AgentLens scores the whole coding-agent trajectory, not pass/fail.

Simon Willison released LLM / Shared entity: LLM / Earlier coverage / Tension

Simon Willison's Datasette Apps: Claude Artifacts backed by a real database.

Linked by a graph relationship (Simon Willison released LLM); both cover LLM; earlier LLM coverage from 2026-06-19.

Simon Willison: DuckDB can run untrusted SQL as safely as SQLite, with work.

Linked by a graph relationship (Simon Willison released LLM); both cover LLM; earlier LLM coverage from 2026-06-18.

LLM uses OpenAI / Shared entity: LLM / Earlier coverage

Simon Willison just called GLM-5.2 the most powerful text-only open-weights LLM, and he has the receipts

Linked by a graph relationship (LLM uses OpenAI); both cover LLM; earlier LLM coverage from 2026-06-19.

Simon Willison released LLM / Shared entity: LLM / Earlier coverage

Simon Willison says AI agents can't be accountable

Linked by a graph relationship (Simon Willison released LLM); both cover LLM; earlier LLM coverage from 2026-07-14.

Simon Willison ships sqlite-utils 4.0rc1 with schema migrations and nested transactions.

Linked by a graph relationship (Simon Willison released LLM); both cover LLM; earlier LLM coverage from 2026-06-22.

The silent safeguards. For the first time, a Claude model will degrade its own answers and not tell you.

Linked by a graph relationship (Simon Willison released LLM); both cover LLM; earlier LLM coverage from 2026-06-10.

LLM 0.32a3 adds tool-call pause/resume, and Fable 5 wrote almost all of it.

Linked by a graph relationship (Simon Willison released LLM); both cover LLM; earlier LLM coverage from 2026-06-10.

Zig formally bans all AI-generated contributions.

Linked by a graph relationship (Simon Willison released LLM); both cover LLM; earlier LLM coverage from 2026-04-30.