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.
-
-
I stand corrected. Thanks
@GaryMarcusThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This is a great juxtapose, very nicely done. To me this contradiction is less a failing of AI or research. And more about the imprecision of the English language.
-
Ie. in a language with 30 shades of "understand", it is far less likely Dr. Sutskever and Dr. Dietterich would have chosen variants that contradict, because I suspect they agree on the meaning if not the word choice.
- 1 more reply
New conversation -
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.