Machine learning, AI, philosophy. Backpropagandist. Father. Contrarian by instinct, academic by training. More interested if legal needs to sign off on it.
Michael Jordan is obviously an all-timer, but Hinton, LeCun, Bengio, Pearl, Vapnik, and Breiman aren’t exactly slouches. To say nothing of Schmidhuber, naturally.
@ArielKwiat I'm sure it's possible. But it seems unlikely it will be the practical solution we eventually converge to (i.e., modern systems will not merely scale, but new components will be added or parts reinvented to solve the current challenges).
@ChrSzegedy@andrewgwils I would not have guessed Millennium Prize problem solving would outpace the ability to make a decent PowerPoint about those solutions, but here we are.
@Austen greatness is locally conserved: positive deviations require equal and opposite corrections (i.e., the conservation of greatness, as established in the Mulligan experiment)
The normal person calls the monster Frankenstein. The half-educated insist, smugly, that Frankenstein is the name of the doctor, not the monster. The well-educated see that Dr. Frankenstein is himself the monster, in a spiritual and literary sense. (And the VP, communicating publicly, is quite right to use the first/third sense).
@pegobry_en "Mad scientist creates horrifying monster" is the far better summary.
It's also better than the "he just needed to be loved" message some people took away from the book.
I'm already nervous about having to file taxes (which I can't do until April 2027 at the earliest and don't have to do until Jan 2028). I just hope that by that point AI will have advanced to the point at which it does our taxes for us.