“不,我发誓这是我写的。”
"No, I swear I wrote this."

原始链接: https://revise.io/blog/06-27-2026/no-i-swear-i-wrote-this

大语言模型已经变得非常强大,现在根本无法分辨内容究竟是出自人类还是模型之手。所谓的“AI 检测器”不过是江湖骗术,人与人之间已丧失信任,现状糟糕透顶。 Revise 的使命是将 AI 优雅地融入写作过程。为此,Revise 编辑器现已支持文档的全流程“回放”功能。我们的历史记录系统用途广泛,而今天又增添了一项新功能:你可以将文档分享给他人,让他们像看 YouTube 视频一样,按顺序观看你的写作、重写和排版过程。 砰!这便是最确凿的“键盘输入证明”,足以证实某段内容确实是由人类亲手敲击键盘创作出来的。

抱歉。
相关文章

原文

LLMs have gotten very good. It's impossible tell anymore if something was written by a human or a model. AI detectors are snake oil. Nobody trusts each other anymore. It's bad out there.

Revise, in its mission to gracefully integrate AI into the writing process, now has full-history "replays" for documents written in the Revise editor. There are many ways our history system is useful, and today there is a new one: you can now share a document with someone and let them watch you write it, rewrite it, reorganize it, all in sequence like a YouTube video.

Boom, definitive "proof of typing" that a given piece was produced by genuine meat-on-keyboard effort.

联系我们 contact @ memedata.com