Dr. David Eagleman

It's very different is the point I'm making.

They both have converged on something that we would call intelligence.

But it's a pretty different structure.

even though AI was inspired by the brain.

Yeah, but that's interesting because Hinton is incentivized to say that, but a neuroscientist... He's incentivized to say that.

People doing AI, of course, are paying a lot of attention to how this is structured like the brain, because before that, people would do things like probability theory or rules.

You know, they would try to do AI by trying to say, OK, if this, then do that.

But when people start doing artificial neural networks, that led to a lot of success.

I'm only pointing out that the artificial neural network looks a lot like the brain on the surface.

You say, hey, you've got units and you've got connections.

But beyond that, there's a lot of differences.

And why are those differences significant as it relates to what's possible?

Because what we've developed is this new species, essentially, that is incredibly impressive, but it ain't a human brain.

It's different than a human brain.

There may be all kinds of similarities, things that we even come to understand are similar, but there are so many differences.

Here's an example.

You know, we humans do one-trial learning all the time, meaning if I say – or when you were a kid and your mom said, hey, Steven, this is a pomegranate, you say, okay, pomegranate, got it.

But you can't – when you're training up an artificial neural network –

like at OpenAI or Gemini or Anthropic, you have to give thousands or millions of examples of everything for it to learn anything.

There's no one trial learning on those systems.