The idea we may have fast continuous learning soon really just worries the shit out of me. I was watching some other clips about models on chips producing 12k tokens/s, it wasn't very good yet, but man once you start getting kind of good at those speeds you can start to do some crazy things.
Cerebras can do like 4000 tokens per second on gpt oss 120b with it's new chip and around 1000 tokens per second on frontier models using its new chip.
Were you talking about Talaas 8b model? They will probably do something like Bonsai or Qwen 27 B next I hope. But it makes might be an old model.
