Paul Lewisohn
Showing 1–2 of 2 transcripts.
- Y Combinator49 min
Gmail Creator Paul Buchheit On AGI, Open Source Models, Freedom
Paul Buchheit, Jared, Harj, Diana, Noam Shazier, Mark Mandelbaum, Mark Blyth, Paul Lewisohn, Zuck Meyer, Melanie Warrick, Gary Illyes, Lyn Alden
Paul Buchheit and Noam Shazier trace Google's evolution from an AI-first innovator to a risk-averse monopoly that stifled tools like Lambda to protect search revenue, while OpenAI emerged as a non-profit counter-movement funded by figures like Elon Musk to keep research open. Buchheit champions open-source models as essential for preserving individual liberty against Big Tech centralization and authoritarian surveillance, predicting that algorithmic efficiency will soon lower barriers for small teams to build AGI. He warns that regulatory overreach like SB 1047 will force excessive censorship and that the future workforce will face displacement by autonomous AI agents capable of deep-faking knowledge work by 2033.
- Jane Street1h 23m
Memory Management with Stephen Dolan
Stephen Dolan, KC Sivaramakrishnan, Ron Minsky, Mark Mandelmann, Simon Hughes, Matt Sullivan, Mark Blyther, Paul Lewisohn, Simon Verstraeteker
Stephen Dolan discusses advanced optimizations in the OCaml language, detailing how generational, incremental garbage collection and software pre-fetching reduce memory access latency by up to 90%. He introduces local types and unboxed representations to enable safe stack allocation and native memory layouts, thereby eliminating heap overhead for temporary values while maintaining automatic memory safety without manual lifetime annotations. The presentation concludes with a roadmap for integrating multi-core garbage collection and lifting GC barriers to support denser data structures, balancing performance with ergonomic simplicity.