>>12601034>anyone else get the impression that the modern style of machine learning is just a massive meme that will end up having a few niche industrial uses and nothing else?Only "lol Idk computer I just use it for facebook whats a browser" tier thinking don't recognize how pervasive they are in literally everyday use
From facial recognition to spam filtering to porn suggestions to amazon "buy this too" shit, pretty much every single task I can think of on the internet has a NN implementation
Further, I'd say >70% of all business, economic, military decisions are guided by ML implementation. It doesn't feel disruptive because it replaced pretty much every single decision-algorithm in almost every field with a better version.
We're brushing the surface, of course; transfer learning is just a barely-scratched surface.
I get the "ML" is a meme, but it seems to be based solely on how many pop-sci articles have "ML" in the title and less on actual data of implementation and usefullness
tl;dr: ML already runs the world behind-the-scene in most decision making processes, and even if we never find any further application, it already controls decisionmaking in real life
