-
@ waxwing
2025-01-27 16:44:30Running deepseek-R1-Distill-Qwen-32B-GGUF using LMStudio on Windows with an nvidia 3090. It's slow, of course, but I'm asking it questions about mathematics and cryptography... the quality of reasoning is through the roof, *so* much better than what I'd seen before. And yes you can get rid of censorship running it locally but I'm less interested in that rn, more just don't want to feed data to their algos.