← Carlos Cardona

AI brought the impostor feeling back

There is a guy on YouTube who seems to know everything. New model dropped this morning, he already benchmarked it. New framework, he shipped a demo. New agent pattern, he has a video, a repo and an opinion. You watch him and a quiet conclusion forms: I am hopelessly behind.

I keep meeting experienced, accomplished people who carry that feeling right now. Founders, senior engineers, people with fifteen years of shipped work behind them, all quietly convinced they are frauds because they have not caught up with this week's AI stack. The impostor feeling was always common in our industry. AI turned it into an epidemic, because the field now changes faster than any individual can absorb, which means everyone is permanently behind on something, which is exactly the condition the impostor feeling feeds on.

So it is worth being precise about what the feeling actually measures.

The impostor feeling comes from comparing your inside against other people's outside, and concluding you do not belong. But the underlying math never changes: there will always be people far ahead of you, and there will always be people far behind you. That was true before AI and it is true now. The only thing AI changed is the visibility and the speed, more people ahead of you are in your feed, and the frontier moves weekly instead of yearly. The feeling got louder. The information in it did not increase, because a feeling that fires for everyone, always, regardless of their actual position, carries no information about you. It measures the size of the field, not your adequacy.

And about the guy on YouTube: keeping up is his job. That is the whole job. He does it for a living, with his full working hours, and the algorithm rewards him for looking effortless. Most people are not doing that. Most people are busy doing their actual work, running their businesses, shipping their products, and squeezing learning into the edges. Comparing your edges to his center is not a measurement, it is a category error.

Here is the part I actually want to argue, because the usual reassurance stops too early. Your years of experience did not lose value in this shift. They gained value.

The new tools collapsed the cost of execution. What they did not collapse, and cannot, is knowing what to execute. When you can direct a fleet of agents and automated workflows, the constraint stops being how much you can do and becomes how well you can decide. And deciding well is precisely what eight, ten, twenty years of experience builds: the judgment to smell a bad architecture before it bites, the taste to know which of five working versions is the good one, the agency to move without being told. A twenty-two-year-old and you have equal access to the models. They do not have equal access to your scars.

So yes, learn the new stack. Genuinely, it is worth it, and the people who refuse will pay for the refusal. But learn it as someone adding leverage to two decades of judgment, not as an impostor scrambling to requalify for a field they already belong to. You are not behind everyone. You are behind some people and ahead of others, like every person who has ever been good at anything. That has always been the deal. The feeling insisting otherwise is not insight. It is just the size of the field, talking.