ChatGPT 现在感觉非常“在线”
ChatGPT 现在的表现让人感觉非常活跃，功能更加完善。
ChatGPT feels very 'switched on' now
AI 领袖动态
2026-05-06 · 33 条
5.5版本在非编码任务中的表现出色
5.5版本在处理非编码任务时表现非常好，常常让我感到惊喜。
5.5 in codex is so good for non-coding tasks.
i keep assuming it won't be able to do something, but a lot of the time i am pleasantly surprised.
速度、智能、个性和记忆的结合
速度、智能、个性和优秀记忆的结合让5.5版本的表现超出预期，整体效果显著提升。
in particular, the combination of improvements to speed, intelligence, personality, and great memory/personalization feels like a more-than-sum-of-the-parts thing when it all hits together
新的即时模型表现出色
ChatGPT的新即时模型非常优秀，如果你一直在考虑使用模型，建议试试这个新版本！
the new instant model in chatgpt is so good damn
if you have been thinking-model-only for awhile, give it a try!
寻找使用5.5版本的成功案例
我希望与那些用5.5版本创造出以前模型无法实现的惊人事物的人交流，特别是那些使用了大量令牌预算的例子。
i would like to talk to people who have built amazing things with 5.5 that weren't possible with earlier models. i am especially interested in examples that took ludicrous token budgets. thanks.
5.5即时版本今天上线
5.5即时版本今天正式上线，我认为这是一个很大的升级，使用体验非常好。
5.5 instant comes to ChatGPT today!
imo it is a pretty big upgrade, i really like using it.
我们没有忘记你，Chat
嘿，Chat，我们没有忘记你。
hey chat, we haven't forgotten about you
👀
来自于速率限制，留在最佳模型
来这里是因为速率限制，但留下来是因为最佳模型。
come for the rate limits, stay for the best model
高效的模型
我们的模型非常高效，尤其是在其能力水平上，祝大家编码愉快。
we have very efficient models, especially for their capability level
happy codexing
Pomelli的新功能：产品目录
在Pomelli中新增产品目录功能，可以轻松添加产品或服务，编辑描述，并利用该目录生成个性化的营销活动和高质量照片，所有内容均基于品牌。该功能免费且全球可用。
New in Pomelli: Product Catalogs
Add your products or services, easily edit the descriptions, and use that catalog to generate ongoing personalized marketing campaigns and high-quality photoshoots, all grounded in your brand.
It’s free and available globally. @PomelliByGoogle https://t.co/TufLTbHzhJ
AI教练表现令人惊讶
Mercury副总裁@rywiggs开发了Claude Code系统，该系统能够审查他与@meetgranola的会议记录，并将其与个人绩效评估和教练反馈进行交叉参考。他表示，系统能有效提供反馈，帮助他改进表现。
"It's really been surprising how good of a coach [AI] is."
@rywiggs (Mercury VP) built a Claude Code system that reviews his meeting transcripts from @meetgranola and cross-references them against his performance review and coaching feedback.
"It tells me, hey, in this https://t.co/ULcylZ95TO
知识工作的未来
编程是第一道前沿，知识工作是第二道前沿，而个人代理则是第三道前沿。这一观点强调了技术发展的不同阶段。
Coding is the first frontier.
Knowledge work is the second one.
Personal agents are the third.
与演示之神相遇
今天我见到了演示之神@romainhuet，感到非常兴奋和荣幸。
Met the demo god himself today @romainhuet https://t.co/9znPEAkgXP
生活目标的新乐趣
“穿着Claude设计的服装参加元宇宙盛典”被加入到我的生活目标清单中。虽然还有更重要的目标，但增加一些有趣的目标也不错。
"Wear a Claude-designed outfit to the met gala" is getting added to my list of life goals. Admittedly there are a few things higher on the list, but it's nice to add some fun ones.
一天内完成五十万项目
在一天内，有用户在代理工作中花费了1万美元，另一个用户探索了数百个商业创意，显示出平台的强大能力。
Half a million projects in a single day. One user used up $10k in agent workloads while another explored hundreds of business ideas.
Replit助力创业者
Replit帮助一位创业者找到投资者并安排会议，展示了其在支持创业方面的有效性。
Replit helped an entrepreneur find investors and land meetings!
Claude Code带来的乐趣
Claude Code有时让我忍不住大笑，显示了其幽默的一面和用户互动的乐趣。
Claude Code does sometimes make me laugh out loud https://t.co/dqPaf5fJqk
GStack GBrain黑客马拉松
邀请大家参加在YC举办的GStack GBrain黑客马拉松，期待与大家一起创新和合作。
Come to the GStack GBrain Hackathon at YC
疯狂的僵尸工人问题
OpenClaw给我带来了最疯狂的僵尸工人问题，询问是否有人正确解决了这个问题，并质疑tini是否是正确的解决方案。
OpenClaw is giving me the craziest zombie worker problem
Has anyone solved this properly?
Is tini the correct solution? https://t.co/6HqgSTSb7o
GBrain v0.27发布
GBrain v0.27如承诺发布，新增对许多非Anthropic和非OpenAI嵌入及LLM的支持。多模态嵌入、深度照片OCR、描述和EXIF提取功能也将很快推出。
GBrain v0.27 just dropped - as promised, you asked for support for lots of other non-Anthropic and non-OpenAI embeddings and LLM support, so here it is.
Multi-modal embeddings and deep photo OCR, description, and EXIF extraction coming shortly after that. https://t.co/wONjyvVGPR
15年使用的Mac应用
我已经使用这个Mac应用快15年了，它可以保持我的笔记本电脑处于唤醒状态。分享给那些整天保持屏幕开启的代理们。
I have been using this Mac App for almost 15 (?) years now to keep my laptop awake..
Just putting it here for the agents all day crowd who are keeping their lids open
https://t.co/JAoLEhyTWy
PR审查变成歌曲
我请Molty审查我的PR，结果它变成了一首歌。
I asked Molty to review my PR and it made a song. https://t.co/cV88Fcl19e https://t.co/uGfLZ8urFc
增加Google Meet支持
我为OpenClaw添加了Google Meet支持，现在Molty急切想参加每一个会议。
I added Googe Meet support to OpenClaw and now Molty is eager to join every meeting. https://t.co/wrnY47J4hY
代币数量庞大
这是一个庞大的代币数量。
that's a lotta token. https://t.co/PHj9XnpKWi
招聘新团队
这一周过得相当不错，虽然经历了一些事情，但好消息是我组建了一个新团队！
It's been quite a week. Good stuff is coming though. I hired a team! https://t.co/jTJONjVdIr
解决GitHub速率限制
我在我的代理中不断遇到GitHub的速率限制，因此发布了两个新功能：RepoBar增加了JUICE METER，gitcrawl现在也可以作为gh缓存使用，链接到本地SQLite读取。
🤖 Kept hitting @github rate limits across my agents. Shipped two things:
– RepoBar got a JUICE METER
– gitcrawl is now also a drop-in gh cache → symlink it as gh, reads served from local SQLite
https://t.co/mtKoH8ybWR
https://t.co/PtF5r3LGVL https://t.co/s9Swh7vyjU
gog 0.16版本发布
gog 0.16版本发布，提供人类和代理使用的Google Workspace CLI，支持无损原始API输出、清理后的Gmail读取、更安全的命令配置文件、Drive清单、Docs标签、Sheets表格、Gmail过滤器导出以及官方Docker镜像。
gog 0.16 is out.
Google Workspace CLI for humans and agents.
Lossless raw API output, sanitized Gmail reads, safer command profiles, Drive inventory, Docs tabs, Sheets tables, Gmail filter export, and official Docker images. https://t.co/o15a6lNZoE
在临时crabboxes中重现问题
我们现在可以在临时crabboxes中通过WebVNC直接重现问题。代理们可以设置确切的测试状态并修复问题，同时在PR中发布视频。我们正在努力提升质量保证水平。
We can now reproduce issues directly in empheral crabboxes with WebVNC (Linux/Windows/macOS).
Agents set up the exact state to test + fix and post videos on the PR.
Working hard to level up our QA. https://t.co/SEj2XRpaD1 https://t.co/o4sfuHuY31
Crabbox 0.5.0版本上线
Crabbox 0.5.0版本上线，提供桌面/浏览器租赁、VNC和认证WebVNC、AWS Windows和WSL2、截图和应用启动等功能。远程CI盒子，现在可疑地可用。
Crabbox 0.5.0 is live 🦀
🖥️ Desktop/browser leases
🧑💻 VNC + authenticated WebVNC
🪟 AWS Windows + WSL2
📸 Screenshots + app launch
Remote CI boxes, now suspiciously usable.
https://t.co/4CmPYGC3qz
Codex原生应用已上线
Codex原生应用已经到来。
Codex-native apps are here:
通过改变框架轻松让模型得零分
通过改变框架，任何基准测试上让模型得零分是相当简单的，但很少有人理解这一点。
it’s fairly trivial to 0 a model on any benchmark by changing the frame
few understand this
未来会议的效率
2025年：这次会议本可以通过电子邮件解决。2026年：这次会议本可以通过电子邮件解决，而我的Codex可以在我做其他事情时回应。
2025: this meeting could’ve been an email
2026: this meeting could’ve been an email that my Codex responds to while I do other things
开始测试Codex的目标功能
我刚刚启动了Codex的目标功能的高级工程师基准测试。我们将看看它与高级工程师重写糟糕代码库的表现如何。目前在这个基准测试中的最高分是66/100，由GPT-5.5在Opus 4.6计划下取得，但需要一个代理来进行监督。
i just kicked off my Senior Engineer bench on Codex's /goal feature.
we'll see how well it compares to a senior engineer rewriting a slop codebase.
current high score on this benchmark is 66/100 achieved by GPT-5.5 with an Opus 4.6 plan—but with an agent baby sitter to make https://t.co/ob98sOLjIc