←── back to feed
/topics/claude-fable-autonomous-work-and-cost-efficiency
Claude Fable autonomous work and cost efficiency
15 items●3 sources●updated 31d ago●trend 0
Claude Fable, an autonomous AI agent, is demonstrating significant cost efficiency and capability in real-world work tasks, including writing most of sqlite-utils 4.0rc2 for $149.25 while identifying and fixing five release blockers, and building complex software projects like games and multimedia applications with minimal human direction.
- sqlite-utils 4.0rc2 mostly written by Claude Fable for approximately $149.25, including discovery and fix of five release blockers
- Fable successfully completed autonomous tasks: downloading Unity, setting up MCPs, and building WebGL games from single prompts
- Simon Willison reports token savings by instructing Fable to use judgment selecting lower-power models for coding subtasks via subagents
- Fable created multimedia project: 10-15 minute film adaptation of 'Last and First Men' using ElevenLabs and Hugging Face APIs
- Concerns about Mythos-class model cybersecurity risks validated by Fable's autonomous work capabilities, per Ethan Mollick
[BLG]blog/rss3
sqlite-utils 4.0rc2, mostly written by Claude Fable (for about $149.25)
Fable's judgement
Fable #6: The Return of the King
[BSKY]bluesky11
Somewhat humbling to have Claude Fable do a final review of some software that you're about to release and have it then find (and fix) FIVE release blockers, for an estimated (unsubsidized) cost of $149.25 simonwillison.net/2026/Jul/5/s...
Hey, Fable: "No like "AAA" in what everyone online thinks AAA is, you know what I mean."
I kept asking Claude Fable to make the game "more AAA" over and over again. The results are... interesting.
The most interesting Fable tip I've heard so far is to let the model use its own judgement as much as possible
Sent out my sponsors-only monthly newsletter for June, which means the May newsletter is now available here github.com/simonw/month... - you can sponsor me through GitHub to stay a month ahead of the free copy!
While it is obviously true that not having verifiable domains makes training models in those spaces difficult... it is also true that models are also getting much better at non-verifiable domains. The frontier is jagged, but less so than I…
Fable: "Last and First Men is out of copyright. I want you to make a movie that features a reading of it with appropriate mixes of animation and images using access to the APIs you have (elevenlabs, hugging face). Give me the first 10-15 m…
The concerns about Mythos-class models and cybersecurity were not, in fact, hype. (As anyone using Fable to do autonomous work has probably recognized)
Fable: "download Unity and set up its MCP." Done.
You really need your own benchmarks. If you are translating hieroglyphics, use Gemini 3.5 Flash. If you are running a vending machine use Opus 4.8.
Fable, one prompt: "build an elaborate game that makes it feel like I'm a brilliant chess player without knowing anything at all about chess. It should make me feel like a grand master. Feel free to go as meta as you want but the more ches…