4 ms·
I think there is a much simpler explanation behind this: training data. Llms have had truck load of training on normal code than a super optimized ones. That'
by mkagenius 2y ago
I think there is a much simpler explanation behind this: training data.
Llms have had truck load of training on normal code than a super optimized ones.
That's why you will see it struggle with nginx configs. And on newer technology where there is less amount of data available.
- userbinator 2y agoIf they trained an LLM on demoscene entries, maybe the results could be better, but I doubt by much.
- simonh 2y agoI just upgraded by AI optimised graphics driver and now every time I launch COD every map is a psychedelic rave trance disco. Help!
- lee_ars 2y agoI can help, but we gotta get through the 20 minutes of scrolling greets first.
- saati 2y agoA lot of demoscene optimizations are sacrificing correctness for speed, and thus not really applicable in most domains.
- jasfi 2y agoRefining and adding to the text and code they train on is a major area of focus at the moment. Things will definitely get better, but it will take time.
- th0ma5 2y agoThis is the key question... "What is the total information in this model and interaction?"