AI Reasoning Models
topic on 5 shows · 8 statements across 7 episodes
the Knowledge Project 
Latent Space 
No Priors 
the MAD Podcast 
the a16z Podcast
8 statements about AI Reasoning Models, every show
Dubois: RL allows AI reasoning models to backtrack wrong paths earlier
“Part of it is the model knowing when it's going down the wrong path. But this is also something that we can that the model can be trained for with reinforcement learning is like knowing, okay, like that seems like not a great path. Let me backtrack and let me …”
George: Legal AI startup Harvey sees user time double
“Users are spending about double the amount in the product as they had before.”
Noah Smith: AI reasoning models hallucinate endlessly and gaslight users
“I mean, the reasoning models still go off in these crazy hallucinations that they'll never, like, admit were wrong, and will just, like, gaslight you infinitely on some crap it made up. Like, they'll still, like, just knowing truth from falsehood.”
Dwarkesh Patel: AI models hallucinate less and are more reliable than humans
“Do they hallucinate more than the average person? I think, like, no, less. They can hallucinate, meaning, like, getting something wrong, and when they push them on it, they're like, no, you know, whatever, and eventually they'll, like, accede if there's, they'…”
Garry Tan: AI reasoning models will drive biotech breakthroughs within a year
“Like, if only by applying, you know, dollars to energy that then goes into these models, will there be, ah, real breakthroughs in, you know, biological sciences, like being able to do new processes or come to a deeper understanding of you know, whether it's, a…”
Ulrich: AI reasoning models increasingly decline to answer when uncertain
“It's not just the accuracy that is improving, but the percentage of time, the significant increase in percentage of time that those models will basically say, I don't know the answer to a question.”
McAteer: AI reasoning models are probably already superhuman at scientific synthesis
“I think that I think it's hard for one person to take in so much data, so much information and like make connections. And I think that's definitely one of the areas where these models, they're probably already superhuman at doing that.”