Sutskever to @NewYorker: "If you train a system which predicts the next word well enough, then it ought to understand"
ML fans when I show GPT-2 has no idea what happens as events unfold over time: GPT-2 is just system for predicting next words; not fair to ask it to understand.
-
-
Indeed. I think it's not widely understood that it's simply not *possible* for semantics to emerge purely from looking at syntax. Trying to do that will always be a losing strategy. Always. Is it part of a solution? Perhaps. Perhaps not. No one knows yet.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.