From previous experience we know that LLMs are great at confidently producing plausible output with problems that are hard to detect and fix without expert level knowledge and experience.

What experiences exactly? Whether the output is correct or just plausible and how many and how severe the problems are depends on the actual problem, and this statement is way too broad. There are also enough experiences where it works. It depends on how you use it.

I don’t see too much of a difference between a LLM producing plausible output as the result of reverse engineering and a human engineer producing plausible output as the result of reverse engineering. Again, we’re not talking about absolute experts doing this for the past 20+ years spending months to do this. For a “regular” engineer doing this, it’s the same best guess.

While I personally think you are right, I don’t think that any “serious” FOSS project will use sources from a fully LLM generated project.

While it’s pretty funny that one of Microsoft’s crown jewels gets shit on with AI, I don’t want to start the pro- or contra AI discussion, but what now? They reverse engineered or rebuild, or however you want to call it, all of Office, Autodesk, hell, for the fun of it do SAP next. Even assuming a perfectly working copy, not getting sued into oblivion, and not the dude going into bankruptcy due to Claude token costs - but what now? There is no way they can manage several repositories of this size alone properly (no, claude doesn’t count). There is no way this gets the community and traction to self-manage.

Apart from a ridiculous invoice from Claude and proof that frontier LLMs can reverse engineer / clone well (…duh), this is a big pile of nothing.

Removed by mod

Removed by mod

Removed by mod

Removed by mod

Removed by mod