. Their “primitive” structure was actually more open.

Modern politics has become an elite ruling class playing musical chairs. The Mongol model was living with, as, and of the people. Somehow that’s the version we call barbaric.

Tolerance That Actually Tolerated

The Mongols had no state religion to protect. They didn’t care what you believed (Buddhist, Christian, Muslim, shamanist) as long as you were loyal and productive.

This wasn’t philosophical enlightenment. It was pragmatism. Conquered peoples who could keep their gods were easier to govern than those forced to convert. The Mongols understood that religious coercion breeds resentment, and resentment breeds rebellion.

But here’s what strikes me: their tolerance was actual tolerance. Indifference to belief, focus on contribution.

Modern Western “tolerance” is something else entirely. It demands affirmation. You must not only accept, you must celebrate. And if you don’t affirm me, I won’t just disagree with you. I’ll work to make you less of a citizen. I’ll remove you from polite society. I’ll destroy your livelihood if I can.

That’s not tolerance. That’s ideological conformity wearing tolerance as a mask.

The Mongol version: Believe what you want. Be loyal. Be productive. We don’t care about your inner life, we care about your contribution.

The modern version: Your inner life is exactly what we care about. Affirm the right things or face consequences.

One of these is tolerance. One is coercion dressed in progressive language.

The State Religion Problem

There’s a reason Mongol tolerance worked that we can’t easily replicate: they had no sacred center to protect.

Western civilization built itself around Christianity as an organizing principle. Not just as belief, but as the framework for law, morality, social structure, and meaning. When you have a sacred center, tolerance means decentering something foundational.

The Mongols faced no such tension. They could afford indifference because nothing was threatened by other beliefs coexisting.

In the United States today, Christianity is being decentered. Whether you think that’s good or bad, the destabilization is real. The people who organized their lives around that center feel it slipping. The people pushing the decentering feel the resistance. Neither side can afford Mongol-style indifference because the stakes feel existential.

Maybe genuine tolerance requires having nothing sacred to protect. Or maybe it requires everything being equally sacred, which amounts to the same thing.

Either way, we’re not there. And pretending we are makes the conflict worse.

Brutality and Clarity

I can’t write about the Mongols without addressing this: they were devastating in conquest.

Entire cities massacred. Terror used strategically. Jamukha, Temüjin’s blood brother turned rival, captured the rhetoric perfectly: “Jump to the smoke hole in the top of the ger… kill them all. Kill the men and the women and the children.”

This wasn’t incidental. It was policy. Resist, and annihilation follows. Surrender, and you’re absorbed peacefully.

The brutality is indefensible by modern ethics. I’m not defending it.

But here’s what I can’t shake: within the empire, there was remarkable order, fairness, and meritocracy. The same system that destroyed resisters protected subjects. The same leaders who showed no mercy in conquest showed genuine care for governance.

This wasn’t contradiction. It was clarity. Total loyalty earns total protection. Resistance earns annihilation. Everyone knew exactly where they stood.

Modern Western ethics rejects this duality. We want consistency across all contexts. We want rules that apply the same way to everyone, everywhere, always.

But maybe there’s something to learn from the clarity, even if the methods are off the table.

How many organizations fail because expectations shift constantly? How many relationships corrode because the rules keep changing? How much anxiety comes from never knowing where you stand?

The Mongols were brutal, but no one was confused. There’s a strange comfort in that, knowing the game, even if the stakes are high.

Why We Can’t Go Back

Here’s the uncomfortable question: Could a Mongol-style system work today?

I don’t think so. Not because meritocracy is wrong, but because we’ve lost the capacity for it.

Western culture, especially in the United States, has drifted into participation awards for everyone. We don’t have the intestinal fortitude for true merit-based accountability. You see it now: as the country moves toward merit-based systems in some areas, there’s tremendous pushback from those who benefit from the current credentialed elite structure.

Merit-based accountability means some people fail. Visibly. Definitively. And we’ve built a culture that can’t tolerate visible failure, not because we’re kind, but because we’re afraid it might be us next.

The Mongols promoted based on demonstrated ability, not seniority or connections. That’s countercultural today. Saying it out loud feels almost transgressive.

But I believe it. In my own business, I’ve seen how merit-based thinking makes organizations better. Loyalty and productivity as the entry ticket, not bloodline, credentials, or ideological conformity. Clear expectations. “Giddy up, you’re part of the group.”

That’s not barbarism. That’s clarity.

The Real Question

History rhymes, but we write the next verse.

The Mongols built something remarkable on principles we’ve abandoned. Not because those principles failed (the empire lasted, in various forms, for centuries). We abandoned them because different power structures emerged. The Church, feudal hierarchies, nation-states, each centralized power in new ways, and those ways excluded people the Mongol system included.

The regression wasn’t going backward. It was going somewhere else. Somewhere that served different interests.

So the question isn’t whether we can return to Mongol principles. It’s whether we can learn from them.

Can we create space for merit that our institutions currently close off?

Can we practice tolerance that doesn’t demand affirmation?

Can we establish clarity that doesn’t require brutality?

Can we build organizations where contribution matters more than credentials?

I don’t know. But I suspect the answers are in the silence after we stop pretending our current system is the pinnacle of progress.

The barbarians knew something we’ve forgotten. Maybe it’s time to remember.

What would change if you judged people by contribution instead of credentials? What would you have to give up?

I’ve been noticing something uncomfortable about my own thinking lately. I collect principles like they’re going to save me. Framework after framework, insight after insight, all carefully documented with proper attribution and cross-references. Ray Dalio on historical cycles. Charlie Munger on market selection. Larry Page on 10x thinking. Jimmy Carr on boredom and insight.

The collection grows. The application lags.

This isn’t writer’s block or procrastination in the traditional sense. It’s something more insidious: epistemic displacement. I’m replacing the hard work of doing with the comfortable work of knowing.

The Seduction of Understanding

There’s a particular satisfaction that comes from understanding something. Your brain releases a little hit of dopamine when a concept clicks into place. You feel smarter, more prepared, better equipped. The feeling is real. The preparedness is often illusory.

Consider how I’ve been thinking about market value in used equipment sales. I can articulate the problem clearly now: if marketing saturation rarely reaches double digits, we’re not discovering true market value, just a realized price from a limited buyer pool. I understand the mechanics. I can explain why different auction companies see different prices for identical assets.

Understanding the problem is not the same as solving it.

I’ve documented the solution framework too: build communities instead of renting attention, find your own lake rather than competing as a better fisherman, apply 10x thinking to reach exponentially more qualified buyers. These aren’t just ideas I’ve read. They’re principles I genuinely believe would work.

So why haven’t I fully executed on them?

The Valley Between Knowing and Doing

There’s a particular kind of work that exists in the space between understanding a principle and actually implementing it. This work is:

• Unglamorous (no dopamine hit for incremental progress)
• Uncertain (understanding doesn’t guarantee execution)
• Exposed (you have to publicly attempt something that might fail)
• Tedious (the “mundane repetitive bullshit tasks” I claimed I wanted machines to handle)

I wrote down that “laziness is a deep indicator from within you that you’re not doing something right with your own life.” Fair enough. But what about the opposite? What about industrious note-taking that feels productive but is actually avoidance behavior dressed up as preparation?

Jeremy Giffon said it, and I dutifully recorded it: “The magic you’re looking for is in the work you’re avoiding.”

The work I’m avoiding is not learning more. It’s not developing better frameworks. It’s not gaining deeper understanding. The work I’m avoiding is the messy, imperfect application of what I already understand.

The Temporal Trick

Here’s what makes this pattern so persistent: collecting knowledge does have value. Understanding principles is useful. The problem is when it becomes a substitute for action rather than a foundation for it.

I have a decision-making framework I learned from my dad: project yourself forward one month, three months, six months, a year. How will you feel about this decision at each point?

Let me apply it to the knowledge-action gap:

One month from now: Will I wish I’d read another framework about market disruption, or will I wish I’d actually tested a new community-building approach?

Six months from now: Will I be proud of my comprehensive notes on business strategy, or will I be proud of measurable progress toward creating my own market?

One year from now: What will matter more, that I understood the theory perfectly, or that I implemented it imperfectly but learned from reality?

The answers are obvious when you force the temporal perspective. Yet in the present moment, acquiring knowledge feels like progress.

Boredom as Signal

Jimmy Carr said something that hit differently when I heard it: “Boredom is an insight creator. The answers you’re looking for are in the silence you’re avoiding.”

I think he’s right, but not in the way most people interpret it. The silence isn’t meditation or stillness. It’s the absence of input. It’s sitting with what you already know and forcing yourself to do something with it instead of seeking more.

We treat boredom like a problem to solve with more stimulation. What if it’s actually your mind telling you it has enough raw material and needs to start building?

The Practice

So here’s what I’m testing: treating knowledge collection like raw materials in a warehouse. Useful, but only if you’re actually manufacturing something with them.

Before I save another principle to my notes, I’m asking:

• Do I already have a principle that addresses this?
• Have I actually applied the principles I already have?
• What’s one concrete action I could take based on what I already know?

It’s not about stopping learning. It’s about establishing a ratio: for every X units of input, Y units of output. Right now my ratio is badly skewed toward input.

The hardest part is accepting that application will be messy. My notes are clean, well-organized, properly attributed. Reality is not. Implementation means dealing with incomplete information, imperfect execution, and visible failure along the way.

But that’s where the actual learning happens. Not in understanding the principle, but in discovering how it bends under real-world pressure.

What This Means Practically

I keep coming back to the “10x thinking” principle. Larry Page’s insight was that trying for 10x better forces you out of incremental optimization into fundamental rethinking.

Maybe I need to apply that frame to my own knowledge-action gap:

• 10% thinking: “I should apply these principles more consistently.”
• 10x thinking: “What would change if I stopped collecting new principles entirely until I’d actually executed on what I already know?”

That second question is uncomfortable. It suggests the answer isn’t more input. It’s disciplined output.

Jordan Peterson says that when you set a goal, your perception reorganizes around it. The world arranges itself as a pathway to that goal. But that only works if the goal is doing something, not understanding something.

The Real Work

History rhymes, as Ray Dalio taught me. People have been confusing understanding with progress forever. Scholars in every era spend lifetimes cataloging knowledge while practitioners stumble forward with incomplete information and learn by doing.

The real insight isn’t about whether knowledge is valuable (of course it is). It’s about recognizing when knowledge accumulation becomes a sophisticated form of procrastination.

Steve Ballmer said teamwork is an implementation detail, not a goal. Maybe understanding is the same: a necessary foundation, but not the destination.

The work isn’t acquiring one more framework. The work is taking what I already understand and building something real with it, even if that something is imperfect, incomplete, and exposes my limitations.

The answers I’m looking for aren’t in the next principle I collect. They’re in the silence after I close my notes and ask: “So what am I actually going to do about it?”

That’s the uncomfortable work. That’s where the magic is.

What principles are you collecting instead of applying? What work are you avoiding by seeking more understanding?

History doesn’t repeat itself. It rhymes.

That’s not a cute saying. It’s the difference between someone who files a situation under “seen it” and moves on, versus someone who does the actual work of figuring out what’s the same, what’s different, and why it matters this time.

Most people stop at the first version. It feels smart. It’s not.

Where This Started

A colleague said something during early COVID that stuck with me: “History repeats itself.” Simple observation, and at the time I nodded along. But as I sat with it, something felt off. The details never line up perfectly. The mechanics shift. The players change.

Then I started reading Ray Dalio and listening to James Holland, and the nuance clicked.

Holland’s work on 1930s German propaganda is the one that keeps me up at night. The techniques from that era show up in today’s information warfare. Both periods grew out of economic crisis (the Depression then, 2008 and COVID now). The playbook is the same: fear, othering, oversimplification. The delivery system changed. Radio became algorithmic feeds. But the human vulnerabilities being exploited? Identical.

Dalio’s contribution is the financial angle. The 2008 crash rhymed with 1929. Both were excessive leverage unwinding. Both triggered by the same psychology: greed on the way up, panic on the way down. But the policy responses were different because (sometimes) people actually learned from the last time around.

The pattern exists. The details don’t copy-paste.

The Real Work

Here’s what makes this useful instead of just interesting. When you spot a historical rhyme, you’ve got to ask five questions:

1. What’s the pattern? Which historical parallel are you actually seeing?
2. What’s genuinely similar? Same human psychology? Same structural conditions?
3. What’s fundamentally different? New technology, new regulations, different global context?
4. What created the original pattern? Not just what happened, but why it happened.
5. How have those causes evolved? The trigger from 1929 doesn’t fire the same way in 2008.

Skip any of those and you’re just doing trivia. “This reminds me of the Depression” isn’t analysis. It’s a cocktail party observation.

The trap is thinking that spotting the rhyme tells you how the next verse goes. Markets crash in rhyming patterns, sure. But each crash has its own speed, its own triggers, its own recovery path. Companies fail in familiar ways (ignoring disruption, clinging to what worked, misreading the market), but each failure teaches you something new about what to watch for next time.

Why We Get This Wrong

The lazy version of pattern recognition feels productive. You see a similarity, you categorize it, you feel like you understand what’s happening. That little hit of “I’ve seen this movie before” is satisfying. And it’s almost always incomplete.

The hard version requires you to hold two things at once: this situation rhymes with that one, AND this situation is different in ways that might matter more than the similarities.

That’s uncomfortable. Your brain wants a clean match. Reality gives you a messy one.

Greed, fear, herd behavior, the cycle of innovation and disruption. These are the recurring melodies. They’re baked into how people operate and they’re not going anywhere. But technology, regulation, and how connected the world is at any given moment? Those change the key, the tempo, the instruments. Same song. Different arrangement. And the arrangement matters.

What To Do With This

The people who actually benefit from historical thinking aren’t the ones who memorize dates and draw parallels. They’re the ones who understand why certain patterns keep showing up, and then pay close attention to where this version breaks from the last one.

They don’t dismiss history. They don’t worship it either. They use it as a starting point and then do the work of figuring out what’s genuinely new.

“We’ve seen this before” is comfortable. It’s also the exact moment you stop learning. The harder question, the one worth asking, is: “We’ve seen something like this before. What’s different this time, and does the difference change everything?”

History rhymes. But we write the next verse.

When’s the last time you caught yourself saying “we’ve seen this before” and stopped there? What did you miss by not asking what was different?

Let’s talk about something every auctioneer thinks about: Are we actually finding true market value?

I had an interesting conversation recently with a colleague about whether auctions reveal what something’s really worth. Sure, they can. But here’s what bothers me. We’re calling it “true market value” when maybe 5–10% of potential buyers ever see our marketing. How’s that the true value?

This isn’t just theory. It creates real problems.

Take used equipment sales. The same machine can bring wildly different prices depending on who’s selling it. A specialized dealer with an established buyer list? They’ve got a huge advantage. Their marketing hits the target every time. They know exactly where their buyers are. But a first-time seller? They’re throwing everything at the wall hoping something sticks. The result: massive price variations for the same piece of equipment.

The obvious answer is to spend more on advertising. But that’s just throwing money at the problem, not solving it.

Smarter, Not More

We need smarter marketing, not more marketing. This means understanding where our buyers actually spend their time and meeting them there. Good strategy can definitely improve our reach. But even the best traditional approach hits a ceiling, and we run into budget limits and timing problems that cap what’s possible.

So if small improvements aren’t enough, we need to think differently.

There’s this concept from Google’s Larry Page about “10x Thinking.” Instead of trying for 10% better, aim for ten times better. Applied to our industry, we stop asking “How can we reach a few more buyers?” and start asking:

“How could we reach 10 times the right buyers for the same cost?”

“What would narrow that value gap by 10 times?”

These questions force you out of the standard playbook. For Grafe, this has meant moving away from paid advertising toward building an organic community. We’re creating an audience that trusts us first and buys second, rather than constantly renting attention in the marketplace.

Fish Where the Fish Are

Charlie Munger had it right: “Fish where the fish are.”

Maybe the answer isn’t being a slightly better fisherman in the same overfished pond. Maybe it’s finding your own lake. Build a dedicated community in your niche, and you create your own market. True market value becomes less of a ghost to chase because you’ve built a place where value actually gets found.

The bottom line: We might never achieve perfect market saturation. But by building trusted communities and targeting the right buyers intelligently, we can get a lot closer to real value discovery than the current spray-and-pray approach that’s easy to fall into.

What market are you fishing in? Are you competing as a better fisherman, or building your own lake?

Source: Jimmy Carr in conversation with Chris Williamson on the Modern Wisdom podcast.

What am I avoiding? Silence, work, something else? Or do I need to seek rather than avoid, and does seeking actually enable me to avoid?

We treat boredom like a bug. Something to fix with another podcast, another scroll, another notification. But boredom might be your mind telling you it has enough raw material and needs to start building. The discomfort isn’t the problem. The discomfort is the signal.

The silence isn’t meditation or stillness. It’s the absence of input. It’s sitting with what you already know and forcing yourself to do something with it instead of seeking more.

This idea threads through Understanding vs. Doing and The Watchman Principle.

Whenever you get cancelled. Dissed. Attacked.

If my life was a movie, this would be the best episode of a 10-part series.

Source: Jimmy Carr in conversation with Chris Williamson on the Modern Wisdom podcast.

You truly need the downs. They create the full story. Which makes me think…

If you’re happy with you today, you’ve got to be happy with everything that’s happened to this point. Because this is what makes me today, me.

Nobody’s highlight reel includes the scenes that actually built them. But those are the scenes that matter. The cancellation, the public failure, the moment everyone turned on you. Those aren’t interruptions to your story. They are the story. They’re the scenes people would actually watch.

The comfortable seasons make terrible television. The valleys make great ones.

This thought became the seed for The Valley and the Namer.

Source: Jeremy Giffon on the Invest Like the Best podcast.

This one cuts. Because it reframes laziness from a character flaw into a diagnostic tool.

If you can’t get yourself to do the work, maybe the work is wrong. Not “wrong” as in unimportant, but wrong for you, wrong for this season, or wrong in a way you haven’t been honest about yet.

There’s also a sneakier version of laziness that doesn’t look lazy at all. It looks diplomatic. Strategic. Wise. The decision to blend in rather than stand out. The choice to stay silent when you should speak. The comfortable act of collecting more knowledge instead of applying what you already have.

That’s chameleon behavior dressed up as prudence. And it’s just as much a signal of misalignment as the person who can’t get off the couch.

This idea shows up in The Watchman Principle (chameleon behavior) and Understanding vs. Doing (industrious avoidance).

I think I first heard this from Dad and have since heard it repeated in various forms throughout my life.

It’s a simple framework, but it cuts through the noise fast. Most bad decisions feel obvious in hindsight because we didn’t bother to simulate the hindsight before we acted. The regret was predictable. We just didn’t slow down enough to predict it.

The power isn’t in the time horizons themselves. It’s in the forced pause. You’re making your future self a stakeholder in today’s decision. That future self has context your present self doesn’t: the emotional dust has settled, the short-term pressure is gone, and the consequences have played out.

The framework also exposes decisions that feel uncomfortable now but won’t generate regret later. Sometimes the right call is the one that stings today but sits well at six months. If future-you is at peace with it, present-you can handle the discomfort.