Not sure I said that, but when I worry about machine learning being fuzzy, unpredictable, and undebuggable, I'm told not to worry because we already don't understand how our systems work. It's already hope and crossed fingers. I sulk.https://twitter.com/rhatr/status/1013812206612475906 …
-
-
Replying to @rob_pike
Sadly even math is like that. Beyond a certain point, it's really hard to get to where you feel like you actually understand every theorem you're relying on, including indirectly. Fortunately their average quality is better than that of library code, tho...
1 reply 0 retweets 4 likes -
Replying to @RichFelker @rob_pike
I guess the difference is, in academic math you might not be able to prove every theorem you use, but someone else has. Whereas the neural nets that come out of training processes might as well be black magic
2 replies 0 retweets 3 likes
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.