The Evolution of Artificial Intelligence

The ability to sort of get a holistic idea of what you're wanting to do, even at the larger prom size is very, very good. And so I think that whatever their strategy is, at least at the moment, in my opinion, it's working better. The other link you sent me during the week was on the subreddit machine learning, uncensored models fine-tuned without artificial moralizing. Has there been any studies about how censorship handicaps are models capabilities? We discussed this last week without actually turning it into text. You've got that side of things. Then you've got models which are trained on relatively smaller denser amounts of data and not going the large.

Transcript

Play full episode

Transcript

Episode notes

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!

Get the app