All Segments
We'll all own big AI models in the future
2 segments available
From my conversation with Nat Friedman, former CEO of GitHub. Watch the full video here: https://www.youtube.com/watch?v=qcvMjoJdck4 Or read the transcript here: https://www.dwarkeshpatel.com/p/nat-friedman Twitter: https://twitter.com/dwarkesh_sp Apple Podcasts: https://apple.co/3oBack9 Spotify: https://spoti.fi/3S5g2YK
Segments Timeline
The Future of Foundation Models
Nat Friedman discusses the future of AI foundation models, predicting a wide-scale proliferation due to their relatively simple training processes, the availability of public data, and rapidly improving hardware. He also addresses concerns about potential concentration in the market due to training costs, questioning whether investing a trillion dollars in a model will yield significantly greater value than a hundred billion dollar model.
"I want to ask you about these foundation models. What is the stable equilibrium you think of how many of them will there be? I think there will probably be wide scale proliferation. They're actually ..."
The Future of Foundation Models
Nat Friedman discusses the future landscape of AI foundation models, predicting a wide-scale proliferation due to the simplicity of training, the availability of public data, and rapidly improving hardware. He highlights that while there may be concerns about concentration due to training costs, the value difference between trillion-dollar and hundred-billion-dollar models may not justify the expense, suggesting a democratization of AI model ownership.
"I want to ask you about these foundation models. What is the stable equilibrium you think of how many of them will there be? I think there will probably be wide scale proliferation. They're actually ..."