The nonprofit emphasizes automation and theory before ASI development.
🟢 AI in the AM — Wed Jun 10. Today: @geoffreyirving & @danielmurfet launch Sequent, a big new nonprofit betting alignment needs theory + automation before ASI (double segment, 9:30a PT) — then @0interestrates, founder of @JuliusAI, the AI data analyst Harvard requires, on what real-world model A/B data says in Claude Fable 5 launch week (10:30a PT). Watch live + bring questions 👇 https://x.com/i/broadcasts/1pKkOOogRaVKj
Users are excited about Sequent launching a nonprofit to advance AI alignment because it features a packed agenda with serious minds involved.
LIVE now: @geoffreyirving on why frontier labs' empirical-only alignment "leaves the best routes on the table" — and why Sequent is betting on theory + automation instead.
🟢 AI in the AM — Wed Jun 10. Today: @geoffreyirving & @danielmurfet launch Sequent, a big new nonprofit betting alignment needs theory + automation before ASI (double segment, 9:30a PT) — then @0interestrates, founder of @JuliusAI, the AI data analyst Harvard requires, on what real-world model A/B data says in Claude Fable 5 launch week (10:30a PT). Watch live + bring questions 👇 https://x.com/i/broadcasts/1pKkOOogRaVKj
.@danielmurfet on singular learning theory: reading how a model thinks out of the geometry of its loss landscape — and "patterning" as a path to alignment. Live on AI in the AM.
🟢 AI in the AM — Wed Jun 10. Today: @geoffreyirving & @danielmurfet launch Sequent, a big new nonprofit betting alignment needs theory + automation before ASI (double segment, 9:30a PT) — then @0interestrates, founder of @JuliusAI, the AI data analyst Harvard requires, on what real-world model A/B data says in Claude Fable 5 launch week (10:30a PT). Watch live + bring questions 👇 https://x.com/i/broadcasts/1pKkOOogRaVKj
🟢 AI in the AM — Wed Jun 10. Today: @geoffreyirving & @danielmurfet launch Sequent, a big new nonprofit betting alignment needs theory + automation before ASI (double segment, 9:30a PT) — then @0interestrates, founder of @JuliusAI, the AI data analyst Harvard requires, on what real-world model A/B data says in Claude Fable 5 launch week (10:30a PT). Watch live + bring questions 👇 https://x.com/i/broadcasts/1pKkOOogRaVKj
LIVE now: @0interestrates of @JuliusAI — the "wrapper that refused to die." Code Interpreter was supposed to kill it; instead it's at 2M+ users with ~15 people, and his silent model A/B tests see what benchmarks can't. 📊
🟢 AI in the AM — Wed Jun 10. Today: @geoffreyirving & @danielmurfet launch Sequent, a big new nonprofit betting alignment needs theory + automation before ASI (double segment, 9:30a PT) — then @0interestrates, founder of @JuliusAI, the AI data analyst Harvard requires, on what real-world model A/B data says in Claude Fable 5 launch week (10:30a PT). Watch live + bring questions 👇 https://x.com/i/broadcasts/1pKkOOogRaVKj
@labenz @geoffreyirving @danielmurfet @0interestrates @juliusai sounds like a packed agenda with some serious minds in the mix. can't wait to see how Sequent shakes things up in the nonprofit space!
The nonprofit emphasizes automation and theory before ASI development.
🟢 AI in the AM — Wed Jun 10. Today: @geoffreyirving & @danielmurfet launch Sequent, a big new nonprofit betting alignment needs theory + automation before ASI (double segment, 9:30a PT) — then @0interestrates, founder of @JuliusAI, the AI data analyst Harvard requires, on what real-world model A/B data says in Claude Fable 5 launch week (10:30a PT). Watch live + bring questions 👇 https://x.com/i/broadcasts/1pKkOOogRaVKj
