This seems nice but I’ll need to read the syntax more carefully to see how capable and convenient it is.

Definitely! Capability vs convenience is tough! Hope you find it useful.

The point about semantic diff viewer sounds very interesting. I think many coding harnesses are not doing this well enough.

I have not studied it properly too. Good that it has both code and note though.

This is sort of clever! It seems to me though you can almost design a small dsl and you don’t need ai at all.

I read your github page and there's mention of continual learning but no argument how catastrophic forgetting is avoided. I think you should explain that better.

"How continual learning works" is the section you are looking for. The main experiment is described there.

Yes I read that one, and that’s why I said the explanation is not clear. Anyway, catastrophic forgetting is theoretically not possible. “ Existing capability must substantially learn new information, so its representation must change, while retaining its still-valid old knowledge”. I think your setup is demonstrating something weaker.

You might be right. I have not measured it on typical CL benchmarks yet. I will try to do that after the reading of the whole corpus. And don't forget there still is a normal slow forgetting happening, ofc. But my main evidence of the network doing something right is an ability to learn from a continuous stream of data. There is no other architecture I know of that can do it on a similar scale.

I think as an experiment your setup is interesting. But in realistic systems even standard evals, i.e. making sure things work well for normal use cases, are already enormous, let alone thinking about catastrophic forgetting.