Maybe I'm wrong, but I'm really skeptical about projects like this. 200000 lines of code addition in 3 months, basically 2000 lines per day. I don't think any human brain is capable to handle such cognitive complexity, unless the code is all plain data get/setters, which any framework is certainly not. For me a good way to use LLM is to learn how to build something big by writing a tiny one with it. But this only gives a shallow understanding of a big system. For any critical complex system you want to maintain for a long time, LLM is in no way a good choice.
> Managing a codebase written by an LLM is difficult because you have not cognitively loaded the entire thing into your head as you do with code written yourself.
Wow you really nail the point, that's what I felt but I did not understand. Thanks for the comment.
I'm wondering if this scheduler is for something like user-space threads. And What is the relationship between such scheduler and go runtime for goroutine and JVM for Java virtual thread?