Tfw I take AI too seriously and explain it too clearly so nobody even gets madhttps://twitter.com/vgr/status/1014307742624780288 …
-
-
Replying to @Meaningness
Only a sensationalist critique can penetrate a sensationalist discourse. Friend of mine has a similar problem: he writes coherently, non-sensationally and with carefully evidence about animal rights and veganism, and is surprised when he gets less traction than he expects
2 replies 0 retweets 4 likes -
Replying to @vgr
I wrote it for actual AI people, rather than singularitarians. Maybe it’s obvious to them? OTOH you’d think they’d do more science if so.
5 replies 0 retweets 4 likes -
Replying to @Meaningness @vgr
I was at a conference a couple weeks ago and many hallway conversation ended with “but of course no one really knows why these these systems work, or we’d also be able to explain <fill in the blank>” My impression is lots of people are tinkering while tinkering is productive.
1 reply 0 retweets 1 like -
Replying to @alexisgallagher @vgr
Yes… I get irritated when I’m pretty sure I can figure out why they work, and that I could teach other people to do this too. It’s dumb to be irritated about this, rather than offering to help, however!
1 reply 0 retweets 0 likes -
Replying to @Meaningness
For example: we might think we know how image classifying CNNs work, because their layers seem to recapitulate successively more abstract feature detectors like from classical computer vision. But if that fully explained it, they wouldn’t be so vulnerable to adversarial attack.
1 reply 0 retweets 0 likes
Yes; that’s probably the most interesting case!
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.