am i really stupid — how to audit the question, in 7 steps

the barista was new and the line was long and i still tipped, which is relevant to nothing except that i was thinking about the question while waiting. how to audit it honestly takes steps. step one is to refuse the word really, which is asking the question to confess to something the question cannot prove.

i’m at my desk now. carla is on the third floor in an annual planning meeting that her calendar lists as “blocking” with no further detail, which is the corporate equivalent of leaving the room and locking the door. i have, by the count i keep running, about an hour before someone from procurement walks past and looks meaningful. plenty of time to investigate.

so. am i really stupid. the question arrived this morning the way these questions arrive — uninvited, dressed as honesty, asking to sit down. i let it sit down. then i started auditing it.

step one: identify the source — often a tweet, often a stranger

before you investigate whether you are really stupid, investigate where the question came from. nine times out of ten it did not come from inside the building. it came in through the phone. a comment. a reply. a quote-tweet from a man with a checkmark and a kettlebell in his bio.

this morning the source was online and unsigned. someone i do not know suggested, in a thread about something i was not a part of, that people who do the thing i happen to do are, broadly, stupid. i read it twice. then i closed the app. then i opened it again. then i went to the pillar i keep on this exact subject — the one that explains, at length, that the entire concept of stupid as a verdict is a category error people use to shortcut having to think about you specifically.

the source matters because the source determines what the question even is. a stranger calling you stupid is not the same data as your mom calling you stupid, which is not the same data as your doctor calling you stupid, which would not happen because doctors say “have you considered” and let you finish the sentence yourself.

step two: separate the source from the verdict

this is the move most people skip. they hear the question, feel the question, and answer the question in the same breath. that’s not an audit. that’s a confession dressed as investigation. there is a difference, and the difference is the entire post.

separating the source from the verdict means asking two questions instead of one. one: who said it. two: is the thing they said even on the table. if the source is “a stranger online with a productivity newsletter”, the answer to question two is, generally, no. it is not on the table. it never was on the table. it was on a different table in a different building and someone airdropped it.

i learned this the way i learn most things — slowly, after hitting the obvious wall a number of times that i’m not going to write down. the feeling of stupidity is not the fact of it, and the fact of it is not the feeling. these are two separate investigations and they get filed in two separate drawers.

step three: a productivity bro tweeted — that is not data

now we get to the specific source of this morning’s question, which i’ll address by category rather than by name because the category is more useful and the name is, at the count i keep running, the eighteenth one this year. productivity_bro tweeted. that is not data. that is weather.

productivity bro types tweet for a living. their job is to produce a daily quantity of confident sentences about how the rest of us are doing it wrong. the sentences are calibrated to provoke. provocation is the product. you reading the sentence and then auditing yourself for an hour is the business model. i am, right now, the business model. i am aware of how that sounds.

the test for whether something is data: would the person saying it stand behind it in a room with you and no audience. tweets fail that test by design. they were built for a room that is only audience. a man performing in front of a metric is not the same as a man telling you something true. one of those is a study. the other is theatre, which, like a shower over four minutes, is theatre.

step four through six: the 3 internal checks and the 3 external checks

once you’ve thrown out the source, you can run the actual audit. it has six checks, in two columns. internal first, then external. do not reverse the order. people who ask the external column first end up with the wrong answer because they are using other people as a mirror that does not know them.

the internal column, which i ran this morning at the kitchen counter before coming to the desk, looks like this. one: in the last week, did i do the thing the question is implying i can’t do. two: in the last month, did i learn one specific thing i did not know before. three: when something went wrong, did i notice it went wrong, or did i need someone to tell me. these are not yes/no in the satisfying sense. they’re yes/no in the “yes, sort of, with caveats” sense, which is the only honest sense available.

the external column, which i ran on the walk over, looks like this. four: do the people who actually know me — not the bros, not the strangers, not the algorithm — treat me like i am stupid in the way the question implies. five: in the last year, did anyone with skin in the game (rent, money, time) bet on my judgment. six: when i told the truth about a mistake, did the room get quieter or louder. quieter means people are recalibrating. louder means people already had their answer.

i ran my six. i’m not going to publish the results because the results are the audit, not the post, and also because i am sitting eight feet from the procurement guy and he reads over shoulders. but i’ll tell you the shape of the result: less stupid than the question implies, more stupid than i would prefer, exactly as stupid as the average person who asks the question, which is the entire trick.

step seven: the close, with rest

the seventh step is the one nobody likes. you close the audit and you rest. you do not re-open it that evening. you do not re-open it on the train. you especially do not re-open it on a monday, because mondays are objectively better than fridays for any kind of investigation but worse for any kind of revisit. the one i ran this morning is closed. the one i’m tempted to run tonight, in the kitchen, is the same audit with worse lighting.

resting after a self-audit feels wrong because the question wants more food. it wants you to keep feeding it. that’s how the trap works. it’s the same architecture as the third yoga mat — you bought the first one because you were going to be a person who does yoga, you bought the second when the first one failed to make you that person, and the third is still under the couch in the kitchen from 2023, possibly evolving. each audit was supposed to settle it. each audit re-opened it. that is not investigation. that is a hobby.

verdict: am i really stupid, the question deflates under audit

so. am i really stupid. after seven steps, the verdict is: the question deflated. it was never about cognition. it was about a tweet from a stranger plus a feeling that arrived dressed as honesty plus a word — really — that was carrying water for the whole sentence. throw out the source, throw out the word, run the six checks, close with rest. what’s left is a person at a desk with forty-three minutes and a microwave at home he respects more than he respects the question.

if you are running this audit yourself, i’d suggest reading the longer take on the word stupid alongside it, because the audit gets faster when you’ve already done the linguistic work. and i’d suggest watching the way an idiot handles this kind of question on a sitcom — fictional idiots run the same audit, they just run it in twenty-two minutes with commercials, which is, by the count i keep running, the right length for most internal investigations.

the question is the trap. the audit is the way out. the rest is just rest.

↑