emmettbt

this seems to be a bit contrarian?

Cool... but this does seem undermined by the fact that Ollama can add support for decision models at any time.

This is so interesting! On a side note it is concerning that a blog post this benign triggered a “safety” check.

I very much agree. There is need for more innovative llm training methods than just give it more data.

proton is nice but their "open source" claim is deceptive as it is only their client that is open source.

Unified memory is also a kicker for crashing a Mac.