how can you tell if you are smart — 4 honest tests

you can tell you are smart, supposedly, by certain signs. i have most of them and also several of the warning signs for the opposite condition. the test is inconclusive. the certified letter on the counter is, possibly, a hint.

writing this from the standing desk, which i am sitting at, because the standing part of standing was a marketing decision i did not consent to. carla is in a procurement walkthrough on the third floor. i have, by the agenda’s look, around fifty minutes.

so the question parked in front of me at 11:34am on a thursday is how can you tell if you are smart — not the iq number you misremember from 2008, not the framed certificate, but the practical version. the one that matters when you’re alone with a fork and the seventh microwave. i have been building tests for years. none, i’m warning you, are reliable.

how can you tell if you are smart: mostly you can’t, not from the inside — the brain that grades the test also wrote it. practical signs show up in behaviour, not self-report: pausing before you speak, knowing what you don’t know, asking better questions than you answer, changing your mind quietly. people sure they are smart usually aren’t.

SMART. IS. NOT. A. SCORE.

i need that on the page before we move. people who treat smart like a number — measured once, framed twice, brought up over a third drink — tend to be running a museum, not a brain. the ones i’ve watched do the actually-smart things mention it less than they mention the weather.

how can you tell if you are smart, the short answer

the short answer is: you watch what you do when nobody is grading you. that’s the whole post if you’re skimming. the long version is me arguing with that sentence from four angles, with four witnesses, none of whom are qualified.

the catch is the one i keep tripping over. the brain doing the watching is the same brain being evaluated. it has skin in the game. mine, on a wednesday, decides i am a small genius. on a friday it decides i should not be allowed to operate a kettle. neither verdict is admissible. but you keep filing them, because it’s the only court available.

this is the same machinery that runs confirmation bias — the trick where the brain only counts evidence that flatters it and drops the rest in a folder labelled misc. once you know that’s happening, the question shifts. it’s not am i smart. it’s am i grading honestly. the second question is harder. the second question is the work.

the bar test mike has been refining

mike, at the corner bar, has a test he updates every time someone new walks in and says something he doesn’t like. mike (who, by his own filing record, has not voluntarily met the tax authority since 2019) considers himself a good judge of brains anyway. current version, one question: “can you describe what you do, in plain words, to a person who is paying for the next round.”

mike does not want jargon. mike wants the sentence you’d use if your job depended on the listener actually understanding. consultants fail hardest. mike has watched a director-level title spend eleven minutes describing his role and never land a verb. mike scored him zero, silently, and ordered him another beer because the verbal failure is its own punishment.

i have failed mike’s test three times. once i used the word stakeholders, disqualifying. once i used alignment, also disqualifying. the third time i said i sit at a desk and process numbers nobody reads, and mike nodded, which was the closest thing to a passing grade he was prepared to award me.

sarah passes a different test, allegedly

sarah — long-distance sarah, marathon sarah, sarah-with-a-pension-she-actually-understands — would pass a different test. hers wouldn’t involve a bar. it would involve mile fourteen of a long run, the point at which the brain stops pretending and goes quiet, and the only thoughts left are the ones that survive the legs. sarah believes the brain at mile fourteen is the real one. the brain at the desk, at the bar, at 11:34am on a thursday — those are, by her account, drafts.

i find this slightly menacing, because i do not run. the closest i get to mile fourteen is the walk from the elevator to the standing desk, which produces no thoughts of any kind. by sarah’s metric i have never accessed my real brain. she once summarised the matter in a sentence: “you can tell someone is smart by what they don’t pretend to know.” i wrote it down. it lives in the drawer.

productivity bro tweets a daily test

productivity bro — sword in his bio, morning routine that begins at 4:18am and never recovers — has been posting a daily smart test for, by the algorithm’s account, the entire month. yesterday’s: “if you can recite your top three priorities for the quarter without checking notion, you are operating at a higher cognitive tier than 94% of professionals.” the percentage is invented. the implication is you should pay him to fix the gap.

i checked. i could not recite my top three priorities for the quarter. i could not, on inspection, recite the quarter we are currently in. mike, asked the same, said his top three were “this beer, the next beer, and not getting a parking ticket.” by his own scoring, mike was also in the bottom 6%, but in a way more honest than any productivity bro thread i’ve scrolled. books on tape are cheating, while we’re keeping receipts on what counts as real input.

the chatgpt test that filtered me out

i did, recently, ask the chat window directly. not because i trust it — i don’t — but because i was curious what the average of the internet would say, served back without a face to argue with. i typed: am i, honestly, smart. the model gave me eight signs. seven i already knew. the eighth: “smart people accept feedback without becoming defensive.”

i became, immediately, defensive. i told the chat window, in writing, that this was a generic statement and the model was not qualified to assess my defensiveness. then i closed the tab. then i opened it again and asked a follow-up that contained the word actually three times. the model did not flinch. i had failed sign eight in front of an entity that does not have witnesses, and yet i felt observed. that, i think, is data.

chatgpt also reminded me, in a sidebar i did not request, that there are eleven subscriptions on my card from services i no longer use, including one called brain-dot-fm i signed up for in 2022 and have not opened since. it offered, for the third time, to audit them. i declined, for the third time. the subscriptions stay — evidence, possibly, that i’ve settled for billed.

if you want a more thorough route through how the brain talks itself into things, the long-form on what confirmation bias actually means walks the mechanism, the deep-dive on cognitive bias as a category covers it when one bias isn’t enough, and the sibling on the four-question version of telling if you are smart has the checklist i scored a 2 on last week.

verdict — the telling is the trap

for a longer route into how a man’s certainty about his own brain collapses on contact with a foreign country, see karl pilkington’s bewildered tour of the world in an idiot abroad, which is, in my private theory, the best long-form documentation we have of the limits of self-assessed brainpower. abroad is the test, and the idiot is the carry-on.

carla is back from procurement, holding a binder, making a face that suggests the binder will not be opened until tomorrow, possibly never. i’ll close this tab and pretend to be processing the variance report on the second monitor.

↑