Post #2200251
2026-03-03 17:16 UTC
@bobmagicii@phpc.social That's how deep learning works, but the inference just puts tokens in front of other tokens, right? I guess it can interrupt its own chain-of-thought, but that's still mostly linear.
Replies (1)
-
@bobmagicii@phpc.social 2026-03-03 18:26
@t_var_s@phpc.social depends how low level and or pedantic we want to be i guess since at the end of the day every cycle of the cpu is still a linear process.