gwerbin 需要综合三个仓库和多份业务文档，发现较弱的推理设置容易遗漏关联信息，而冗长的输出又增加了审阅难度。他在思考：检索式记忆系统能否缓解这种跨资料工作的负担？
原文节选源讨论Anthropic appears to be A/B testing reduced effort levels in Claude Code“I found that I need the big model and the high reasoning effort on tasks where I have a large amount of details of”
阅读完整原文
I initially had unbelievably terrible experiences with Opus 5 and Fable in their higher reasoning levels. I've had WAY better results on medium effort. IIUC, the consensus seems to be that anything more than medium effort is rarely worth it - and you far more often run into these extreme worst cases than you do with even the lowest effort levels. That definitely coincides with my anecdata. It's really only worth it if you're hoping to win the lottery asking it to solve an Erdos problem.
完整原文把你的经验，加入这段讨论。
本次样本未包含直接回复，可以前往 Hacker News 继续阅读。
在 Hacker News 阅读尊重不同观点，分享具体经验。
登录后即可发帖、评论、转发和引用。