You're the one hallucinating now

We used to complain about AI hallucinating. But it’s gotten pretty good now. So good, you no longer really hear people complaining much about AI hallucinations.

But now we got a different hallucination problem — the people using the AI.

Because of how AI now is so good, so easy, and so agreeable, you start hallucinating that whatever it produced would work well.

The fancy LinkedIn post that ChatGPT wrote for you.
That market research Gemini scoured the web for you.
This flawless code that Claude Code vibecoded for you.

And that product, SaaS, and business you built in a day with AI, you start to hallucinate that it will hit the market and start printing money.

But everyone could tell that your LinkedIn post was AI slop.
The market research was based on one comment in Reddit.
The code that worked, introduced more debt than you realised.

And somehow, the product didn’t sell. A few friends tried it. No one paid for it. And you wondered what went wrong.

You went wrong.
You hallucinated the outcome.

Because the output was so so easy, you thought the outcome would too.
Because you outsourced the work to AI, you also outsourced judgement.
Because the AI sounded so confident, so you convinced yourself it is right.

You hallucinated.

So now the hallucination radar has to flip. Instead of looking out outwardly for AI hallucinations, you have to watch yourself inwardly for hallucinations. You’re the easiest to fool, by yourself. And when equipped with AI, you’re 100x, 1000x, more likely to be fooled.

And trust me, you will hallucinate.
It’s not a matter of if, but when.

You’re the one hallucinating now.