4/ In ed, we have a variety of idiotic "alternate" education schemes (and a few good ones) because trust in the core institutions is gone.
Conversation
Replying to
5/ But messed up ed-science had a big unintended effect: IQ became the anchoring concept in AI research, a truly bizarre event.
5
3
9
Replying to
6/ IQ is a singular, metric-based scientifically meaningless model of cognition (reduced to a legible "intelligence") created for HR needs..
2
7
10
Replying to
7/ Anchoring as generative an idea as "artificial cognition" in a pragmatic, unscientific HR instrument leads predictably to silly AI.
1
3
3
Replying to
8/ The idea of a singular superintelligence that beats humans on one or even a handful of historically arbitrary metrics is NOT significant
1
1
2
Replying to
9/ The very idea of "intelligence" as a model of cognition shares common roots with the _testing_ of the intelligence aspect of cognition
1
6
6
Replying to
10/ This is ultimately why the idea of the Singularity is so meaningless.
1
3
3
Replying to
11/ Going back to the nutrition analogy, it's like worrying a robot will beat all humans on cholesterol and lung capacity tests.
1
2
2
Replying to
12/ Buzzfeed 2020 headline: "Superhealthy Artificial Organism beats humans on cholesterol test on national TV"
1
4
1
Replying to
13/ So what's a better way to think about AI: Pinker was right that Moravec's Pardox is the most significant idea in AI so far. Here's why
2
4
4
Replying to
14/ Moravec's paradox revealed that early AI was more about anthropomorphism in problem choices by early pioneers than about cognition
Replying to
15/ Due credit to pioneers, but it is no accident that they pursued same metrics that a) inform IQ b) they aced to become academics
2
2
Replying to
16/ Current AI research is extremely exciting not just because it is higher powered but because it is getting away from anthopomorphic roots
3
8
6
Show replies
