The Gradient

Peter Lee on how simple ideas underlie practical systems across computing fields:

I really enjoyed this conversation with Peter Lee, President of Microsoft Research. We spent time on his early work in compiler generation and type theory, Microsoft’s “Sparks of AGI” paper and its claims, and how he thinks about his work’s impact.

I spoke with the legendary Manuel and Lenore Blum, renowned for their contributions to theoretical computer science and mathematics, about their early influences and their current work on the Conscious Turing Machine.

I spoke with Kevin Dorst, Professor of Philosophy at MIT, about why your (political) opponents might be more rational than you think.

David Pfau on the deepest kinds of research questions: "if you can discover something where you make people ask a question they didn't even know to ask in the first place, that's the most exciting thing for me."

"The deepest ideas come from studying simple stuff, and the most impactful ideas are the deepest ones" — David Pfau on finding the ideas that scale up and why his own work on Geomancer isn't yet finished.

I had a great conversation with David Pfau, Staff Scientist at Google DeepMind and one of the most thoughtful researchers I’ve ever spoken to, about: - Doing impactful research - Deep learning for computational quantum mechanics - Representation theory and learning to factorize manifolds

I spoke with Dan Hart and Michelle Michael of the NSW Department of Education about: - Developing NSWEduChat, an AI-powered chatbot - Delivering effective education to students - Developing technology with speed and effectiveness, *while* implementing governance measures

I spoke with the brilliant Kristin Lauter, who leads AI at Meta FAIR Labs North America, about: - Elliptic curve cryptography - Cryptographic protocols - ML on encrypted data - Attacking post-quantum cryptography with AI I admire her work a lot, so this was an honor. Enjoy!