AI Told Me to Buy. I Hesitated. It Tripled.

I want to tell you about a mistake that still bothers me a little.

I’d asked Claude to help me think through a sector tied to AI infrastructure — the kind of industry that benefits no matter which specific company wins the broader AI race. The analysis came back clear: fundamentals improving, demand accelerating, conditions genuinely favorable.

My reaction was basically: “Makes sense. But this has already run up a lot. I’ll wait for a pullback.”

It didn’t pull back. It tripled.

I never bought in.

The Excuse Disguised as Patience

A lot of people would say the lesson here is “the AI was right and you should have listened.” But that’s not actually what went wrong.

What went wrong is that I hadn’t done my own work yet.

“I’ll wait for a pullback” sounds like discipline. In my case, it was a translation of something less flattering: I hadn’t calculated what I thought the companies in that sector were actually worth. I hadn’t built a framework for judging whether the price was already too high or still reasonable. I had a conclusion from Claude, but I hadn’t internalized it into my own judgment.

So once the price had already moved, I panicked a little — not because I’d made a wrong call, but because I had no way to answer the only question that actually mattered: is this price still worth paying?

If I’d done the work — estimated a fair value range, figured out where my margin of safety would sit, formed my own view of the industry logic instead of borrowing Claude’s — I could have answered that question regardless of what the price had already done. Instead, “wait for a pullback” was just a polite way of avoiding a decision I wasn’t ready to make.

The Rule I Built From It

Here’s what I took away, and it’s become one of the few hard rules I actually follow now:

If your reasoning for not acting is “I’ll wait for a pullback,” stop and ask yourself: have you actually calculated a price you’d be willing to pay?

If the answer is yes — you have a number, you’ve done the valuation work, you’re genuinely waiting for the market to meet your price — then it’s a real plan.

If the answer is no, you’re not waiting for a pullback. You’re just not finished thinking yet, and the price has nothing to do with it.

I didn’t have a number. I had a feeling that the run-up made me uncomfortable, dressed up as a strategy.

Why I’m Not Telling You What to Buy

I’m deliberately not naming the sector or the companies here, and not because of any legal disclaimer reflex. It’s because the specific industry isn’t the point. By the time you read this, that opportunity is long gone, and chasing it now would be exactly the kind of mistake I’m describing — acting on someone else’s old analysis instead of doing your own current thinking.

What’s still useful, months or years later, is the pattern: AI is very good at giving you a clear-eyed read on fundamentals. It is not a substitute for having your own framework for what something is worth. The gap between those two things is exactly where I lost out — not because I was wrong, but because I wasn’t ready to be right.

I run a small Chinese-language account now where I write about exactly this kind of moment — the real, sometimes uncomfortable thinking behind value-investing decisions, not just the wins. This was one of the stories I told there first, translated and a little more honest here, because I think the lesson holds in any language.

Have you ever talked yourself out of a decision using a reason that sounded smarter than it actually was? I’d like to hear it — the embarrassing ones are usually the most useful.