hugetim

8 ms·

1.
ARC-AGI-2 human baseline surpassed(lesswrong.com)
9 points by hugetim 9mo ago|1 comments
2.
by hugetim 9mo ago
Contrary to what the leaderboard lists as the human score, their technical paper implies a human baseline of ~48%.
3.
by hugetim 9mo ago
Am I crazy to think these models have actually surpassed human performance on ARC 2? https://www.lesswrong.com/posts/DX3EmhmwZjTYp9PBf/ai-perform...
4.
by hugetim 2y ago
The package Jeremy Howard uses for Claudette, nbdev, actually does address the notebook pain points of version control, testing, integration, and deployment: https://nbdev.fast.ai/ (Debugging is already a strong point of de
5.
by hugetim 2y ago
Curious how folks think this measures up: https://www.answer.ai/posts/2024-06-23-claudette-src.html
6.
by hugetim 2y ago
A nice source of middle- and high-school aged problems: https://www.usamts.org/ One of the joys of high school was discussing these with a friend (after submission---but one could forego the contest and just do the circle t
7.
by hugetim 2y ago
For me it was this curriculum, which should be better-known: https://www.elementsofmathematics.com/ One place it is offered in-person is the St. Louis area where I grew up: https://megsss.org/about-us/
8.
by hugetim 2y ago
He's already written a bit here, starting in the second paragraph under "Why": https://about.fastht.ml/components
9.
by hugetim 2y ago
But is it reasonable to consider charlie worse than alice?
10.
by hugetim 3y ago
Clarify your thinking about what it means to ask "why" here: https://plato.stanford.edu/entries/questions/#PraAppExpCon
11.
by hugetim 3y ago
I'd love to see them implement a good liquid democracy system that would offer a better "middle ground between 'take every vote' and 'let Blackrock decide'."
12.
by hugetim 3y ago
I'm guessing you're not very familiar with Lessig's work: https://perma.cc/5XQ8-MBWA?type=image Or maybe you just have a different theory of change? His seems to be that our best shot is to use whatever lever