3 ms·
> We're trying to pretend AI/ML systems will make better decisions, and faster, but they always fail catastrophically at the edge cases in ways humans never do.
by ASalazarMX 5y ago
> We're trying to pretend AI/ML systems will make better decisions, and faster, but they always fail catastrophically at the edge cases in ways humans never do.
The pretense is that AI can handle the bulk, routine cases. Not keeping enough people to handle the edge cases is the real mistake.
- Arrath 5y agoSomething seems to get lost in translation from the premise of "increase efficiency and decrease your team size by letting the AI have a first pass at the data" and the execution of "well we have this fancy AI we can just fire everyone who used to touch this data"
- mcphage 5y ago> The pretense is that AI can handle the bulk, routine cases. Unfortunately, determining whether something is a routine case, is often not a routine case.