/discussion

**The Day I Realized Nobody Was Coming to Expose Me** For almost four years I woke up convinced I'd be found out that week. I didn't know by whom, and that was the strangest detail. There was no inspector, no committee, no furious client with a dossier in hand. It was a vague sense that at any moment someone would open the repository, read the code I'd written and say out loud what I already knew. That this wasn't the work of a professional. Nobody said it. The four years passed, and then a few more. It's worth starting by taking the name apart, because the name is already half the problem. "Impostor syndrome" sounds like a diagnosis, something with a leaflet and a treatment, and it's nothing of the sort. Pauline Clance and Suzanne Imes described the phenomenon in 1978, in a paper about high-achieving women who couldn't internalize their own success, and they called it the "impostor phenomenon." Phenomenon. It never made it into the DSM, it isn't a recognized disorder, and anyone selling eight-week courses to "cure" it knows this perfectly well. The word syndrome sells better. When you stick a clinical label on something that is, in large part, a reasonable reaction to unreasonable environments, you move the problem from the environment into the person's head. That suits a lot of people. So I distrust almost everything written on the subject, and that includes, for what it's worth, what I'm about to write here. There's a number that shows up everywhere, that roughly seventy percent of people have been through this. It comes from a 2011 review by Sakulku and Alexander, and anyone who reads the original finds enormous ranges, student samples, different scales measuring similar but not identical things. Someone grabbed one end of it and made it a headline. I'm not saying the number is false. I'm saying I don't know what it measures, and when a statistic works equally well in a LinkedIn post and in a self-help book, it usually measures very little. But the thing exists, I don't dispute that. I know it from the inside. The first episode I remember clearly was in 2013, on a Thursday, at 11:40 p.m. I was cleaning up an old sessions table in a production database, a gesture I'd made a thousand times, and MySQL threw me an ERROR 1451 (23000), the foreign key constraint one, telling me I was trying to delete something another thing depended on. I fixed it in ten minutes. Anyone with two years of experience would have. And still I spent the following night convinced that a real programmer would never have gotten to that point, that a real programmer would have done the migration differently, with a prior step I didn't know and everyone else did. The existence of that hypothetical step tormented me for weeks. I never found it. I think it never existed. This sounds like a small anecdote and is, in a way, the whole mechanism. The person doesn't compare what they did with what was reasonable to expect. They compare it with an imaginary professional, built out of every good thing they've seen in everyone else, with none of the mistakes those others also made and never told them about. It's an opponent you can't beat because it has no flaws. I played against him for years. Now the part that irritates me about the friendly approaches to this topic. They say the feeling of fraud is a distortion, that the person is much better than they think and you just need to show them the evidence. In many cases that's true. In too many, it isn't, and nobody says so because it sounds cruel. There are fields where the surface area of knowledge is so vast and changes so fast that feeling you know little is the correct reading of the situation. Web development is the example I know best. Between what gets written in JavaScript, what runs on the server, what the database does underneath, what the browser decides to do with your CSS and what the search engine decides to do with your site, nobody masters all of it. Nobody. Whoever says otherwise is selling something, or has never looked at what they don't know. So the person who feels like a fraud because they don't understand exactly how the TLS handshake works isn't suffering from a cognitive distortion. They're noticing, accurately, a real gap. The mistake is thinking the gap disqualifies them, when every colleague has gaps of a different shape. Here I hesitate a little, because if I take this argument to the end, it says the feeling is just calibration. And it isn't only that. There are genuinely competent people, with measurable results, who feel this with an intensity that corresponds to nothing external. People who get promoted and end up worse. For them, the "real gap" explanation is a cheap consolation. The mechanism is something else, and it has much more to do with how success gets attributed, always to luck, to timing, to the client being easy, than with the size of the knowledge. But both things coexist, and I think anyone writing about this should admit they don't know which weighs more in each case. I don't know which one weighed more in me. It's the second-order effects that interest me, because they're the ones nobody counts. The first is extra work. The person who feels like a fraud compensates with hours, and in a very specific way that you can't see from outside: they never hand in the first version. They review, rewrite, add tests nobody asked for, document what nobody will read. On the surface this looks like diligence and is rewarded as such, which feeds the cycle, because the praise confirms the extra effort was necessary. If I needed three times as long for it to be good, and they said it was good, the logical conclusion, inside the head of someone who suffers from this, is that without the extra effort it would have been bad. The experiment of handing in the ordinary version is never run. You never find out the ordinary version would have passed too. It's a study with no control group, and the subject is the person themselves. I once measured it. For a quarter, in 2017 or 2018, I logged how much time I spent on tasks after they were already functionally finished. The average came to 37% of total time, polishing time that no user would see and no colleague would review. I stared at the spreadsheet in silence for quite a while. I didn't change anything right away, of course. Knowing the number didn't switch off the habit. The second effect is worse, and it's silence. Someone who feels like a fraud doesn't ask questions, or asks questions so over-prepared and so apologetic that the other person gets lost halfway through. There's a cruel logic here. Asking means risking exposing the gap, and the gap is precisely the evidence the person believes someone is looking for. So they stay quiet, lose two days solving alone a problem that a thirty-second question would have dissolved, and then feel vindicated for having solved it alone. I saw this on teams I worked with, and I saw it in myself. On a project in 2016 I spent more than a week fighting an nginx configuration, an intermittent 502 Bad Gateway that only showed up under load, because it cost me too much to ask the one colleague who knew the answer. It was a buffer parameter. He fixed it in four minutes when I finally showed him the logs, and didn't even comment. I'm the one who held on to the comment he didn't make for months. (By the way, I never forgot his face when I showed him those logs. The face of someone thinking about something else. Lunch, probably. This interests me more than it should: how often our worst hour is just another person's hour of thinking about lunch.) The third effect, the third-order one, is the one that seems the most treacherous to me and shows up least in the literature. It's what happens to money and ambition. Someone who feels like a fraud doesn't negotiate salary. It isn't shyness, it's an almost bookkeeping-like conviction that they're already getting more than they deserve, that asking for more would hasten the moment of discovery. And there are employers who know perfectly well how to recognize this profile, even if they don't put it that way. The person who thanks you for every raise like it's a favor, who takes the extra project without compensation because "it's a learning opportunity," who never says no. Being cynical, and today I'm a bit cynical, the feeling of fraud is an economic resource for people who manage people. In most cases it isn't created on purpose. But it gets exploited. I've seen managers cultivate it by omission, avoiding clear feedback, keeping the person unsure whether the work is really good, because an insecure person works more and asks for less. It doesn't need to be malice. Laziness about giving honest feedback produces the same effect. Now, my own version of this changed shape when I became a founder. I thought it would get better. I thought that, being the owner, there was nobody to answer to, so nobody to fool. It was the opposite. When you run a project alone, a community, a platform with real users, there's nobody to confirm the decision was good. No code review, no retrospective, no colleague saying "you did that well." There are numbers, and numbers are ambiguous. A week where organic traffic drops twelve percent could be a Google algorithm update, could be seasonality, could be me having broken something in the titles. With nobody around to give me context, the interpretation my brain picks by default is always the third. It's me. I broke it, and now they'll see. And the most perverse part is that success doesn't fix it either. An article that does well, a page that climbs to the top, a traffic spike on a Sunday afternoon at 4:12 p.m., and the first reaction isn't satisfaction, it's suspicion. It was luck. It was a topic that happened to land. Anyone could have written that. I've looked at a graph going up and felt a physical discomfort, as if I were receiving money that wasn't mine. I have to stop here for a second to say something that contradicts what I said earlier a bit. I said knowledge gaps are real and that feeling limited is often an accurate reading. I stand by that. But the solo founder has an additional problem that isn't about knowledge: it's about witnesses. Nobody saw. And the feeling of fraud lives largely on a lack of witnesses. In a job, even a bad one, there's someone who saw what you did and can say, for better or worse, that it existed. Alone, the work is so invisible that it becomes easy to doubt it was good, or even that it happened the way I remember. I don't know if this is a general observation or just an excuse of mine. It stays here. Now, what actually helps. Here I'm going to be biased, because I have opinions and no desire at all to balance them against the opposite thesis. The usual advice doesn't work. "Acknowledge your achievements," write a list of the things you did well, they say. I did that. A sheet with twenty-three items, printed, which I kept in a drawer. I read it twice. By the end of the second reading, my brain had already come up with an explanation for each of the twenty-three. It was favorable context, it was a colleague who helped, it was a generous deadline. Achievement lists only work on people who didn't need them. Someone who suffers from this has an extraordinarily efficient internal machine for discounting evidence, and giving it more evidence is giving it more raw material. What pulled me out of the worst place, I won't say cured, I'll say pulled, were three small things and none of them is glamorous. The first was reading other people's code. Really reading, line by line, the code of open-source projects I admired, and discovering it was full of TODOs, commented-out hacks, two-hundred-line functions and variable names like "tmp2." Respected software, used by millions, written by people who also improvised at two in the morning. This isn't news to anyone who's been in the trenches, but reading with the eyes of someone looking for proof that others fake it too is different from reading with the eyes of someone looking to learn. The second was starting to ask earlier, deliberately clumsily. Without preparing the question, without apologizing. "This doesn't work, I don't know why, can you take a look?" I discovered that people like being consulted. They like it quite a lot, actually, more than they like watching someone solve things alone. The third was harder to explain: I accepted that the feeling wasn't going away and stopped waiting for it to go. That last point deserves more care. I spent years waiting for the moment I'd finally feel legitimate, like someone waiting for a train. First, when I had two years of experience. Then five. Then, when I had a product with users. Then, when I had a product that made money. The target always moved one step ahead of me, and I think this is the most characteristic trait of all of this, more than the anxiety: legitimacy is always the next stop. It's never the current one. When I understood the pattern, I didn't feel relief. I felt a sort of boredom, like when you discover the trick of a magician who's less good than you thought. What remained was a more practical relationship with doubt. I still wake up, some mornings, with the feeling that today's the day. Only now I treat it the way I treat a monitoring alert with frequent false positives: I log it, check whether something is actually wrong, and if not, carry on. It isn't elegant. It isn't an inner transformation. It's maintenance. A note for people who lead teams, because it's where I think the most could be done and the least is. If you have people reporting to you, what helps them most isn't telling them they're good. That goes in one ear and out the other, and in someone who suffers from this it's even counterproductive, because it sounds like courtesy. What helps is being specific and dryly factual. "The way you structured the migration avoided the downtime we had in March." No adjectives. And above all, telling your own mistakes, out loud, without drama, as working information. I learned more from a former boss who admitted in a meeting that he'd deleted, in 2011, an entire bucket of client files by mistake than from any workplace wellbeing training. He said it in the same voice he used to tell the time. I never looked the same way again at the idea of a "flawless professional," because I'd just watched one describe his flaws like someone reading an invoice. There's one thing, though, that I distrust and almost nobody mentions: the feeling of fraud is also useful. I say this with some discomfort, because it sounds like romanticizing suffering, and that isn't the idea. But people who never doubt themselves are, in my experience, considerably more dangerous in technical fields. I once worked with a guy like that, not a shadow of doubt, who deployed to production on Friday afternoons and said "everything's under control." A dose of distrust in yourself is what leads someone to write the test, read the log to the end, double-check the query before running it in an environment that matters. The problem isn't the doubt, it's the dose, and the inability to switch it off when it no longer serves any purpose. I don't know exactly where the line is. It probably sits somewhere different for each person, and it moves with tiredness, with sleep, with the month you're having. One more thing needs saying about the large number of people who, now that the topic is fashionable, use the label as an identity. Some say "I have impostor syndrome" with the same lightness as "I'm terrible in the mornings," as a nearly endearing personality trait, a socially acceptable weakness that, curiously, only the competent confess to. I'm a little suspicious of this. Not of the pain, which is real, but of the social use of the concept, which sometimes works as a discreet way of asking for compliments or faking humility. I've caught myself doing it. At a dinner, talking about my work, saying "honestly I don't really know what I'm doing," waiting for someone to contradict me. Someone did, and I felt a little worse, because I realized it was a question disguised as a statement. I think what bothers me, deep down, is the clean narrative. The one that says the feeling has an origin, a mechanism and a solution, and that following five steps gets you to the other side. My experience wasn't like that. It was slow attrition, with relapses, in which what changed wasn't the feeling but the distance between the feeling and what I did next. Before, the doubt decided for me: I didn't publish, didn't ask, didn't charge, didn't launch. Today the doubt still speaks, but it no longer has a vote. It talks, I listen, and I do what I was going to do. I sometimes think about the day I started writing this, and the work it took to begin, and the fact that I rewrote this introduction four times. Four. If that isn't the perfect example of what I described, I don't know what is. The version that stayed is, probably, the same as the second. I never found out about that prior step in the 2013 migration, by the way. Even today, now and then, I catch myself looking for it.