{"id":48981329,"user":"uncivilized","time":1784565873,"time_ago":"a day ago","type":"comment","content":"

Regarding the latter, are you able to obtain good results from an LLM? With graphics programming work, I find LLMs only help in cases where the task would take me 5 mins or so.","comments":[{"id":48981411,"user":"cmrdporcupine","time":1784566297,"time_ago":"a day ago","type":"comment","content":"

Absolutely.","comments":[{"id":48981809,"user":"uncivilized","time":1784567992,"time_ago":"a day ago","type":"comment","content":"

Interesting. Do you happen to have a public example?","comments":[{"id":48982443,"user":"cmrdporcupine","time":1784570482,"time_ago":"a day ago","type":"comment","content":"

Of what specifically?","comments":[{"id":48985421,"user":"uncivilized","time":1784584853,"time_ago":"a day ago","type":"comment","content":"

A CUDA kernel, compiler optimization, or anything of similar complexity","comments":[{"id":48985633,"user":"cmrdporcupine","time":1784586209,"time_ago":"21 hours ago","type":"comment","content":"

Feel free to look, but don't bother replying if you have something negative to say, don't feel like having my day ruined by negativity.

https://github.com/rdaum/mica/tree/main/crates/relation-kern...

https://github.com/rdaum/mica/tree/main/crates/vm

https://github.com/rdaum/pagebox/blob/main/crates/wal/src/wa...","comments":[{"id":48986118,"user":"uncivilized","time":1784589484,"time_ago":"20 hours ago","type":"comment","content":"

Thanks for sharing. Looks interesting. Was all the code AI generated, or is it a mix?","comments":[{"id":48987064,"user":"cmrdporcupine","time":1784597593,"time_ago":"18 hours ago","type":"comment","content":"

that stuff in particular -- AI generated with heavy heavy prompting and up front design work and post-implementation testing

I have CUDA work here somewhere too but I have the repository private right now","comments":[{"id":48991969,"user":"uncivilized","time":1784640061,"time_ago":"6 hours ago","type":"comment","content":"

Very cool. Thanks for sharing.","comments":[{"id":48993036,"user":"cmrdporcupine","time":1784645225,"time_ago":"5 hours ago","type":"comment","content":"

Here's what is key: never rely fully on the "intelligence" of the model. Build a workflow and/or tooling that allows for empirical improvement. e.g. for performance tuning I have a benchmarking framework and a container/server that contains the differential results available via MCP for the agent to observe as it works. Using /goal and a clear destination you want to get to, it will literally grind for hours and use that help. Doesn't help with architectural stuff, but it does help with producing efficient code.","comments":[],"comments_count":0,"level":8,"url":"item?id=48993036"}],"comments_count":1,"level":7,"url":"item?id=48991969"}],"comments_count":2,"level":6,"url":"item?id=48987064"}],"comments_count":3,"level":5,"url":"item?id=48986118"}],"comments_count":4,"level":4,"url":"item?id=48985633"}],"comments_count":5,"level":3,"url":"item?id=48985421"}],"comments_count":6,"level":2,"url":"item?id=48982443"}],"comments_count":7,"level":1,"url":"item?id=48981809"}],"comments_count":8,"level":0,"url":"item?id=48981411"}],"comments_count":9,"url":"item?id=48981329"}

Comment by uncivilized