There are cases where Codex is infinitely slower than Claude for a task with a /goal command, because Codex simply refuses to work.
See https://t.co/GRDHkqucMS
If you're building your own agent project with the need to debug agent resoning content, avoid Codex.
You might not have wondered if today's coding agent refuses to execute a simple legit task, but here it is:
https://t.co/GRDHkqucMS
Codex refuses to extract the "reasoning" field from OpenCode database of an open source model trace. With /goal command it even runs indefinitely.
I think this matters because software engineering is starting to look more like research:
in a world of abundant code generation, bottlenecks shift toward validation, synthesis, and autonomous iteration.
Akari is my attempt to build toward that future.
https://t.co/3vQiYoeE2T
I spent last week on vacation in Tahiti.
Meanwhile, my autonomous research agent produced 5,000+ commits.
My productivity didn’t drop. It kept going while I was away.
The bottleneck of my output is no longer my working hours.
Open source repo below.
https://t.co/4Oo6xxBVDF
The most surprising part was not that it could generate code.
It was that it completed experiments I wanted to do myself without explicitly assigning the exact tasks.
That is the real promise of autonomous agents:
not just faster, but making output less bounded by human hours.
🎉 From 0 to 20,000,000 3D creations—how did we get here?
❤️ With YOU: Millions of Meshy 3D Creators!
💡 2024 was incredible: 3 version updates, 100+ new features, endless love (and spicy roasts), and we became the ONLY 3D tool on A16Z’s Popular AI Tools Survey!
👉 Watch our full 2024 Wrapped and unlock your Personalized 2024 Summary now on https://t.co/mpI4Bej05k.
🎉 Oh, and one more thing— We’re celebrating with a 20% holiday discount on your first month or year for all plans!
🚀 Use the code 'Meshy2024' at checkout, and grab this deal before it’s gone!
#3dmodeling #My3DWithMeshy2024 #AI
Meshy generates quad meshes with edge flow by default. To obtain a triangle mesh without edge flow, one needs to intentionally set generation config away from default.
Congratulations to the Edify team
@NVIDIAAI on the release of their text/image-to-3D model
Besides the enhanced visual fidelity, I think the most impressive thing is how clean the quad mesh is compared to other state-of-the-art text/image to 3D models! This has significant implications for downstream editability, particularly in reducing the need for mesh retopology for 3D artists and animators.