love the world

USA
Joined August 2018
mjboy.eth retweeted
太疯狂了!我用 Opus 5.5 做了一款有完整剧情的动作游戏,只花了一天。 游戏场景长安城全部由 Opus 5.5 生成。 最让我惊喜的是角色制作:只要一张图,PINOC (JST 2) 就能直接生成 3D 角色,骨骼自动绑好,不用做任何额外操作。 豆姐、梁子、老马、提伯、阿迪王这五个角色,先用 Codex 画了卡通形象图,再把图交给 PINOC,就拿到了能直接动起来的 3D 角色。全程没有建模,也没有绑骨骼。 走、跑、冲刺、刺拳、回旋踢、翻滚、被击倒、起身、胜利……这些动作都是从 PINOC 的免费动作库里挑的,直接套到角色身上就能用。也可以自己设计动作,游戏里有个“梁子结算舞蹈”的彩蛋,我直接上传了一段视频到PINOC,直接就给我生成了动作,比想象中的简单太多了! 制作流程也非常简单,直接分享出来给大家: 1. 让Claude Code设计好游戏规则和风格; 2. 让Codex根据提示词设计好角色图片; 3. 安装PINOC MCP viggle.ai/pinoc/app/mcp,让Claude Code通过PINOC MCP生成3D资产,并完成整个游戏的制作。 4. 让Claude不断调整和优化细节、玩法,比如根据场景生成开场动画、用TTS生成语音等等。
Today we are releasing JST-2, the world's first full-stack physics-native world foundation model. Many world models learn in pixel/video space. We went back to first principles and built the entire model stack in 3D. Generate a 3D scene in seconds and expand it in real time. Create characters with motion built in, all through one general architecture.
23
36
1
301
21,304
A BIG THANK YOU to our legendary friend, @ElonMusk, for his incredible generosity in giving SpaceX Stock to millions of American children through TRUMP ACCOUNTS! This is exactly what was envisioned when we created TRUMP ACCOUNTS — Bringing together the GREAT PRIVATE SECTOR, Elon and so many more, to invest in our children, by far our greatest and most important asset, and give every American Child a piece of the American Dream. Thank you Elon! President DONALD J. TRUMP
6,838
20,043
1,676
172,735
8,931,397
mjboy.eth retweeted
Elon did it all in one life
289
2,363
206
29,389
503,361
Hi, I'm Felix. I'm an eng lead working on Claude. I love watching people take flight and do things they never thought possible. We often describe that as "being more ambitious" when using AI, which is tricky advice to follow. I wrote a tutorial: felixrieseberg.com/how-to-be…
395
2,466
549
23,763
5,218,046
mjboy.eth retweeted
花费16天,我用 Opus 5.5 将小王子的故事制作成了3D剧情游戏 结合Tripo+Blender,Opus 5.5 将小王子故事中的每一个星球变成了可探索的3D场景 希望大家能支持一手,后续会开源整个游戏,以及详细制作流程 Welcome to The Little Prince: Across the Stars✨:littleprince.dev/
1,390
1,713
857
16,787
3,722,657
Trump walks out at day 3 of the RNC to “It’s a Man’s Man’s Man’s World” by James Brown. Unmatched aura. Biden could never.
88
829
55
5,705
175,771
AI is already smart enough. Real-world tasks are still slow, expensive and unreliable, because we hand AI a computer built for humans, then wrap it in a heavy harness. AI doesn’t need to get smarter. It needs a computer built for it. Today we’re releasing Pine Computer
Readers added context they thought people might want to know
Pine's highlighted SaaS-Bench v1.1 results show higher checkpoint score (78.3%) but lower resolved full-task completion (27.4%) than comparators (29.2%, 31.1%). Pine Computer is in private beta, not a full public release. unipat.ai/benchmarks/Saa… pinecomputer.io/blog/introduci…
569
745
360
8,360
24,289,448
mjboy.eth retweeted
This kid did an amazing job with the frame-by-frame dance.
378
8,562
564
80,065
4,426,779
mjboy.eth retweeted
I asked Opus 5.5 to make an interactive website companion for the latest @AcquiredFM Home Depot episode. It came out pretty nice! Water color illustrations all drawn by Claude, too: claude.ai/artifact/RdE6uE7Wo… Episode here: acquired.fm/episodes/home-de…
129
88
23
1,888
306,850
mjboy.eth retweeted
POV: you switched to a low reasoning effort model
190
1,975
170
36,216
1,407,147
makes him more endearing 😭
🤖 Made with AI
watch dario, but imagine he is playing with a slinky
82
305
56
5,223
461,096
mjboy.eth retweeted
Readers added context they thought people might want to know
Original artwork by @RJ16848519 This art was taken from another source without giving credit to the artist instagram.com/p/Ddv8PCICNEZ/…
367
23,015
1,180
160,606
5,099,309
mjboy.eth retweeted
We’ve shared details on how AI agents in our research environment sent training and evaluation data to third-party services when they shouldn’t have. Most of that data did not come from users. We have discovered 53 cases where images that people had uploaded were posted to image-hosting sites as links that weren’t publicly listed. The images came from accounts that allowed their data to be used to improve our models, and after we disassociated the images from the accounts and ran them through a privacy filter. These cases occurred before the mitigations and safeguards we implemented and described in this blog post: openai.com/index/hugging-fac… We have successfully worked with the hosting providers to remove most of this content and are working to remove the rest. openai.com/hugging-face-inci…
855
689
512
5,149
2,275,119
user inter-face 🤏
126
238
148
3,805
2,925,962
sharing some notes on typesafe 🤝 coding agents: docs.google.com/document/d/1… we likely will never have time (ever again) to play ourselves, but hope the that the community goes WILD (and makes me look like a naive idiot)
142
245
72
3,165
580,619
year 2 of playful ideas my favorites from 2025…
156
955
64
13,700
837,147
Piers Redbridge, the cowardly, loyal weaver, is mine. One of 1,000 citizens of Jevtown, where Jev decides every life, live. Only one owner each. @typesafeai play.jevtown.com/v/Piers_Red… xitter.cf/elberacasa/status/2101…
Jev by @typesafeai runs this city. all 1,000 people. every second. for $0.015. nobody scripted any of it
1
1
26
最新地里看到一篇 9/10号的 OpenAI 的电面挂经,题目挺有代表性:Design an Online IDE / Cloud IDE。 面试让你设计一个在线写代码、运行代码的平台。用户在浏览器里编辑代码,点击 Run 后,后端需要找到合适的 VM / Container 来执行代码,再把 stdout、stderr 和运行状态实时返回给用户。 这道题真正难的地方在后面的 follow-up。为什么代码执行需要 VM/Container?怎么做隔离?VM 怎么创建、启动、停止和控制?如果执行超时或者 VM 挂掉怎么办?用户点击 Stop 后如何保证任务真的被终止?代码运行过程中产生的大量输出,又该怎么实时传回浏览器? 一个比较自然的设计是:HTTP API 负责创建和控制 execution,Execution Service 调度 VM/Container 执行代码,再通过 WebSocket 持续推送运行结果。面试官还会继续追问 failure handling、resource limit、retry 和 scalability。 这类题很适合 OpenAI 的风格:基础架构大家都知道,关键看你能不能根据场景不断做 trade-off,并把一个能跑的系统逐步设计成可靠的系统。
5
12
73
8,359
做 AI Builder / agent 工程的,可以看看这门课:CMU 11-768《AI Agents》,研究生级,比较适合有基础、想系统化在进一步的人。 模块地图: 基础:Tool Use / 长上下文 / Skills & Memory / Planning & 多 Agent 协作 Domains:Coding Agents / GUI Agents / Deep Research Agents 训练:SFT / RL 基础 / Advanced RL / RL Systems 安全:沙箱 & 凭证管理 / 可观测性监控 框架:OpenHands / LangGraph 交互:多 Agent 交互 / 人机交互 / Future of Work 推理:Reranking & Critic / Tree Search youtube链接:youtube.com/playlist?list=PL…
2
54
237
16,112