Viewing profile — gsjbjt

That's awesome - can you say anything about what datasets this was trained on? I assume something specifically conversational?

Site: https://dynalang.github.io/ Tweet: https://twitter.com/realJessyLin/status/1687515845281624064

The point is that LLMs can’t backtrack after deciding on a token. So the probability at least one token along a long generation will lead you down the wrong path does indeed increa…

What are the usual ways to discover discord servers (also through search, or some other means)?

Nice post! I work on NLP and I think a lot of ideas in this post resonate with what I find exciting about working on the intersection of language + the real world: large text datas…

> My theory is that taste is one quality that separates the academics from the business people. Academia doesn't necessitate a lot of taste. If you have it, great. If you don't hav…

This (made by a friend) is exactly that idea! https://www.anishathalye.com/2019/12/12/constraint-based-gra...

I’m working at Lilt (lilt.com) right now, and this is exactly why I’m excited about the company ;)

Yeah, it's an excellent course. That is typically the unfortunate downside of self-studying courses, but AFAIK the tests aren't that sophisticated - you can mostly test it yourself…

David Spivak and Brendan Fong have been doing lots of work at MIT to evangelize applied category theory. See their recent books/classes: Seven Sketches in Compositionality: An Invi…

Implementing it is the fun part! You might like going through the Raft lab in MIT's distributed systems course: https://pdos.csail.mit.edu/6.824/labs/lab-raft.html + Jon Gjengset's…