[pvo50555]

There was a post a few weeks back (or a reply to a post) showing an app entirely made using an LLM. It was like a 3D globe made with 3js, and I believe the poster had created it locally on his M4 MacBook with 96 GB RAM? I can't recall which model it was or what else the app did, but maybe someone knows what I'm talking about?

[jckahn]

That was with Qwen 32B, which is still the best coder model for the size. You could run that just fine with a 36 GB RAM Mac.

[pvo50555]

Is it actually able to generate all the files and directory structure, etc? Or did the author just take all the responses he got from multi-shot prompting and eventually assemble it into the final product? I believe I used LlamaCoder some time ago to build the scaffolding of an app, but I'm not sure about the state of that project now.

[jckahn]

I imagine _a lot_ of iteration was involved regardless of the model.