...until you get rug-pulled (like with Fable recently) because the model you're depending on is proprietary, then you have fun with no model at all.
The sooner Chinese labs catch up the better it is for all of us, even if you don't use their models, as they're the ones who define the baseline capability that cannot be taken away from you (no one is going to limit/remove access to an LLM if you can get equivalent/better unrestricted access from Chinese open-weight models).
alt Hacker News
...until you get rug-pulled (like with Fable recently) because the model you're depending on is proprietary, then you have fun with no model at all.
The sooner Chinese labs catch up the better it is for all of us, even if you don't use their models, as they're the ones who define the baseline capability that cannot be taken away from you (no one is going to limit/remove access to an LLM if you can get equivalent/better unrestricted access from Chinese open-weight models).