Jeffrey Ladish

speaker

1,006 appearances 1 recordings 1 series first heard Apr 2025 last heard Apr 2025

Jeffrey Ladish’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

No recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.

Appearances

newest first · ▶ plays the moment

"The Cognitive Revolution" · Reward Hacking by Reasoning Models & Loss of Control Scenarios w/ Jeffrey Ladish of Palisade Research, from FLI Podcast · 2 Apr 2025 podcast

And then based on whether it got the answer correct or incorrect, they sort of gave it reward or sort of a down vote, like yes, more of this or no, less of this.

And very quickly you got to AI systems which were not just good at programming, but but among the best in the world.

So their their latest what this is this people called this like a reasoning model or a thinking model.

And because it's sort of been trained via via via sort of trial and error.

I think of this as sort of trial and error learning.

And you know, this is a very important part of how humans learn, right?

Oh absolutely.

If I'm if I'm in a math class

I need you know, I need to I need to read the textbook, but I also need to try a bunch of practice problems and you know, get graded and see what works.

And so

So the you know, uh a few months after O one was released, they they made a new model called O three.

There was never an O two 'cause it was trademarked.

Man, the naming of these things is crazy, but whatever.

O three got

It was better than 99.8% of competition programmers on this on the platform Code Forces, which is sort of like a platform.

It's actually at Palisade, we use this program to screen engineers.

We're like, go to this website, do a bunch of programming challenges, and we'll and the the site will grade you, and we get the score, and then we can tell how good they are based on how well they do at this at this test.

And this is the s the test where like, you know, O3 did better than

Most top open AI engineers.

And so I'm like, wow, you know, just with a little, you know, just with this a little bit of trial and error learning, on top of sort of having read all the textbooks, you now have models that are getting really good at learning by trial and error, being able to solve problems in the real world.

Showing 121–140 of 1,006 · page 7 of 51