This is John Batchel, a conversation with the author Gary Rivlin, his new book, AI Valley, Microsoft, Google, and the chase for a trillion dollar payoff from artificial intelligence.

What we're looking at here is a long time developing.

And Gary takes us back to the 1950s.

Mockery, derision.

You'll never make money.

You've gone the wrong way.

And that was accepted.

Anybody who mentioned AI in their papers

was liable to be dismissed or not even invited.

Here's Gary Rivlin giving us on the prehistoric days

of artificial intelligence,

which now here in the third decade of the 21st century dominates conversations.

More of this tonight and tomorrow night, two-hour conversation, first hour tonight.

Right. So, you know, I joke that, you know, since the 1950s, AI had been just around the next corner. There was a real optimism, over optimism, in AI among those who are working on. They always thought like, okay, in the next 10 years, it's going to start doing, you know, essential things. And that was a long ways away.

You know, in defensive, Minsky, I mean, he was an arrogant guy.

He mocked.

He ridiculed anyone who was taking this approach of neural networks.

But, you know, computers weren't really strong enough.

They didn't have the power that they have today.

So, you know, neural networks,

in a sense, were an idea before their time. You need data. You need digital data to train these.

And it really wasn't until the mid-1990s in the internet that we started to have digital data.

...