Jeez guys. Fix this problem once in self-driving cars and it's fixed forever for all such cars. Fix 1000 such "bugs" and those cars become *immensely* better than your average human driver. You're so missing the point, it's like you're criticizing a baby's ability to speak.
This is a BEAUTIFUL example of how current commercial-grade AI can fail to incorporate basic priors about the world, sometimes with deadly consequences. Thank you, @PositiveExisthttps://twitter.com/DiveraAI/status/1025810395527831554 …
-
-
-
Unfortunately that's not how machine learning works. There's no explicitly written program in there where you can just add that as a condition. You have to hope your test data covers all the cases.
- 4 more replies
New conversation -
-
-
Gee whiz ... how hard could it be to program in the idea of persistence of matter through space and time?
-
Very hard.
End of conversation
New conversation -
-
-
The same is true for any car currently equipped with adaptive cruise control, or automated emergency braking. It sounds like a glaring flaw. Nope. These systems are designed to ignore static obstacles because otherwise, they couldn't work at all. https://www.wired.com/story/tesla-autopilot-why-crash-radar/ …
- 1 more reply
New conversation -
-
-
Did you guys see the announcement of the new custom ASIC Tesla will be releasing, 2000FPS at full resolution! I bet Jim Keller leaves when they roll it out to work on custom AI ASIC's for other applications.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
It is not only incorporating priors but in general working in real-world conditions
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.