If Opus 5.5 was a woman i will never cheat on her
i made this FPP open world game in 7hrs just with Blender and ThreeJS
i think the only limitations there are now is just your imagination
Opus has to be the best model for its weight class and the fact that output stayed consistent since release thank you anthropic
J A Z I I 分享用 Opus 5.5 制作的 Lantern Hallow,并公开了工具组合和在线试玩入口。
1️⃣ Tripo 生成 3D 对象,Blender 制作动画,Godot 作为引擎
2️⃣ 据作者自述,制作约 8 小时,成本为 108 美元;这属于个人案例,不能视为通用报价
3️⃣ 作者提示试玩需要较好的硬件
#AI游戏#Godot#Blender
lantern hallow made by opus 5.5
asked it to make a small cozy game and after about 8 hours of cooking it came back with this
tools it used:
> tripo for 3d objects
> blender for animations
> godot as game engine
overall cost was $108, which is about 3 session limits on claude pro plan
you can play it here (gonna need good hardware): https://t.co/pEjl4djb1v
Sac 用 Opus 5.5、Tripo 和 Blender 制作了《The Little Prince: Across the Stars》,把《小王子》的星球做成可探索的 3D 场景。
1️⃣ 作者介绍开发耗时 16 天
2️⃣ 原帖提供在线游玩入口和演示视频
3️⃣ 游戏与详细制作流程计划后续开源,目前原帖尚未给出源码
#AI游戏#Claude#Blender
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✨:https://t.co/qr4wR0jLu3
RareUI is now one prompt away.
Introducing RareUI MCP - browse, discover, and add beautifully crafted UI components directly using your AI agents.
https://t.co/tbgxFSTRCk
You already gave the LLM the answer. So why is it writing it out from scratch?
Introducing a new AI model architecture: Constrained Decoding Models (CDMs)
LLMs were built to write. Most production workloads just need an answer: fast, consistent, and affordable.
CDMs are purpose-built for 1 task each. They read your input once and skip the token-by-token generative loop that makes LLMs slow and expensive.
Not just a decision model, but a family of 3 CDMs live today:
• Classify: define your labels and get back a calibrated probability + reason for each. Used for routing, tagging, and intent detection.
• Extract: describe the entities in natural language and get back well-formed JSON with confidence scores + citations. Used for web scraping, document processing, and PII redaction.
• Summarize: pass in long text, images, or PDFs and get back a consistent summary under 1,000 tokens. Used for context compaction, metadata generation, and transcript digests.
Built for production:
• 1M tokens of multimodal (text + image) input, 31x larger than Jev
• 77.2% on DecisionBench vs Jev's 72%
• 200x faster, 467x cheaper than frontier LLM APIs
• $0.05/M input tokens. Output is free (yes, even for summarize)
Live on your favorite platforms today:
@OpenRouter@vercel_dev@LiteLLM@LangChain@TrustedRouter@itsArthurAI@useRouterPlus
Get 50M input tokens for free (for the first 50 hours) at link below 👀
微软开源 Code Video Model 模型:用代码生成视频
微软研究院推出了一套新的视频生成方案:Code Video Model(代码视频模型)。
它让 AI 生成视频少一些“开盲盒”,先用写代码的方式,明确画面布局、动作轨迹和物理运动
到了后面的细节打磨阶段,就撤掉代码参考,根据参考图片去还原真实质感和细节
学电影拍片,“先画火柴人骨架,再让画师上色”
这套技术借鉴了传统电影工业中的前期可视化预览(Previs)机制:先用简易三维场景预演机位、演员走位和镜头节奏,确认安排后,再投入正式拍摄与后期制作。
你给它一句提示词,它先写一段代码(用类似网页 3D 引擎),在虚拟空间里把舞台、桌椅、角色和摄像机机位精确摆好,连运动轨迹和物理碰撞公式都算好。
这时会先跑出一个类似“低配游戏画面”的动态小样。如果你觉得椅子放歪了,改一行代码立刻就能纠正。
有了低精度的代码视频后,怎样把它变成逼真的最终成片?
这里存在一个关键挑战:如果直接把代码视频作为参考条件全程输入给视频扩散模型,AI 容易连同它的简陋几何和平坦材质一起模仿,导致成片仍有“塑料感”。
为了解决这个问题,研究团队提出了一种无需重新训练的调度机制:早期步数条件注入(Early-step Conditioning)。
只在扩散采样的最前几步让模型参考代码视频来“定构图和动作骨架”;到了后面的细节打磨阶段,就撤掉代码参考,全力根据参考图片去还原真实质感和细节。
Introducing Selfdrive.
The AI growth team for your iOS app that gets you users while you sleep.
It handles Apple Ads, ASO, SEO and UGC creators for you. All of it.
https://t.co/2UCiidgwrq
Introducing Apple HomeView!
It recognizes up to six people, connects seamlessly across your Apple devices, and makes managing your home easy with Siri AI. All in a stunning, elegant design with privacy built right in.