Why DeepSeek, Kimi, GLM, and s1 are shaking up a hundred-billion-dollar industry — and how anyone can do it
A fantastic write up, with evidence, stunning visuals and a roadmap to do it yourself. Good job!
Personally I’d rather tweak the system prompt of the existing well prepared and tested LLM than train or distill a new one
Thank you Fabio. So kind. Here here! I whole heartedly agree. Keep it simple whenever possible. Models have become extremely capable that good (system) prompts (Inc of memory) take you far enough depending on scope of projects/work etc…
A fantastic write up, with evidence, stunning visuals and a roadmap to do it yourself. Good job!
Personally I’d rather tweak the system prompt of the existing well prepared and tested LLM than train or distill a new one
Thank you Fabio. So kind. Here here! I whole heartedly agree. Keep it simple whenever possible. Models have become extremely capable that good (system) prompts (Inc of memory) take you far enough depending on scope of projects/work etc…