Anthropic最近在Claude Opus 5上做了一件看似疯狂的事:移除了80%的系统提示词,而评测分数不仅没跌,反而稳如泰山。
这记耳光…
别再把模型当傻实习生哄了
Anthropic最近在Claude Opus 5上做了一个让很多Prompt工程师后背发凉的动作:移除了高达80%的系统提…
The Shelf Life of System Prompts: Why Less Is More in the Age of Opus 5
For years, we’ve treated large language models …
系统提示词的保质期到了:80%指令删掉不降分,你还在给AI当保姆吗?
最近 Anthropic 在 Claude Opus 5 上做了一个大胆的实验:…
The Shelf Life of System Prompts
For years, the indie developer mindset around AI integration was built on a foundation…
如果你还在死磕那些“保姆级”系统提示词,可能该醒醒了。
最近Anthropic在Claude Opus 5上做了个大胆的实验:直接移除了80%的系…
The Silent Tax on AI Outputs
For years, the default mindset among developers and creators was that AI models were fragi…