4 ms·
Don't count your chickens before they hatch. I believe that the odds of an architecture substantially better than autoregressive causal GPTs coming out of the w
by AS04 1y ago
Don't count your chickens before they hatch. I believe that the odds of an architecture substantially better than autoregressive causal GPTs coming out of the woodwork within the next year is quite high.
- suddenlybananas 1y agoWhy do you think that?
- 9rx 1y agoHow does that equate to "winner take all", though? It is quite apparent that as soon as one place figures out some kind of advantage, everyone else follows suit almost immediately. It's not the 1800s anymore. You cannot hide behind poor communication.
- VerminOctopus1 1y agoWhy do you believe this? Do you know researchers actively on the cusp or are you just going off vibes?