全部 AI 动态

https://x.com/i/article/2108757732417572864

https://x.com/i/article/2109541953155313664

以下是如何用 Claude Opus 5.5 + Blender + Unreal Engine + Meshy 构建这样一款游戏…… 诀窍是给每个工具只分配一项任务，但有一种聪明的设置方式： - 第 1 步：安装 Claude Code、Unreal Engine 5 和 Blender 5.1+，除了 Claude 订阅外全部免费 - 第 2 步：通过 Blender 的 MCP 插件连接 Blender，并通过 Meshy MCP 配合 API key 连接 Meshy - 第 3 步：先用 Opus 头脑风暴 5 个游戏创意，选一个，让它在开始构建任何东西之前先扩展这个创意 - 第 4 步：告诉它你的 GPU，这样它就能把画质推到你的机器能运行的极限 - 第 5 步：Meshy 生成道具和角色，Opus 将它们导入 Blender 并进行清理 - 第 6 步：Opus 把所有内容带入 Unreal：关卡、光照、HUD、小地图 - 第 7 步：试玩，然后每次提示添加一个功能：战斗、敌人、任务 谁负责什么： > Meshy：3D 模型 > Blender：清理、摆放、材质 > Unreal：光照、物理、实际的游戏 > Opus：在一个聊天中运行这三者 而真正的成本…… 一次完整首次构建大约花了 4 小时，以及每周 Max 订阅约 19% 的额度 把第一关做好，之后的关卡会快得多，因为资源已经存在了

You’ve seen me build all kinds of 3D websites, and many of you have asked how to create something like that yourself 👀 So we’re making it happen I’m teaming up with @tripoai and @NickDevFE, the creator of img2threejs, for a live workshop where we’ll show you exactly how to do it. Here’s what you'll learn 👇 ✴️ Generate detailed, high-quality 3D characters with Tripo AI ✴️ Reconstruct 3D models as editable Three.js code using img2threejs ✴️ Modify geometry, customize individual components, and refine model details ✴️ Rig characters, create, repair AI generated animations, and VFX ✴️ Integrate everything into an immersive, interactive 3D website using Three.js And if you attend, there’s something FREE waiting for you, along with another surprise we think you’ll love 💛 🗓️ October 12 at 14:00 UTC 📍 Live on Tripo AI’s Discord Join us: https://discord.gg/84xJc7QzW And if there’s anything you’d like me to cover during the workshop, let me know below :)

Never trust a black-box AI total on messy order sheets. Did it drop duplicate rows or refunds? WorkBuddy cleans the CSV, itemizes every gap, and gives you a searchable receipt. Bonus: message your bot from your phone to run it on your PC. 👇 https://x.com/i/article/2108890527861297152

斯坦福大学课程 CS329A: Self-Improving AI Agents 课程 9 讲视频都已发布在 Youtube，第 1 讲 52 万观看，第 9 讲 1.5 万。趁假期坚持看完 9 讲视频，你就超过了中间的 50 多万人，和连第 1 讲都没看过的无数人 😂 开玩笑，虽然 AI Agent 已经强大到很多人觉得不需要再学习，什么问题都可以直接问，到顶级大学的基础课程，系统学习下来，还是会很有收获，全局认知的提升。 课程主页： https://cs329a.stanford.edu/ 视频列表： https://youtube.com/playlist?list=PLangBM27OtEA&si=khxaTCYv_ARHBgBl

https://x.com/i/article/2108757732417572864

pstack now includes 2 skills i recommend everyone use or copy: /create-verification-skill https://github.com/cursor/plugins/blob/main/pstack/skills/create-verification-skill/SKILL.md /maintain-verification-skill https://github.com/cursor/plugins/blob/main/pstack/skills/maintain-verification-skill/SKILL.md if you don't already have one, /create-verification-skill creates a skill that teaches your agent how to run, control, and debug your app. it also creates something i call the feature map: a map of all the features in your app and how to get to it and use it from a user's pov. it allows agents to navigate and use the app just like a real user which greatly increases its ability to verify its own work. but of course, this feature map goes out of date very quickly. run /maintain-verification-skill as a daily automation with cursor cloud agents. this skill will check your app for changes and keep your feature map up to date. having a strong verification skill will end up becoming critical infra for your team and become the foundation of boosting everyone's productivity and quality and also enable more interesting automations to be built. install https://cursor.com/marketplace/cursor/pstack and you'll get auto updates for my skills. enjoy!

Asking Claude to review your code is a slot machine. Every review finds brand new bugs. Review again: more. Review a tenth time: it still finds something. It never runs out of bugs to find... so how do you decide your code is actually done?

what did you get done this week with @bot?

https://x.com/i/article/2108375208570818560

build a bot https://botpfp.com

ChatGPT launched their version of Generative UI yesterday. We took it apart tldr; the agent never write raw code. It emits DIL (a new language), the server compiles and streams it, the client renders it using a component set so that it feels native https://x.com/i/article/2108220775568404480

build a bot https://botpfp.com

AI做PPT还是得用ppt master这个开源项目啊，目前已经五万多star了 测了很多个开源项目，还是这个效果最好。无论是给Codex Claude还是Workbuddy，都很好用 开源地址 https://github.com/hugohe3/ppt-master

https://x.com/i/article/2108311949260054528

if you want to get every last bit of performance out of your local setup, you need to know a bit about inference 🥵 but we've got you covered, shipping conceptual guides 🔥 here's the first one about prefill, decode, KV cache what you should optimize for 🙌🏻

you now get Claude API credits with your MAX plans ($100, or 200 matching your plan) every month, use this to build more personal AI for yourself! I made an AI homepage that's generated everyday based on sites I read and replaces my 'new tab' page in Chrome

Our 5-person sales team was drowning before we built Rasp, an AI sales agent using OpenRouter's Ori. Every inbound lead gets researched, many hear back in under 60 seconds, and we save ~600 hours/month. Cost: ~$30/day and getting cheaper by the week. https://openrouter.ai/blog/case-studies/how-an-ai-sales-agent-saved-our-sales-team-600-hours-a-month

How do always-on, proactive agents like @Bot work? Watch my lecture at Stanford CS146S: 2:35 How we got here 6:13 What changed in the models 10:06 Inside an always-on agent 18:07 The harness 27:26 Context engineering 34:26 Where this is going

you now get Claude API credits with your MAX plans ($100, or 200 matching your plan) every month, use this to build more personal AI for yourself! I made an AI homepage that's generated everyday based on sites I read and replaces my 'new tab' page in Chrome

holy fuck. so auto compact basically loads last 1M tokens into context by default on claude? and here i thought they were summarising and reducing the context loads. no wonder people used to burn anywhere from 2%-15% 5h limits simply for auto compacting their sessions.

build a bot https://botpfp.com

You can now train your own Decision model like Jev locally! We increased Qwen3.5 0.8B’s aggregate accuracy from 20.7% to 74.3% across 3 decision benchmarks - on just 4GB VRAM. Turn any LLM like Qwen3.8, Gemma 4 into decision models with our open-source Unsloth repo. We fine-tuned with a Clef head using Unsloth and LoRA (r=64) for one epoch, increasing downstream accuracy from 30–37% to 78%. GitHub: https://github.com/unslothai/unsloth Guide and Notebooks: https://unsloth.ai/docs/basics/train-your-own-decision-model-with-unsloth