一个挺实际的判断:智能体操作电脑现在比人慢,与其盯着看,不如租几台便宜的云机器让它们并行跑,思路可以参考。
#computer use
OpenAI 出了 GPT-6.1 Terra,能看懂屏幕、直接帮你操作各种软件,价格还比之前便宜,干整天活更扛得住。
Ethan Mollick 分享了一条给 Codex 电脑操作用的提示词,做智能体操作的朋友可以直接拿去试试。
The tutorial lowers the entry barrier for running a quantised computer use agent locally, showing that NVFP4 weights can drive real browser automation rather than sitting idle on disk. Developers building browser agents can follow the OpenEnv, Helium and Hugging Face Sandboxes stack to replicate the setup. It also illustrates how open-weights models are closing in on proprietary computer use offerings.
Ethan Mollick 发推提醒:AI 已经能替你操作电脑了,很多老套路得重新想想。适合关注 computer use 方向的人。
沃顿教授 Ethan Mollick 发了条短评,说现在 AI 都能替你操作电脑了,有些老套路撑不久。适合关注 agent 安全的人看看讨论。
The rotation strategy shows how heavy users stretch AI subscriptions beyond single-account limits, and tier selection between medium and high reasoning changes both speed and quality. Practitioners paying for multiple coding subscriptions can copy the setup, including routing Codex's computer use through Claude. The spending arc, from a S$20 plan to two US$250 accounts, is itself a datapoint on how fast coding-assistant budgets scale.
The practical detail is the tier-splitting pattern: pairing fast and high reasoning modes within one subscription to stretch quota, then offloading capabilities another product does better through a bridging tool. Developers burning through Claude or Codex limits can copy this setup to run coding, device control and multi-environment verification on a single desk. The Parallels-and-remote-Windows arrangement is especially relevant for anyone building cross-platform software from a Mac.
Browser Use 团队把 Mac 自动化操作工具 MacOS Harness 更新到 0.2,加了虚拟光标,还能在控制浏览器时自动切到 Browser Use,玩智能体操作电脑的可以试试。
博主现身说法吐槽 ChatGPT 干活要人催、代码打补丁、computer use 还费 token,考虑过真实使用体验的可以对照看看。
写 Claude 智能体的可以升级 SDK 了,点击、打字这些循环不用再手写,SDK 直接帮你跑,省不少胶水代码。
OpenAI 拉上法务软件公司 Ironclad,专门练 AI 操作电脑处理合同,做 computer use 的可以看看他们的评估思路。
OpenAI 的 Codex Desktop + Astra 是我目前用过最强的电脑操作 Agent,作者把为什么要用 GUI 当通用接口讲得很透,做 Agent 的都该看看。
HyperWrite 创始人 Matt Shumer 发了条推文,说 Claude Opus 5.5 若配上 Codex 那种电脑操作能力就没对手了,看看业内人怎么比这两家的长处。
Tw93 把 Mac 上 611 款软件的 AI 自动化实测做成了可查清单,关键发现是无障碍属性能让 AI 定位元素更准,独立开发者很该看看。
Mole's improved uninstall functionality benefits Mac users by thoroughly removing software residues, with AI handling complex installation and cleanup processes.
朋友遇到视频播放卡顿的 bug,ChatGPT 的 computer use 能帮你定位问题并解决,比你自己摸索快多了。