Comment by nirui
4 days ago
One question here: Why you don't see a lot of good American open weight models on the market? I'm guess it's about money? Could there be ways to lower that cost? Sharing idle GPUs through network (like Folding@home, BOINC), open and fair market for training data, open source training framework, and such?
This whole Chinese AI thing makes me giggle at these American AI companies, but the ultimate solution/long-term goal is still to make the technology universal. I don't think China alone can make it happen.
Also, if Americans can too offer open weight models, that could change the logic of their laws, making it more open rather than protective.
Money is a big part of it. The notable thing about Chinese open models is their size. Most open source models out of US labs have not tried to compete with the enormous flagship models you see from frontier labs now. Cohere, Arcee, Poolside, etc all put out open source models, and of course, Google has their Gemma line, Meta has their open models, OpenAI released an open weight 120b model, etc. The incentives are just tricky for running a massive training run for a fully open model. Monetizing can be hard, as if you want to monetize inference, you'll be competing against platforms like Fireworks etc that will also be serving your model.
Thinking Machines just released an enormous open model though, and I'm hopeful to see more of this globally. I believe open models are emphatically a good thing. But I do sympathize with smaller labs trying to figure out the economics.