4 ms·
Unfortunately I tend to agree with Atwood on this one. Machine learning systems often demonstrate amazing results initially, but then seem to lag in that “last
by freetime2 5y ago
Unfortunately I tend to agree with Atwood on this one. Machine learning systems often demonstrate amazing results initially, but then seem to lag in that “last mile” compared to their human counterparts. The most irritating example that comes to mind for me is my HomePod - more than a decade after Siri was commercially released, it still regularly wakes up and interrupts my meetings with “I’m sorry I didn’t get that” when I say “I see”.
I think self-driving technology is probably in the same boat. It’s amazing what it can do, and it is definitely useful today in many contexts. But it will be a long time before it is able to achieve parity with human drivers in the enormous number of different contexts that humans are able to handle with relative ease.
For any engineers working on self-driving cars, though, I hope you prove me wrong! I would love it if my next car came with a level 5 self-driving option.
- tim333 5y agoIf you look at AI type computer technology in general it tends to get steadily better while human performance remains fairly constant. So in chess or go you can see the computer players steadily get better in the rankings until one year they get better than human. A few years before some people say these things are rubbish they'll never get there but the rate of progress is the thing to look at. With self driving, things like Teslas can drive from A to B but are a bit rubbish and need interventions to avoid crashing but I think you'll find the level of that keeps increasing year on year till one day they get safer than human.