What Founders Get Wrong About Habits

By Momar Lissa Ndiaye ("MLN"), Founder & CEO, weyoga Inc.

I write this one from inside the chair. I have spent years in markets building systematic strategies — machines for making decisions repeatable, auditable, and pattern-aware — and then watched myself, and nearly every founder and allocator I respect, run the self that operates those machines with no system at all. This essay is the Stakes movement's operator chapter: what the people who decide for a living get wrong about their own patterns, and what the framework of this series says to do about it.

Start with the word in the title, because the mistake begins there. Ask a founder about "habits" and you will hear about mornings: the workout, the routine, the deep-work blocks, the phone in the other room. An entire literature has trained operators to think of habits as behaviors of the body and calendar — things you install, stack, and optimize. All fine, and all beside the point. The habits that determine a company's trajectory are not behavioral in that sense at all. They are decisional: the founder's recurring way of framing options, weighing evidence, and choosing — Essay 6's machinery, running at the top of the org chart, unexamined. The founder who hires the same dazzling wrong profile three times does not have a routine problem. The founder who re-enters the same commitment spiral every spring cannot cold-plunge their way out of it. The habit literature optimizes the operator's inputs — sleep, focus, energy — and leaves the operator's decision function exactly as it found it. Founders instrument everything except the instrument making the decisions.

The irony, once seen, is almost comic, and Essay 8 named its general form: every company worth admiring has better memory infrastructure than its founder. But the operator's version is sharper, because founders built that infrastructure on purpose. You maintain dashboards on revenue, funnels, burn, latency. You run post-mortems on incidents and retros on sprints. You keep an investment memo so the fund's reasoning is auditable later — because you know, professionally, that un-audited reasoning drifts and repeats. And the single highest-leverage decision system in the company — the one that chose the market, the hires, the term sheet, the pace — has no dashboard, no ledger, no retro, no memo. Its decisions ship straight to production with no review and no record. If any team in your company operated the way you operate, you would fire the manager. The founder's decision function is the last unobserved system in an otherwise fully instrumented company.

The costs are not hypothetical, and every operator reading this can supply their own instances; the pattern tax (Essay 8) simply compounds fastest at the top, because the top's decisions have the most leverage and the least oversight. Three species recur so reliably they deserve names. The repeated hire — the same wrong profile, differently dressed, because the founder's evidence selection (Essay 6) collects charisma the same way every time, and no record exists in which the resemblance could be seen. The cycle mistimed — the founder's own eighteen-month arc of conviction and depletion, projected onto the company as strategy: expansion decided at the peak, retrenchment at the trough, both experienced as clear-eyed reads of the market. And the board-room precedent — the allocator whose "case-by-case" decisions rhyme across a decade, visible to every LP who reads the full sequence and invisible to the person producing it, one perfectly reasoned case at a time. In each species, note what did not fail: intelligence, information, effort. What failed was the absence of the check — does today's entry match a known pattern? — that the founder's own company performs on every deploy.

Now the steelman, from the toughest voice in the room: this is what boards, co-founders, and executive coaches are for — the external observer already exists at the top. Partly true, and worth pricing honestly with Essay 12's economics. The board sees quarterly snapshots, not the sequence; it audits outcomes, not formation, and it arrives structurally in the post-mortem tense. The co-founder is inside most of the same patterns, and often selected — by the founder's own framing machinery — for compatibility with them. The coach is the real thing, the artisanal recognition of Essay 9 — and is therefore rationed exactly as Essay 12 described: hours a month, capitalizing for the first year, non-portable, and priced so that the operators who most need the function at seed can least afford it. The observer exists at the top the way the good therapist exists in the population: genuinely, occasionally, and by luck and wealth. That is not an argument against the humans. It is the operator's case for the infrastructure — the behavioral ledger (Essay 11) as the founder's missing internal system: decisions logged with their stated reasons, outcomes attached, recurrence checked, the one line surfacing in the formation window — this hire resembles the last two; the stated reason is the same — before the offer goes out, not in the post-mortem after the exit interview.

And because this essay is written for people who allocate capital as well as deploy it, close the loop that Essay 5 opened. Investors underwrite founders' pattern quality constantly — it is half of what "founder judgment" means in a memo — but they underwrite it from the outside, inferring the sequence from references and track record. The founder who keeps the sequence — who can show, not claim, that their decision function is instrumented, that the third hire was checked against the first two, that the spring commitment passed a recurrence review — is offering diligence-grade evidence of the one asset every memo gestures at and none can verify. Self-recognition, made legible, is underwritable. That sentence will sound strange for about five more years, so spell out what it would mean in practice, because it is genuinely new territory. Today, founder judgment is underwritten through proxies — references, track record, the performance of pattern-awareness in a pitch. A maintained decision ledger changes the evidentiary class: the sequence itself becomes producible, the way audited financials once converted "trust me" into "verify me" and created modern capital markets in the process. One can imagine the descendants — diligence that reviews decision infrastructure alongside data rooms; terms that price a founder's recurrence-checked judgment the way they price clean books; funds whose edge is underwriting the operator's pattern quality directly rather than inferring it. Whether any of this arrives in five years or fifteen, the direction is set by the same logic that has driven every prior expansion of what capital can see: markets reprice whatever becomes legible. The founder's decision function is the largest asset in early-stage investing that has never been legible. It is about to be.

What founders get wrong about habits, then, is the altitude. The habit is not the morning. The habit is the choosing — and it is the only system in the building still running unlogged, unreviewed, and unrecognized, at maximum leverage, by an operator who instrumented everything else precisely because they knew better. The tooling this series describes was, in a sense, designed for everyone. But it was designed first for the person in that chair — because no one alive pays the pattern tax at a higher rate, and no one is better positioned to understand, professionally, why the ledger works.

One essay remains. It steps back from every audience — user, researcher, operator — to the widest claim of all: what this framework says the future of AI actually is, and why the name on it will not be the one everyone expects.

Part of The Recognition Layer →

Momar Lissa Ndiaye ("MLN") is the Founder & CEO of weyoga Inc., a Delaware company. — weyoga.ai · mln@weyoga.ai