Fable's judgment: efficiency tips Simon Willison learned from the Claude Code team
Original title:Fable's judgement
AI summary
After speaking with the Claude Code team at AIE, Simon Willison suggested letting Fable (and Opus) work with their own judgment rather than hard-coding behavior. For example, simply letting Fable decide when to write tests works better than giving specific rules. To cope with upcoming price increases and save Fable tokens, another tip from Jesse Vincent is to tell Fable to delegate smaller tasks to lower-power models (Sonnet for substantive implementation, Haiku for mechanical changes), while keeping judgment, auditing and data synthesis tasks in the main loop. Willison has stored the prompts in Claude Code memory files, with good practical results and a noticeably slower rate of Fable token consumption.
Key Takeaways
- After speaking with the Claude Code team at AIE,Simon Willison suggested letting Fable (and Opus)…
- For example,simply letting Fable decide when to write tests works better than giving specific rules.
- To cope with upcoming price increases and save Fable tokens,another tip from Jesse Vincent is to…
- Willison has stored the prompts in Claude Code memory files,with good practical results and a…
Quick Facts
- Published
- Updated
- Source
- Simon Willison 博客
- Category
- Tutorials & Tips
- AI score
- 73
Simon 从 Claude Code 团队得到的实战技巧:别硬性规定 Fable 怎么写测试、用哪个模型,让它自己判断。他实测这条 prompt 能明显节省代币消耗,Fable 涨价前偷时间的利器。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Fable's judgment: efficiency tips Simon Willison learned from the Claude Code team. July 4, 2026. https://new.nat6.net/en/news/cmr5blxyu00ykslc7kiaaxqam/