The Confident Ones Don't Know Either

In 1945, the physicists calculated the probability that the Trinity detonation would ignite the atmosphere. Set off a chain reaction. End all biological life on Earth.

The number came back very small. Probably fine.

"Probably" was carrying the weight of every living thing on the planet.

They detonated it anyway.

Oppenheimer watched the explosion and thought of the Bhagavad Gita: Now I am become Death, the destroyer of worlds. He said that after. Not before.

The builders of the most consequential technology in human history didn't know what would happen. They built it anyway. That's just what builders do.

Sam Altman is the most visible voice in AI today. He has testified before the US Senate. He gives interviews constantly. He projects direction and purpose.

He has also said, publicly: "I feel totally confused about what society looks like" when describing a future with highly capable AI. On X, he wrote: "near the singularity; unclear which side."

That's not a hedge buried in fine print. That's the man building it.

Altman still needs capital. He needs engineers. He needs you to believe the project is worth running. His confidence is partly fuel for the mission. That's not dishonest. But it's not the same as knowing.

I had lunch with Mo Gawdat last year.

Former Chief Business Officer at Google X. One of the few people alive who has actually been inside the rooms where some of these decisions are made. Smart. Thoughtful. Credible in a way that most people talking about AI simply are not.

He has said publicly that people should not be having children right now.

I pushed back. Isn't that the point of being human? Connection. Life. The present moment. If there's a utopia on the other side of this, and I believe there could be, then the answer isn't to stop living.

He listened. Then: your job as a parent is to provide a safe environment for your child. What if you can't?

For the record: I don't have children yet. We were talking about an idea. He treated it like a real question anyway.

That's not a hedge. That's a worldview shaped by proximity to real uncertainty.

I hold Bitcoin. I bought it years ago knowing almost nothing about the underlying technology. Smart people I respected seemed to believe in it. Some version of that was enough for me to act.

I still hold it. I still couldn't fully explain it to you if you asked. I'm in it anyway.

If the people building AI don't know. If the most credible insider warns against having children. If I can't even explain my own largest positions. Then what exactly is the confident newsletter selling you?

Here is what happens when AI research gets published.

A frontier researcher publishes a finding. A journalist summarizes it. A newsletter picks that up. A podcaster reads the newsletter. A LinkedIn creator watches the episode and posts their three takeaways. You scroll past it on a Friday afternoon and sound interesting at dinner.

By the end of that chain, the signal is gone. The confidence isn't.

And we follow anyway. Not because we're uninformed. Because in a system this large and this fast, the loudest available signal is the default. The signal is already empty before it reaches you. Nobody is stupid. The chain is just doing what chains do.

I hold their stocks. Through indices, passively. I'm in the system too. What I don't do: make individual bets on which company wins based on what someone said on a podcast or in a newsletter. That distinction is everything.

I invest in AI infrastructure. I believe the direction is inevitable.

I don't know what's coming.

Neither do you. Neither does almost anyone.

If you want to know what I do with that: I wrote about it here.

Everyone sounds certain. Nobody is.

I wrote this with AI. Which probably tells you something.