Fable is the ceiling.

After testing out Fable I had a stark realization about words I heard months ago. The team at anthropic said it had been using a model such as this one to code their models. If this is true, or to be trusted I believe this model is very close to the ceiling of what AI from anthropic will produce. I think we are going to hit what I like to call the "China factor" in AI very soon, where China will most certainly outpace American frontier models in the next years. Chinese AI developers do not use AI to improve their models aside from the distillation of american models. models like GLM or Kimi could easily outpace American frontier models given the pace we've already seen. I think its more than likely. Chinese devs are also forced to work with worse conditions for creating AI due to their lack of top tier compute compared to America, and theyre still mostly keeping pace which in itself is already unbelivable. I think we're at a point where American AI will stagnate out, minor improvements every release not small or large jumps.