AGI - AI Dynamics

Top scores on Terminal-Bench 2 went from ~25% → 75-80% in just 4 months.

For Benchtalks #1, @vincentsunnchen sat down with @alexgshaw to dig into what happens when your benchmark gets solved before you're ready for the next one.

Key takes:
→ The terminal is the right… pic.twitter.com/XkWlJT8SKT
— Snorkel AI (@SnorkelAI) 31 mars 2026

Top scores on Terminal-Bench 2 went from ~25% → 75-80% in just 4 months. For Benchtalks #1, @vincentsunnchen sat down with @alexgshaw to dig into what happens when your benchmark gets solved before you're ready for the next one. Key takes: → The terminal is the right abstraction for agentic AI → Harbor exists because benchmarking and RL at scale are infra problems → "Benchmaxxing" is real; the defense is shipping harder tasks faster → TB3 is coming, and they want your hardest unsolvable problems "We need 1000x more benchmarks than we have right now" — @alexgshaw

→ View original post on X — @snorkelai, 2026-03-31 19:21 UTC

31 March 2026

Claude’s rapid evolution toward autonomous agent capabilities

By

@karpathy

–

31 March 2026 21h07

They've released a number of features already heavily hinting on nudging Code into Claw directions, i.e. it's a speedrun of

→ View original post on X — @karpathy

31 March 2026

LLMs as CPUs: Statistical Processing and Agent Operating Systems

By

@karpathy

–

31 March 2026 20h59

LLM = CPU (data: tokens not bytes, dynamics: statistical and vague not deterministic and precise)
Agent = operating system kernel

→ View original post on X — @karpathy

31 March 2026

AI policy’s failure of imagination on employment

By

@emollick

–

31 March 2026 17h55

All AI policy is haunted by a failure of imagination. It is either nothing happens or apotheosis, people can’t seem to conceive of any other futures. It is amazing to predict massive AI development and expect nothing major to change in employment or productivity.

→ View original post on X — @emollick

31 March 2026