Agreed @GaryMarcus , which is why there is so little attention on solving the hard problems that you and Ernest describe & are still outstanding in state-of-the-art. Building a taller ladder wont get us to the moon, ie. todays #AI can't give us #conversationalai incrementally.
This tweet is just so wrong. Economists GPT-2 answers aren't more coherent than humans, they are (a) effectively cut & paste from human answers & (b) filtered by human for coherence, each chosen from 5 options
@erikbryn corrects below, but sexy news travels 75x faster than truthhttps://twitter.com/erikbryn/status/1199040038476156928 …
-
-
- 1 more reply
New conversation -
-
-
@erikbryn you know this isn't revolutionary dont you? It's not leading us to AGI. This is training on data that leads to regurgitation of text.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I published a book written by GPT-2 earlier this year. You may find it amusing: http://www.shardcore.org/shardpress2019/2019/05/16/algohiggs/ …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
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.