Human-level AI is often forecast 20 years in the future. I think that's overly optimistic, by a large factor, and I am willing to bet on it
Show this thread
-
Here's a version of the test criteria that I believe focuses on the sensorimotor control aspects only.https://docs.google.com/document/d/1rQb9Flkt9En8rxu9s98hdwDS4XCRe7e9qzM_Ka4Itfo/edit?usp=sharing …
V. reasonable, but I don't think the proposed tests really capture the goal. They're more narrow; passing them wouldn't require cat-level AI
-
-
Whatever practical test we have should emphasize generalization power, adaptability, not only agility & 3D navigation
-
Ideally the tests could be run on actual cats too, so we can compare success rates and have a concrete "cat-level threshold" for the tests
End of conversation
New conversation -
-
-
Ah, I interpreted your one tweet more narrowly to focus on sensorimotor control instead of general cat AI. I missed the whole thread.pic.twitter.com/fczubEPSoP
-
I’m not sure general cat AI is really well definable as a test. I’m also not sure we’ll recognize superior AI as superior when we see it.
- Show replies
New conversation -
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.