gold retweeted
crazy...that i made this with GPT Image 2.5 and Seedance 2.5
nobody is teaching you the actual workflow behind these AI realistic videos.. so here it is,
full breakdown, every prompt included👇
gold retweeted
I optimized my Dream Loop skill to give you Astra's 3D magic for 80% cheaper!
This demo was one-shot, in 25 minutes, for 7% quota on ChatGPT Plus. Yes, the $20/mo one.
3 days of banging my head on the wall and burning Astra tokens. Sharing it for free. Y'all better build some cool stuff with this
Link: github.com/achimala/dream-lo…
Details, demo prompt, and cost breakdown are in replies.
To be clear: It's not perfect. The demo is a small interactive scene, not a full game. For full games, your best bet is to build it with smaller models, then use this skill to improve the graphics. This is the best I could do, but it's not going to fix the fact that Astra is expensive and Plus quotas are tiny
gold retweeted
🚨 Architects aren't going to like this.
Someone just open sourced a full 3D building editor that runs entirely in your browser.
No AutoCAD. No Revit. No license key.
It's called Pascal Editor. Built on React Three Fiber and WebGPU, so it renders straight on your GPU at near native speed.
What's actually inside:
A full building hierarchy. Levels, walls, zones, slabs, all editable in real time.
An ECS style core where every object updates through GPU accelerated systems.
Zustand state with real undo and redo.
Next.js frontend, so it ships as a web app and nothing gets installed.
Change tracking that re-renders only what moved, not the whole scene.
And the part that got me:
You can stack floors, isolate one level, or pull a whole section out of the building. Select a zone, drag a wall, reshape a slab. In 3D. In a browser tab.
17.5k stars. MIT license. Zero dollars.
gold retweeted
Less than a minute to generate this shot with MiniMax H3 Max
via Partner Nodes.
It works quite well for fast video-to-video experiments, especially when starting with a basic blockout + reference images.
The workflow also includes an option to upscale to 2K.
To try this workflow, link below 👇
gold retweeted
A single image. An entire 3D world. And every important object is editable.
Hyper3D WorldGen turns a single scene image into a structured 3D environment with:
- Independent assets
- Spatial relationships
- Physical properties
- The freedom to move, replace, refine, and interact with them
This is scene-level 3D generation.
🧵👇
gold retweeted
This 3D dragon was made in less than 30 minutes.
GPT-6 Astra built the production-ready creature and its environment entirely in Blender by using Higgsfield plugin.
gold retweeted
Astra was able to recreate my Studio in Blender from 5 photos + 3 panos,
I shared the Blender files 3 days ago, it took 11 minutes initially, folks said it was fake... well.
I had Astra make a web viewer for it, so you can check it out yourself, and download the model too.
gold retweeted
I fed GPT-6 Astra 9 crappy photos of my studio and it was able to understand the spatial arrangement, stitch them together and create a fully interactive 3D model of the space.
Absurd. I'll link it below so you can check it out for yourself:
gold retweeted
Unreal Engineで、数千人規模の群衆をリアルタイムに表示できる「GaussianCrowds」が開発中。
大量の人物を軽量に描画しながら、照明や影の変化にも対応。スタジアムやライブ、戦場など、大規模な群衆表現で活躍しそうです。
投稿者：technomancyai
gold retweeted
"Hey Astra, make me a Mac app. It should render a 3D iPod that you'll build in Blender, then using the original iPod UI and interactions I want to be able to visualize all my Codex threads on it."
15 minutes later: Boom.
What's even in this model? What time we're living in!??
gold retweeted
GPT-6 Astra is now available to all Pro, Enterprise and Business Premium users in ChatGPT Work and Codex. It’s also live in the API.
It's a phenomenal model and I love every second working with it!
Astra follows instructions more closely, so revisit your prompts and skills:
> Make “can you…” mean “do it”: finish authorised work and prepare a reviewable result before seeking approval.
> Audit AGENTS.md and skills. Ask it to identify the exact file and rule blocking progress. Prioritise your request over skill guidelines.
> Specify short paragraphs, plain language and phrases to avoid.
> Ask for subagents on independent work when it improves speed or quality.
> Run relevant checks. Repeat or expand only for changes, failures or unresolved concerns.
To push it to it's limits, I gave it a prompt to build a real-scale video fly through of nürburguring track in blender and it did not disappoint!
Enjoy! ✨
gold retweeted
If I understand correctly, the insight here is very clever and might even be a glimpse of the future where all the usual workflows we took for granted become unbundled.
Basically, the idea is to completely unbundle pose detection, segmentation, and all the usual heavy stuff, and outsource it to external image editors (ex: gpt-image-2) for just a single frame. That single frame is all you need.
By dropping the pose/mask/face pipeline entirely, the model itself can be fast. A workflow innovation, implemented as a finetune.
Introducing Viggle-Animate, our first open-weight model.
Built on MiniMax-H3, it replaces a character in a video from one repainted frame. No pose skeleton, mask or text prompt at render time.
3 forward passes. 5s of video in 26s.
Try now: huggingface.co/viggle/viggle…
gold retweeted
🔥最新：GPT-6 Astra拿下Terminal-Bench 4.0第一！GPT-6 Astra + Codex才是现在的王炸组合！
58.2% vs Fable 5.1 的 57.9%。
分数只高 0.3 个点，成本却只有第二名一半：$3.3k vs $6.2k。
再看第三名 Opus 5：51.8%、$6.0k、6.5B tokens。
Astra 用 1.5B tokens 就把分数拉开 6.4 个点。
Agent 时代比的不是谁更贵，是谁用更少的钱把终端任务做完。
Codex harness + Astra，效率赢了！
榜单：tbench.ai/
#GPT6Astra #TerminalBench #Codex #OpenAI #ClaudeCode #Tibo
🚨最新：Codex 教，Claude 学！Codex、Claude双双给用户重置了额度！
Tibo 已发完重置卡，Astra 提前一天上线，Plus/Pro/Business 用户全部银行假期重置，还补了一句“你们其实没那么有耐心，但没关系”。
同一时间，Claude Code的 Lydia 也给所有Max用户重置了周限额。
社区第一反应：终于有 Anthropic 版 Tibo 了。
这说明什么？Tibo 已经把「对用户好」做成了整个 AI 编程工具的新标准。
祝大家 Astra 日快乐！
#Astra #GPT6Astra #Codex #ClaudeCode #Tibo #OpenAI #Anthropic
gold retweeted
🔥绝了！GPT-6 Astra在24分钟内用Three.js把整座杭州搬进网页，能逛能飞能切换昼夜！
刚上线的GPT-6 Astra，直接用Three.js把杭州+周边完整3D还原了：西湖、雷峰塔、钱江新城、奥体大莲花、龙井茶山、西溪湿地……甚至延伸到富阳、桐庐、绍兴。
拖动旋转、滚轮缩放、点击地标自动飞过去，还能切换白天/日落/夜景，开自动漫游随便逛。建筑细节、茶田垄沟、水面小船、桥上车辆都有，不是静态图，是真正可交互的微缩杭州！
杭州的朋友赶紧认认看，你能认出几个地方？👀
Astra简直是隐藏的城市规划大师啊！
#GPT6 #Astra #ThreeJS #AI建模 #杭州 #3D场景 #AI编程 #把城市搬进网页
xitter.cf/nft_chen/status/209613…
🔥最新：GPT-6 Astra拿下Terminal-Bench 4.0第一！GPT-6 Astra + Codex才是现在的王炸组合！
58.2% vs Fable 5.1 的 57.9%。
分数只高 0.3 个点，成本却只有第二名一半：$3.3k vs $6.2k。
再看第三名 Opus 5：51.8%、$6.0k、6.5B tokens。
Astra 用 1.5B tokens 就把分数拉开 6.4 个点。
Agent 时代比的不是谁更贵，是谁用更少的钱把终端任务做完。
Codex harness + Astra，效率赢了！
榜单：tbench.ai/
#GPT6Astra #TerminalBench #Codex #OpenAI #ClaudeCode #Tibo
