Someone just told me that machine learning has not produced true breakthroughs in the last 20 years. Dude, I just bought a little cylinder for 20 bucks on Amazon that argues with my children about the weather and sings them their favorite songs from Spotify
Self driving cars were around since the late 1980ies, this is really mostly incremental progress.
-
-
Backprop has been around since the 80s as well, thanks to researchers like Hinton. That way I can call deep learning and all of its results incremental. Which isn't necessarily incorrect, but for your point I think SIRI/ALEXA are pretty boring and meh.
-
1960ies. But yes, you are of course correct. I was still impressed that $20 buys you a futuristic bluetooth speaker that can give an intro to quantum mechanics and offer a reasonable response to a certain family member's request to poop its pants.
End of conversation
New conversation -
-
-
So your FB friend might be lamenting that part. If this person is waiting for secret AGI sauce rather than zettaflopping our way into true AI,.. I get what they mean.
-
Being able to obliterate chess and go in less than a day looks quite impressive to me. I also did not expect translation and vision to become so good so fast. End to end learned image annotations and GANs are early days but a big deal.
- 1 more reply
New conversation -
-
-
Anyhow, I think the distinction here is that we've gained computational power, access to shitloads of data and learned to apply and play with lessons mostly learned in the 80s and 90s... that much is true.
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.