This is a headgum podcast. I don't know what to think I don't know what to say.

Yeah, but that's all right.

That's okay.

I don't know anything.

Hello and welcome to Factually. I'm Adam Conover.

Thank you so much for joining me on the show again.

Artificial Intelligence is supposed to revolutionize the world and our economy.

The major platforms are spending tens of billions of dollars on it just this year alone

to gain an advantage.

It's one of the largest infusions of cash into a

specific technology in the history of Silicon Valley. But what we're seeing

all this investment, what we're not seeing yet is anything approaching a revolution.

The companies tell us that AI is about to get massively better,

that this is just the beginning, and that a mega powerful artificial general

intelligence that will transform work and our lives is just around the corner.

But there's a big problem. Large language models require material written or generated by humans to train on.

And a recent study found that they might already written or generated by humans to train on.

And a recent study found that they might already

run out of training data in just the next couple years,

which means those models won't be able to get much better.

And in the meantime, all of the AI companies are burning cash.

ChatGPT loses $700,000 every day that it's running. Open AI's recent release of GPT 4O

...