sorry, been sick. *looks suspiciously at "low temperature pasteurized cream"*
This was ~a couple years later: web.archive.org/web/20001212...
I could elaborate on my first reaction but, like, the principal component is I'm a lot more with him than his haters
This was ~a couple years later: web.archive.org/web/20001212...
I could elaborate on my first reaction but, like, the principal component is I'm a lot more with him than his haters
The Low Beyond
web.archive.org
November 29, 2025 at 2:27 AM
sorry, been sick. *looks suspiciously at "low temperature pasteurized cream"*
This was ~a couple years later: web.archive.org/web/20001212...
I could elaborate on my first reaction but, like, the principal component is I'm a lot more with him than his haters
This was ~a couple years later: web.archive.org/web/20001212...
I could elaborate on my first reaction but, like, the principal component is I'm a lot more with him than his haters
there was a good argument against "poverty of the stimulus"-type claims 30ish years ago, en.wikipedia.org/wiki/Rethink... (changed my mind somewhat)
and arxiv.org/abs/1104.5466 from 2011 was also forward-looking though it didn't envision pure Stack Moar Layers vs. trying to understand the workings
and arxiv.org/abs/1104.5466 from 2011 was also forward-looking though it didn't envision pure Stack Moar Layers vs. trying to understand the workings
Rethinking Innateness - Wikipedia
en.wikipedia.org
November 26, 2025 at 6:42 PM
there was a good argument against "poverty of the stimulus"-type claims 30ish years ago, en.wikipedia.org/wiki/Rethink... (changed my mind somewhat)
and arxiv.org/abs/1104.5466 from 2011 was also forward-looking though it didn't envision pure Stack Moar Layers vs. trying to understand the workings
and arxiv.org/abs/1104.5466 from 2011 was also forward-looking though it didn't envision pure Stack Moar Layers vs. trying to understand the workings
It’s actually short and I thought rather sweet, if you enjoy an absurd setup. I don’t remember it clearly. Different topics afair
November 25, 2025 at 11:51 PM
It’s actually short and I thought rather sweet, if you enjoy an absurd setup. I don’t remember it clearly. Different topics afair
can't deny this sound inductive reasoning. (personally wouldn't call the source 'God', that character gets mixed up a lot)
November 25, 2025 at 5:47 AM
can't deny this sound inductive reasoning. (personally wouldn't call the source 'God', that character gets mixed up a lot)
so much more lyrical! I'll check it out.
November 25, 2025 at 5:45 AM
so much more lyrical! I'll check it out.
I'd side with him though that it'll all be populated in the future (wish I were confident they'll be us)
November 25, 2025 at 5:44 AM
I'd side with him though that it'll all be populated in the future (wish I were confident they'll be us)
when I read what I guess was Eliezer Y's first webpage in the 90s I was like, well... what a strange mix of obviously bright and... foolish. Heard a bit later he was 17
November 25, 2025 at 5:32 AM
when I read what I guess was Eliezer Y's first webpage in the 90s I was like, well... what a strange mix of obviously bright and... foolish. Heard a bit later he was 17
oh, great! this one? web.viu.ca/johnstoi//ka...
Kant: Universal Natural History and Theory of the Heavens
web.viu.ca
November 25, 2025 at 5:24 AM
oh, great! this one? web.viu.ca/johnstoi//ka...
Opus 4.5 needed one hint, and didn't confabulate a different error. Really good by past standards.
(I haven't tried recent OpenAI models.)
This was an easy-to-grade test whether these AIs are good enough to improve my own work. I don't employ them that much so far.
(I haven't tried recent OpenAI models.)
This was an easy-to-grade test whether these AIs are good enough to improve my own work. I don't employ them that much so far.
November 25, 2025 at 4:33 AM
Opus 4.5 needed one hint, and didn't confabulate a different error. Really good by past standards.
(I haven't tried recent OpenAI models.)
This was an easy-to-grade test whether these AIs are good enough to improve my own work. I don't employ them that much so far.
(I haven't tried recent OpenAI models.)
This was an easy-to-grade test whether these AIs are good enough to improve my own work. I don't employ them that much so far.
curious remark from a writer imaginative enough to compose en.wikipedia.org/wiki/Univers...
(which admittedly I haven't read; all the Kant I've cast my eyes over was as much of a slog as everyone says)
(which admittedly I haven't read; all the Kant I've cast my eyes over was as much of a slog as everyone says)
Universal Natural History and Theory of the Heavens - Wikipedia
en.wikipedia.org
November 25, 2025 at 2:17 AM
curious remark from a writer imaginative enough to compose en.wikipedia.org/wiki/Univers...
(which admittedly I haven't read; all the Kant I've cast my eyes over was as much of a slog as everyone says)
(which admittedly I haven't read; all the Kant I've cast my eyes over was as much of a slog as everyone says)
fwiw Feynman apparently took something close to that as a goal: digital.archives.caltech.edu/collections/...
Richard Feynman’s blackboard at time of his death
digital.archives.caltech.edu
November 24, 2025 at 5:46 PM
fwiw Feynman apparently took something close to that as a goal: digital.archives.caltech.edu/collections/...
embarrassing for humanity to stall out at the level of carving chips out of stone, like cavemen
or maybe an unexpected triumph of wisdom, to pause short of superintelligence we don't yet understand
one or the other?
or maybe an unexpected triumph of wisdom, to pause short of superintelligence we don't yet understand
one or the other?
November 22, 2025 at 10:30 PM
embarrassing for humanity to stall out at the level of carving chips out of stone, like cavemen
or maybe an unexpected triumph of wisdom, to pause short of superintelligence we don't yet understand
one or the other?
or maybe an unexpected triumph of wisdom, to pause short of superintelligence we don't yet understand
one or the other?
Seventy-Two Letters - Wikipedia
en.wikipedia.org
November 22, 2025 at 10:23 PM
can confirm about alliterative msgs & a PhD, to some extent
(doubt my word game is on the same level)
(doubt my word game is on the same level)
November 22, 2025 at 4:51 PM
can confirm about alliterative msgs & a PhD, to some extent
(doubt my word game is on the same level)
(doubt my word game is on the same level)
It's barely hot enough to glow. You can see a candle flame is higher temp because it's yellow with a hint of blue in the middle -- same order as the rainbow. This inspired Betsy Ross's red, white, and blue! White instead of yellow because the color of heat is a broad mix of rainbow colors. . .
November 21, 2025 at 3:17 PM
It's barely hot enough to glow. You can see a candle flame is higher temp because it's yellow with a hint of blue in the middle -- same order as the rainbow. This inspired Betsy Ross's red, white, and blue! White instead of yellow because the color of heat is a broad mix of rainbow colors. . .
I meant your evals of Lecun and Hinton. But nm
November 20, 2025 at 5:44 PM
I meant your evals of Lecun and Hinton. But nm
I think Lecun is wrong about a lot too
and yet
and yet
November 20, 2025 at 5:36 PM
I think Lecun is wrong about a lot too
and yet
and yet
maybe I'll regret asking, but why do you think these two guys are dumb and dumber
November 20, 2025 at 5:29 PM
maybe I'll regret asking, but why do you think these two guys are dumb and dumber
(as a very amateurish graphics programmer sometimes back when Pixar was young)
November 20, 2025 at 5:09 PM
(as a very amateurish graphics programmer sometimes back when Pixar was young)
Neat. I assume the toy at Target did not actually do things, but I noticed I'd be tempted by one that could
November 20, 2025 at 5:07 PM
Neat. I assume the toy at Target did not actually do things, but I noticed I'd be tempted by one that could
like a toy Pixar lamp? I went by something like that at Target the other night
November 20, 2025 at 2:45 AM
like a toy Pixar lamp? I went by something like that at Target the other night
no idea! fun to think it could go well with some nontraditional style of actuators if they don't need to be fast & precise (make up for the latter with computation)
November 20, 2025 at 2:33 AM
no idea! fun to think it could go well with some nontraditional style of actuators if they don't need to be fast & precise (make up for the latter with computation)