Hello and welcome back to Inside Tech.
I'm Tom Standage, deputy editor at The Economist,
and with me as usual is my co-host Alex Hern, our AI writer.
Hello, Tom.
Last year, there was a lot of enthusiasm about artificial general intelligence, or AGI,
which is roughly when an AI can do everything better than humans can.
A lot of people in the field felt that some kind of breakthrough
and a big leap in AI's capabilities was within reach.
In recent months though, I think it's fair to say that
there's been a cooling off of enthusiasm about AGI
and a recognition that if you want to make AI more capable and more reliable,
then simply scaling up existing approaches is not enough on its own.
Further innovations and some kind of change of approach is going to be needed.
And lately, there's been a lot of excitement around one approach in particular known as world models.
And that's what we're going to talk about today.
So Alex, to kick things off, I've got some examples of recent statements about world models
and how wonderful they are from luminaries in the field.
So first off, here we've got Yann LeCun.
So obviously, he's just left Meta where he was the chief AI scientist,
but he's one of the three godfathers of AI.