@imedwardluo 谢谢回复！另外，刚刚cursor更新了，有个agents window的界面，看着和codex desktop比较像。或许有一些codex的feature适用于cursor。
中文
Aoran Wang
291 posts
Aoran Wang
@wang_aoran
Researcher @ Shanghai AI Lab | @DAAD_Germany AInet Fellow | Previous @SquirrelAi_Edu @BoschGlobal @uni_lu @KITKarlsruhe @TongjiUni
Esch-sur-Alzette, Luxembourg Katılım Aralık 2018
500 Takip Edilen146 Takipçiler
@wang_aoran Cursor 暴露的权限审批不会自己分级，所以会有这种情况，我一会更新一个版本会优化一下这个情况：
1. 应该会更加适配你设置好的权限逻辑，
2. 提供一个设置开关如果还有额外审批的问题，可以在设置里面把 cursor 的关掉。
感谢提出反馈哈哈！
中文
最难过的是原来那条帖子上有好多大家反馈的建议、bug、新需求我都还没来得及回，如果大家有任何问题欢迎直接 DM 我或者在这里回复，我都会一个个看的 😭
Edward Luo@imedwardluo
可能是这几天熬夜没睡觉做产品脑子做傻了...想补一条 Vibe Island 新帖子的时候，不小心点错删掉了自己有生以来最火的一条 30w+ 的帖子... 😭 发之前完全没预料到帖子会这么火，还有这么多大佬都自发帮我转发，真的非常感激各位大佬，也非常感谢大家喜欢我的产品，容纳它还有很多瑕疵和小问题。我在疯狂修 bug 优化中，很快就会上线更多有趣有用的功能！ 只保留了一张截图纪念，原贴内容如下 👇🏻
中文
Codex 还是很讲武德，虽然这周没有被「恶意」reset，但 rate limit 0% 时，还是会把最后一个 thread 里的 session 跑完，不会主动中断，这就让我不用担心接近 0% 时任务中断不容易复原
这种做法，建议所有 AI Agents 普及！ @OpenAIDevs
中文
New site to InnoClaw: github.com/SpectrAI-Initi…
Aoran Wang@wang_aoran
Still working on it to make it better. 🎉🎉🎉
English
Still working on it to make it better. 🎉🎉🎉
GitHub Projects Community@GithubProjects
What if your filesystem became an AI research engine? InnoClaw does exactly that: • Self-hosted AI assistant • Chat with your files using RAG • 200+ scientific tools built-in Not just answers actual research workflows
English
@CVPR @openreviewnet Please check the site as it keeps crashing.
English
Here we go again ...
Qwen@Alibaba_Qwen
🚀 We're thrilled to unveil Qwen3-VL — the most powerful vision-language model in the Qwen series yet! 🔥 The flagship model Qwen3-VL-235B-A22B is now open-sourced and available in both Instruct and Thinking versions: ✅ Instruct outperforms Gemini 2.5 Pro on key vision benchmarks ✅ Thinking achieves state-of-the-art (SOTA) performance on multimodal reasoning tasks ✨ Key breakthroughs: 🖥️ Visual Agent: Operates GUIs on PC/phone — understands buttons, calls tools, and completes real-world tasks (SOTA on OS World) 💻 Visual Coding: Transforms screenshots into code (HTML/CSS/JS, Draw.io) — true "what you see is what you get" development 📚 256K+ context (scalable to 1M) → supports 2-hour videos and multi-page long PDFs 🌍 32-language OCR with enhanced robustness for blurry, tilted, or rare characters 📐 Advanced spatial reasoning: 2D → relative coordinates, 3D grounding, occlusion handling, and perspective understanding 🧠 Thinking Mode: Leading performance in STEM/Math — enables deep causal reasoning 🔤 Text capabilities rival top-tier LLMs — a solid language foundation powering its multimodal excellence From "seeing" to "understanding", from "recognizing" to "reasoning & acting" Qwen Chat: chat.qwen.ai/?models=qwen3-… API: #5540e6e52e1xx" target="_blank" rel="nofollow noopener">alibabacloud.com/help/en/model-…
Blog：qwen.ai/blog?id=99f033… ModelScope: modelscope.cn/collections/Qw… HuggingFace: huggingface.co/collections/Qw… English
too many things to release. don't know how to plan. alright, let it be.
English
@yann @Ming @jiajun @Katerina @tali @Lerrel @jack @nicklas @sherry @boyi @yilun @prithvijit Let’s build worlds that work — physically, interactively, and reliably.
SEE YOU THERE! 🌐🚀 #ICCV2025 #WorldModels #EmbodiedAI #RIWM #Simulation2Reality
English
🌍✈️ Ready to explore the real and interactive future of AI?
Want to go to Hawaii??
Let’s gooo to ICCV 2025 for the first ever RIWM workshop!! 🎉
🚨 Reliable and Interactable World Models is coming! 🚨
——-Threads——
English
🚨 We discovered a surprising side effect of Reinforcement Finetuning (RFT): it makes LLMs more confidently wrong on unanswerable questions.
We call this the hallucination tax: a drop in refusal behavior that leads to overconfident hallucinations.
🧵 1/n
English
@wang_aoran @icmlconf Wow that’s really great! Congrats! I think ARR should follow this as well
English
