And what do you see as that kind of like as the tension there? Or do you think that is a false choice? Do you think that's the wrong way to think about like work? where and why and when to implement kind of an AI-driven or an agentic process, if you will?

So I think the field right now is too immature for people to be trusting AI to do the jobs that many companies are asking the AI to do.

And I think customer service is a perfect example.

These AIs are making mistakes, to your point.

There is liability, and some of these mistakes have made companies look bad, terrible, in fact, in the media.

Companies have looked at the AI as a way of cutting costs, and in some cases have looked fired or let go of all the people in their customer service and brought in AI agents.

And as your listeners obviously know, there's a reversed trend there.

but

Mm-hmm

... program that's gonna give me all the answers, and clearly that's wrong.

But the trust level between the student and the program was such that they weren't questioning the results.

They believed it just because this is, like, amazing.

And that was unsettling because as a researcher, I know how the algorithm works, and I know that it can't be trusted to be accurate.

What we've seen since then, and this will sound rather cynical, and the AI companies probably don't like the word I use, but, uh, they've applied a lot of Band-Aids.

So the errors are still there, but they put on all these extra, um, patches, fixes to cover scenarios that people point out that are embarrassing.

Mm-hmm.

Mm-hmm.

A lot of things that people are doing today I couldn't do with Chinook because in order to announce that the game of checkers was a draw, there had to be zero error, no possibility of error.

Now, there was one point where I had a file that got corrupted, and it turned out there was one bit that had got corrupted in a f- in a file.

And going back in the logs, the bit got corrupted two years ago.

Well, calculations build on top of calculations, and I could not tolerate any error, so all those computations had to be redone.

Today's AI has a tolerance for error.

ChatGPT has educated us, I put the word educated in quotes, that it's okay to have errors.

Terms and conditions apply.

And so I guess obviously my first question is, why? What would a reset look like?

First of all, we're rushing headlong into something that we don't know and we don't understand.

AI is unlike any other technology we've seen before.

It impacts everything.

We're also at a point where we're in this hockey stick phenomenon where technology, new technology is being invented literally every day.

Exponential growth.

And we know... already that there were problems that AI is causing.

Maybe you're just holding it to too high of a standard.