Meta Muse Code Is HERE – Spark 1.2 & Meta’s NEW Coding Agent!
- Source
- youtube.com
- Author
- Bijan Bowen
- Date
Context
A hands-on review of Meta's Muse Code and Muse Spark 1.2 shows why a generated application's first successful build still needs a playthrough. In a skateboard-game test, the reviewer reported a camera getting stuck near buildings and a board that did not move independently of the rider.
After feedback, the reviewer reported improved rider movement and board behavior, while encountering another problem during a rail-grinding attempt. This supports a narrow finding: the system could revise several concrete defects after human inspection, yet the result still needed checking. Because the review tested the model inside its accompanying coding tool, it does not isolate how much of the outcome came from either component.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- multimodal — A model that works with more than text — reading images, audio, or video, and sometimes generating them too.
“These agents remain active throughout each session rather than being spawned for individual tasks, which is something that I have not seen as often.”
Bijan Bowen
“part of testing these AIs is basically just giving them imperfect real life things exactly like this”
Bijan Bowen
“I'm not huge on like testing things that are partially discounted cuz that's not the actual genuine price they're going to be.”
Bijan Bowen
“I think pricing-wise, this may be a little too weak to be competitive, at least in terms of the performance noticed in these specific tests.”
Bijan Bowen
“I kind of like seeing that because the models are independent and have their own styles, I guess could be said.”
Bijan Bowen
Checking sign-in…
Loading comments…