Now of course, it's not that hard to understand: the ability to generate hype and the ability to generate meaningful progress / shape the future are entirely orthogonal
Show this thread
It's funny how at any given time what's hyped in AI is largely meaningless while what's actually momentous is largely under the radar (AlphaFold was a rare exception to this rule)
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What are some examples?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What are the things that are momentous but under the radar right now?
-
I’d argue for Quantization and model shrinking through bayesian analysis. Take giant models, shrink them down with very marginal loss in accuracy, and then be able to scale them very practically.
- Show replies
New conversation -
-
-
This calls for a list of what these are perhaps...
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Subtweet about Perceiver?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
So if I have to understand between the lines of this tweet it's the Codex demo is just Hype? There is already github Copilot which is being used by thousands of developers and which clearly increases their productivity. For you this is not significant progress?
-
Try asking for the Technical Preview to judge for yourself.
End of conversation
New conversation -
-
-
Non-AI/media people can't really tell when progress has been made. They only have outcomes to work with and it's understandable. But the effect is pronounced in AI due to the alarming rate innovation is happening.
-
AI people cannot usually tell either.
End of conversation
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.