|
|
|
|
|
|
by thot_experiment
711 days ago
|
|
|
most of the significant generative AI runs on a potato, and runs well on any modern GPU, obviously SotA models have worth but I have not touched a paid service since Llama was leaked and I use AI a ton in my day to day, for the rote replacing stack overflow type shit I used ChatGPT for a local copy of llama, gemma, mixtral or whatever is more than enough |
|
|