Fable 5.1 hands-on: faster and cheaper, but "coding is solved" rings hollow
vboykis · x · 2026-09-03
Justin Duke published a hands-on review of Anthropic's Fable 5.1, which he tested via Conductor and the Claude Code web interface. His verdict: it's faster and cheaper than Fable 5 for the work he does.
But he has two reservations:
- Anthropic couldn't articulate what kind of work Fable 5 couldn't do that the new model can — the boundary between "LLM can do this" and "LLM cannot" feels amorphous, suggesting no clear step-function improvement
- The product experience is still rough: after already trying the model, both the Claude web app and mobile client bombarded him with full-screen modals promoting Fable 5.1, despite Anthropic having the data to know he'd already used it
His core argument: every time a frontier lab employee claims software engineering is a solved problem, he hears it like someone saying cooking is a solved problem. vboykis added in her retweet that writing code is barely solved, and "walking around the app" is missing from the RL training loop entirely.
Related event: Fable 5.1 Review: Faster and Cheaper, But Coding Not 'Solved'(2 posts)→
More from coding & agent
- Cursor's harness beats VS Code, but hooking it to local Ollama models needs workarounds — walkingriver · 2026-09-03
- Baseten ships remote MCP server and skill, cutting coding agents' token usage by up to 57% — baseten · 2026-09-03
- davidad: rollout rubrics in verifiable domains should check whether the model ran the verifier — davidad · 2026-09-03
- Dev's Claude Code sales plugin sent 1,097 emails in 85 days: 0 paid — No_Job_9995 · 2026-09-03
- Ripple Edit: a vibe-coded ComfyUI extension that moves node groups to free up space — GeroldMeisinger · 2026-09-03
- Context windows grow, agents drift: PetsApp CEO to demo prompt ablation live — dscape · 2026-09-03