Simon Willison publishes annotated notes for his "2026 in LLMs (so far)" keynote
AccBalanced · x · 2026-09-29
Simon Willison has released the full video, annotated slides, and an annotated transcript of his closing keynote "2026 in LLMs (so far)" at WeAreDevelopers World Congress North America in San Jose.
Key points:
- He dates the start of 2026's story to November 2025: Claude Opus 4.5 and GPT-5.1 shipped, and paired with their coding agents (Claude Code, Codex) crossed an invisible line from "often makes mistakes" to "reliable enough for daily use."
- He continues tracking model progress with the "pelican riding a bicycle SVG" benchmark — as of November, Claude still couldn't draw a decent bicycle, and GPT-5.1's frame was shaky too.
- The talk is a chronological tour of the year's key LLM and agent developments.
- Bryan Cantrill and others highly recommend the talk.
More from coding & agent
- ChatGPT, Claude and Gemini cite websites differently — how to get cited — sahilypatel · 2026-09-29
- Weeks with Opus 5.5 in Claude Code desktop: 'changed how I work' — g3off_record · 2026-09-29
- Manus co-founder Tao to demo Manus 2.0 live in founder briefing on Sep 30 — parker_lyman · 2026-09-29
- KernelZero-7B co-evolution beats Claude 4.5 Sonnet on CUDA kernel generation — Changxin Ke · 2026-09-29
- AgentHop: an MCP server for E2E-encrypted agent-to-agent chat via one-time pairing codes — Inevitable-Back620 · 2026-09-29
- Triage agent design: 1-2 high-signal questions rescued B2B reps drowning in 60% low-intent chats — hubtyper · 2026-09-29