What comes after AI?
At the 3rd annual Caltech–Broadcom Symposium, Broadcom's @CharlieKawwas joined @Caltech leaders at the new Ginsburg Center to explore quantum computing, agentic AI, and the critical role of photonic connectivity in scaling next-gen compute.
Read the recap: https://t.co/eM6nbTz2le
At the KITP workshop on AI for Physics, I was asked to review one of 722 OpenAI papers.
The argument turned out to be fairly elementary. I could probably explain it to physics undergraduates in an hour.
There are some interesting steps, however.
https://t.co/913bp5qEBK
I was brought to tears on Sunday when the last milestone cleared. My father, a mathematician who passed in Jan, introduced me to the giants of group theory as a kid. CFSG defined my 80s college years. I can't believe I somehow got to be part of this human achievement.
This is a podcast about whether AI could discover the laws of physics before we do. Hirosi Ooguri @PlanckScale, chair of Caltech’s Division of Physics, Mathematics and Astronomy, joins for his first English podcast to explain why AI has moved faster in mathematics than in physics. He fears it could solve physics in a way we cannot understand. We discuss why quantum field theory lacks a firm mathematical foundation, quantum gravity as the “final boss,” the swampland distance conjecture he proposed with Cumrun Vafa, and his estimate that, left to humans, verifying string theory could take 2,250 years. The conversation also covers the PhD thesis numbers that turned up in a Mathieu group table twenty years later. What an honor.
https://t.co/ZwSg1eVs4B