Is AI Scaling Dead? — With Gary Marcus
Listen to the original audio · Transcribed with NVIDIA Parakeet · Why times can differ in your player
What's going to happen now that AI research houses are coming up against a scaling wall? We'll find out with the leading AI skeptic, author, and founder, Gary Marcus. And I am thrilled to be joined by Gary here in studio today. Gary, great to see you. Great to see you.
Welcome to the show. Thanks for having me. So the genesis of this episode is that I did an episode with Mark Chen from OpenAI about GPT 4.5. And you come into my DMs and you say, listen, I want to give a rebuttal. Scaling is
basically over and it's not exactly what OpenAI has said. Now, for those who don't know about the scaling laws, uh, basically the idea is that the more compute and data you put into these large language models, the better they're gonna get, basically predictably, linearly. Well exponentially was the idea. Right. And so the context here is
We already have this episode, so unlocking it is cheap and permanent: the full transcript (53 minutes, 10,876 words), segment and word-level timestamps, and .txt, .srt, .vtt and word-level JSON downloads, as many times as you like, forever.
No account needed, we create one from your checkout email. With credits it is 1 credit; packs start at 5 episodes for $3.49, and monthly plans are cheaper still.