Get the latest tech news
In defense of shallow technical knowledge
Whenever a new piece of technology comes out (these days, mostly AI) I go to some effort to understand it. Usually I end up writing a post about it, so I can be…
It has been very useful to understand what language models actually are, and I’ve spent a lot of time doing that (coding LLM inference from scratch in Javascript and Ruby, for instance). That way we don’t confuse people who aren’t familiar with the points above and just want to try the LLMs out, while still letting developers have control over what they’re building with the models. The whole point of this process is to be able to have useful intuitions about these technologies (e.g. “oh, text diffusion would be a good fit for the AI box in Google search results because it has to be very fast and has a mostly fixed size).
Or read this on Hacker News