Grifting is authority you haven't earned

Context

Symptoms

The tell isn’t what someone claims. It’s that the confidence doesn’t move.

Real knowledge is uneven - someone who has actually built things is certain about the parts they burned a weekend on and vague about the parts they’ve only read (you can watch the confidence change as the conversation crosses that terrain). When a person is equally sure about the system they ran for three years and the one they saw a conference talk on last month, the confidence has stopped tracking the knowledge and started tracking something else.

The other signs follow from that. Answers arrive complete and immediately, with no “let me check”. Proximity gets narrated as ownership, so having been in the room becomes having made the call, and having been on the team becomes having built the thing. And the detail thins out on the second question rather than the first, because a good summary survives one question from anybody.

What isn’t a symptom: not knowing something, saying so, being new, being wrong in public, or holding a position you can’t yet fully defend and saying that’s where you are. Nor is ambition. Committing to something hard that you might not pull off isn’t overclaiming (it’s how anything difficult ever starts) and I’d far rather work with someone who reaches too far and misses than someone who only ever promises what’s already in the bag. All of that is the work.

The unstated claim

The sharpest version of this contains nothing you could point at and call a lie.

Nobody stands up and announces “I am an expert in distributed consensus”. They just talk about it the way somebody who has implemented it talks about it, and everyone updates accordingly. Go back afterwards and every individual sentence is defensible. The claim was never made - it was left lying around where people would pick it up, and technically nobody put it there!

Which is why “did they actually lie?” is the wrong test, although it’s the test most of us reach for first. Authority doesn’t travel in claims, it travels in register: how much you hedge, what you volunteer, whether you say “I think”, and what you do with the question you can’t answer. You can mislead a room comprehensively without saying one untrue thing, and the room will go away and make its decisions on the impression rather than on the sentences.

The engine

I’ve watched this closely in enough people, at enough companies, to be reasonably confident about the shape of it (although this is one person’s observations, not research). What’s usually underneath is an insecurity, a chip on the shoulder, a felt need to be more than you currently are. The performance is downstream of that. Almost nobody sets out to defraud a standup! They set out to not be found out, and the overclaiming is what that looks like from outside.

Which raises a question I haven’t settled: does framing it as insecurity rather than dishonesty let people off too easily, or is it just accurate? My feeling is that it’s accurate, because the alternative framing (that these are simply bad people) predicts behaviour badly and suggests no remedy at all. But I’m aware that “it comes from a wound” is an extremely comfortable thing to believe about something that costs other people real money and real weekends, and I don’t fully trust myself on it.

There’s a second engine, and it involves no performance at all. Some people just don’t know how much they don’t know. That’s the Dunning–Kruger effectDunning–Kruger effectThe finding that people with low ability in a domain tend to overestimate it, because the judgement needed to rate your own competence is the same judgement you currently lack. It describes a miscalibration the person is genuinely unaware of, which makes it the harder version to fix rather than the more forgivable one.en.wikipedia.org ↗, and the mechanism is nastier than it first sounds: the judgement you’d need in order to rate your own competence in a domain is the same judgement that domain is still busy teaching you, so the less of it you have, the better your position looks from inside; the error hides behind the thing that would reveal it. Nobody in that state is grifting - they’re reporting honestly on a picture that happens to be wrong.

I don’t think that gets anybody off the hook though, because the cost doesn’t care. A decision made on overstated depth goes the same way whether the depth was overstated on purpose or by accident, and the accidental one is arguably the harder problem (the person can’t see it from where they’re standing). That’s what makes “know what you don’t know” a skill you develop rather than a virtue you either have or you don’t.

The cost to a team

The reason this belongs in a field note and not only in a diary: engineering runs on people reporting their own confidence accurately.

Every technical judgement is a self-assessment being passed upward as an input. “Yes, that’ll hold at ten times the load.” “No, that failure mode can’t happen here.” “This is a two week problem.” When somebody conveys ninety percent and the real number is forty, they don’t only get their own work wrong, they corrupt every decision made on top of it by people who had no way to discount it. The damage turns up months later in a design chosen on a capability that never existed, and by then it’s nobody’s fault in particular (which is the worst kind of failure to have).

It also drags on everybody honest. Calibration is a public good - it only pays if it’s the norm. In a room where the confident answer wins, the person who says “I haven’t done this, here’s my best guess and here’s what would change my mind” is the one who looks weak, and the incentive quietly points the wrong way for everyone.

That’s why the hard line in ZFN-29Field Note · currentZFN-29 — Blameless culture, taken seriously — and its one hard lineWhen something breaks, support the person, don't blame. Run post-mortems with ceremony and learn at every level — software, org, culture, process, even solo. Blameless protects honest mistakes, not dishonesty: evading or blaming gets coached; hiding evidence is a firing offense.Why it's cited here: Where the line actually falls. Blameless culture protects honest mistakes and stops at dishonesty, and deliberately overstating what you know is that dishonesty before the incident rather than after it.Open ZFN-29 → sits where it does. Blameless culture protects honest mistakes and stops at dishonesty, and deliberately overstating what you know is that same dishonesty committed before the incident rather than during the post-mortem. The unwitting version isn’t dishonest at all, although it lands on the same people in the same way (which is the whole reason the standard has to be wider than the accusation).

The same pattern outside tech

None of this is a technology problem - it’s the same mechanism as people living in houses they can’t afford and driving cars they can’t afford: a projection of success that isn’t there, maintained at real cost, because the alternative is being seen at your actual size. The resume and the leased car are the same move with different props.

Utah

Living in Utah is what made me want to write this down. There was intense cultural pressure to appear perfect, and underneath it there were grifters everywhere. Multi-level marketing schemes, products that didn’t exist, people selling things that were never going to be delivered. The performance of perfection and the grift seemed to feed each other, and being surrounded by both at once made the relationship between them impossible to ignore.

I don’t know which way the causation runs, and I’m not going to pretend otherwise in a note about exactly that. Does a culture demanding the appearance of success actually produce more grifters, by making the gap between how you’re doing and how you must appear to be doing too wide to admit? Or does it just hide the ones it already had, by making the performance normal enough that nobody looks twice? Those are different problems with different fixes and I can’t currently tell them apart. It’s the part of this I most want to dig into.

Recommendation

Attach the provenance to the claim. Not the hedging ritual of “I could be wrong” bolted onto the end of a confident paragraph, but the actual source of the belief, up front. “I ran this in production for two years.” “I’ve read the spec but never implemented it.” “This is a guess from the shape of the problem.” Those are wildly different statements and they shouldn’t sound alike; it costs you one clause, and it tells everyone downstream how much weight the sentence will take.

Learn where your own edge is, and be humble past it. This one applies whether or not you’re doing any of this on purpose, and it’s a skill rather than a virtue (which means you can practise it). Go one question deeper than is comfortable and notice the point where you start reconstructing rather than remembering. Notice the topics where you’ve only ever heard yourself talk, and never had to make the thing actually work. Ask the person who does know, and pay attention to how much of their answer you couldn’t have produced. What you’re looking for is the specific discomfort of finding a floor where you assumed there was more room, and the useful response to it is curiosity rather than embarrassment.

Stretch the goal. Promise the ambitious outcome if you want to. Take the project you’re not certain you can land, commit to the hard date, reach well past what you’ve done before. What you don’t get to stretch is how well you understand the ground you’re standing on, because that’s the part everybody else is navigating by. “I’ll find a way to make this work” and “I’ve done this before” are completely different sentences and only one of them is a claim about the past. That is exactly the distinction I got wrong at nineteen: what went on that resume wasn’t ambition about what I’d deliver, it was fiction about what I’d already done.

Treat “I don’t know” as a complete sentence. It’s cheap on a Monday and ruinous in an incident channel six weeks later. Following it with “here’s how I’d find out” is better, although the finding-out is a separate offer and not the price of admission for having admitted the gap.

Separate what you built from what you were near. Proximity isn’t authorship. Having been on the team is worth saying - it just isn’t worth saying in the words you’d use for having designed it.

Judge by age and stage. A junior overstating themselves in an interview is a person under pressure with no track record to stand on. A principal doing it is making a choice with alternatives available. Same act, genuinely different thing, and the response should differ too: the first is worth coaching and the second is worth taking seriously.

Keep the word narrow and the standard wide. This is the part I’d get wrong first if I weren’t careful, and it’s the same failure ZFN-27Field Note · currentZFN-27 — Don't tolerate assholes — but be strict about what one isDon't tolerate assholes — people who demean, belittle, punch down. But filter hard on the word: disagreeing, raising ideas, or opening a competing PR isn't being an asshole, it's the work. Assholes attack people; colleagues attack problems. Don't let the label silence dissent.Why it's cited here: The same discipline applied to a different label. There the argument is that 'asshole' has to stay narrow or it silences dissent; here it is that 'grifter' has to stay narrow or it punishes people for being new.Open ZFN-27 → warns about with “asshole”, where a useful word for a real behaviour turns into a weapon the moment it gets applied loosely. Someone inexperienced isn’t a grifter. Someone confidently wrong isn’t a grifter - that’s just being wrong, which everybody who ships is, regularly. Someone reaching for a hard goal isn’t a grifter. And someone who genuinely can’t see the edge of their own knowledge isn’t one either.

That last one is where the word and the standard come apart. “Grifter” is an accusation of intent, so keep it for the cases where there’s intent, and remember you usually can’t see intent from outside anyway (which is a good reason to be slow with the label and quick with a direct question instead). The standard doesn’t move with the label though. The obligation to know what you don’t know lands on the person who isn’t doing it on purpose exactly as hard, and telling them is a kindness rather than an accusation. Don’t use “well, they don’t realise” as a reason to leave it alone.

Impostor syndromeImpostor syndromeThe experience of doubting your own competence and feeling like a fraud despite clear evidence of achievement. It is the same instrument miscalibrated in the opposite direction, authority genuinely earned but never felt, and it distorts a team's picture of who knows what just as effectively.en.wikipedia.org ↗ is the same instrument miscalibrated in the other direction: authority genuinely earned but never felt (people hedging claims they’ve comfortably paid for). It distorts the team’s picture of who knows what just as effectively as the Dunning–Kruger case does, which is the tell that what we’re actually after here is calibration rather than confidence. The goal was never for people to claim less. It’s for the claim and the knowledge to be the same size.

Underneath all of it is something closer to a discipline than a trait. Feynman’s name for it in Cargo Cult ScienceCargo Cult Science (Richard P. Feynman, Caltech commencement address, 1974)Feynman's address on scientific integrity, which he describes as a kind of utter honesty, a leaning over backwards to report everything that might make you wrong rather than only what supports you. Its most quoted line is the whole discipline in one go: you must not fool yourself, and you are the easiest person to fool.calteches.library.caltech.edu ↗ is a kind of utter honesty, a leaning over backwards, reporting the things that might make you wrong rather than only the ones that make you look right. His line is the whole note in a sentence: you must not fool yourself, and you are the easiest person to fool. Note that it’s aimed squarely at the case where nobody is lying to anybody. The performed version does end up there eventually (a story you told other people until you believed it) but plenty of people start there, having never told anyone anything false in their lives. Either way there’s nobody left checking, which is why this has to be a discipline you run on yourself.

I have this on my about page as one of the lines I try to hold to: I speak the truth, even when it costs. This note is mostly that line applied to the case where the truth is about me: what I actually know, what I’ve actually done, and how far past either of them I’m tempted to talk. Choosing to be accurately who you are rather than who you wish you were turns out to be a daily practice and not a decision you make once.

Consequences

Easier: Technical judgements start to mean something, because they carry a known confidence. Being wrong gets cheaper, since a guess you flagged as a guess missing is information rather than a failure. Juniors will say “I don’t know” in a room where seniors say it first. And you find out what your team actually knows, which is the input every other decision quietly depends on.

Harder: You’ll lose some rooms. There are hiring processes, sales calls and promotion committees that reward the performance directly, and calibrated honesty reads as weakness in every one of them. That’s a real cost and I’m not going to pretend otherwise. ZFN-41Field Note · currentZFN-41 — Make the technical choice you can put your name behindWhen two options would both work, it's fine to choose the one your experience lets you put your name behind, and to resist pressure that isn't about the merits — vendor incentives, politics, top-down preference. Owning the call costs capital up front; being right repays it.Why it's cited here: The thing unearned authority makes impossible. You can only put your name behind a call your experience actually covers; performing the experience gets you the signature without the backing.Open ZFN-41 → makes the same trade and gets the same answer: the capital you spend being straight is repaid by still being believed on the fifth thing, and there’s no way to earn that back once you’ve spent it the other way.

New obligations: If you want people honest about the edge of their knowledge, you have to make that edge survivable. React to “I don’t know” as information rather than as a shortfall, or you’ll train it out of the room in about two weeks and be left with a team that tells you what you wanted to hear (which is a much worse problem, and a quieter one). And apply it to yourself before you apply it to anybody else, because without that all of this is just a licence to interrogate other people about a standard you’re exempt from.

References

• ZFN-26Field Note · currentZFN-26 — LLM-assisted content needs no disclaimer, only a human who can back itDrafting engineering content with an LLM needs no disclaimer — inside a team or codebase that already shares the co-sign norm. The obligation is human co-signing. Writing for strangers inverts it: there, stating your policy builds trust rather than diluting a default.Why it's cited here: Co-signing is this position stated as a rule for published work. The co-sign is exactly the claim that you can back what went out under your name.Open ZFN-26 → - the co-sign: what it means to be able to back what goes out under your name.
• ZFN-27Field Note · currentZFN-27 — Don't tolerate assholes — but be strict about what one isDon't tolerate assholes — people who demean, belittle, punch down. But filter hard on the word: disagreeing, raising ideas, or opening a competing PR isn't being an asshole, it's the work. Assholes attack people; colleagues attack problems. Don't let the label silence dissent.Why it's cited here: The same discipline applied to a different label. There the argument is that 'asshole' has to stay narrow or it silences dissent; here it is that 'grifter' has to stay narrow or it punishes people for being new.Open ZFN-27 → - keeping a label strict so it names a behaviour instead of silencing people.
• ZFN-28Open problem · currentZFN-28 — Capability without understanding: brute-force LLM PRsAn open problem: people brute-force PRs with LLMs in domains they don't understand, taking on more than their knowledge supports — and the struggle that used to teach them is smoothed away. How do we stop un-understood code without killing learning or banning a good tool?Why it's cited here: The code-level form of this, already open. Shipping a change you can't explain is claiming authority over it, whoever or whatever wrote the diff.Open ZFN-28 → - the same failure at the level of a diff, and still an open problem.
• ZFN-29Field Note · currentZFN-29 — Blameless culture, taken seriously — and its one hard lineWhen something breaks, support the person, don't blame. Run post-mortems with ceremony and learn at every level — software, org, culture, process, even solo. Blameless protects honest mistakes, not dishonesty: evading or blaming gets coached; hiding evidence is a firing offense.Why it's cited here: Where the line actually falls. Blameless culture protects honest mistakes and stops at dishonesty, and deliberately overstating what you know is that dishonesty before the incident rather than after it.Open ZFN-29 → - blameless for mistakes, hard line at dishonesty.
• ZFN-41Field Note · currentZFN-41 — Make the technical choice you can put your name behindWhen two options would both work, it's fine to choose the one your experience lets you put your name behind, and to resist pressure that isn't about the merits — vendor incentives, politics, top-down preference. Owning the call costs capital up front; being right repays it.Why it's cited here: The thing unearned authority makes impossible. You can only put your name behind a call your experience actually covers; performing the experience gets you the signature without the backing.Open ZFN-41 → - owning a call your experience actually covers.
• ZFN-50Field Note · currentZFN-50 — See something, say something — you owe a report, not a repairTwo standing duties beyond your assigned work: see something, say something, and leave things better than you found it. But the duty is to report, not to repair — route it to the owner and move on. Security issues always go to security, including ones you fixed yourself.Why it's cited here: The same muscle, pointed outward. Saying 'I don't know' and saying 'I think this is wrong' both cost something in the moment and are both owed.Open ZFN-50 → - the adjacent duty: saying the thing that costs something to say.
• Cargo Cult ScienceCargo Cult Science (Richard P. Feynman, Caltech commencement address, 1974)Feynman's address on scientific integrity, which he describes as a kind of utter honesty, a leaning over backwards to report everything that might make you wrong rather than only what supports you. Its most quoted line is the whole discipline in one go: you must not fool yourself, and you are the easiest person to fool.calteches.library.caltech.edu ↗ - Feynman on utter honesty as a working discipline.
• Dunning–Kruger effectDunning–Kruger effectThe finding that people with low ability in a domain tend to overestimate it, because the judgement needed to rate your own competence is the same judgement you currently lack. It describes a miscalibration the person is genuinely unaware of, which makes it the harder version to fix rather than the more forgivable one.en.wikipedia.org ↗ and impostor syndromeImpostor syndromeThe experience of doubting your own competence and feeling like a fraud despite clear evidence of achievement. It is the same instrument miscalibrated in the opposite direction, authority genuinely earned but never felt, and it distorts a team's picture of who knows what just as effectively.en.wikipedia.org ↗ - the same miscalibration in both directions.

Changelog

• 2026-09-17: Sharpened the thesis after feedback, which is a bit on the nose for this particular note. Ambition was wrongly bundled in with the failure mode: aggressive goals and missed targets aren’t overclaiming, only overstated depth of understanding is. Added The unstated claim, since the version worth worrying about usually contains no false statement at all. And rewrote the Dunning–Kruger treatment, which previously excluded the unwitting case; it costs exactly as much as the deliberate one, so the standard covers both even though the word “grifter” shouldn’t.
• 2026-09-17: First published as a Field Note.