"lawyerly" - Yes, I agree completely. Function doesn't follow form. Prof Dave's screeds
often seem like he's arguing from the dictionary. (I don't think he has a PhD. I could be wrong.)
But his, perhaps naive, hard line feels like the opposite of what Fridman does, which is befriend
even the most toxic people in the name of "love". We all sit on a spectrum of tolerance
of intolerance.
And, FWIW, I also like Sabine's parasocial personality more than P. Dave's.
C'est la vie.
Sabine's moral stand - IDK. If she wasn't monetizing that moral stand, I'd
agree. But she is monetizing it. And even if, for now, the fact that she makes
money from hyperbole, ... even if that hasn't coerced her moral stand toward
the reinforcement, she's human. And we humans are susceptible to reinforcement.
Is it really a moral stand if you're getting paid to take that stand? She's a
professional reactionary, not so different from a lawyer or politician, however
wonky.
Anyway, we'll see where her advocacy arcs over time.
On 11/20/24 09:52, Santafe wrote:
It’s funny.
I have some sympathy to Sabine’s position, and I “like” her more than Dr. Dave,
on a couple of counts. Not that any of the practical concerns raised in this
thread are empty or should be dismissed. They just seem to me (for my own
concerns) like lower-order matters.
The through-line I see in a string of videos of this kind — including another
one that comes up on feeds to this one,
maxresdefault.jpg
Science is in trouble and it worries me.
<https://www.youtube.com/watch?v=QtxjatbVb7M>
youtube.com <https://www.youtube.com/watch?v=QtxjatbVb7M>
<https://www.youtube.com/watch?v=QtxjatbVb7M>
— is that there are fewer and fewer consequential _ideas_ seeming to be
created in physics, and although I am not in touch with all of science, I have
enough traffic in enough areas that I think something like this is true more
broadly.
One of the symptoms of this that I have been discussing with people for some
years is that, increasingly, the Nobels in physics have really been awarded for
engineering. In a few cases, like the Higgs, they were for an idea many
decades old, which just took a lot of time to bring within reach of
measurement. Not that the Nobel matters much one way or another, but it is
okay as a kind of cultural barometer. Every year they have to come up with
_something_, and what they do come up with can reflect what the current
sociology has on offer.
I am stunned to realize that the last one I paid a little attention to was
2021, and of those three, Parisi is the one who I think really matters in an
idea sense (as opposed to important application domains). Even for Parisi,
among the several things he has done, I think glasses stand out as the thing
that brought a new idea into existence, that was really, deeply hard to have
seen and then been able to develop. For the first decade or more, it looked
like total voodoo (replica methods, specifically), and I say that even having
used a little shred of that math to do something many years ago (localization
in acoustics). Parisi could see through that house-of-mirrors and manage to
develop it, for a long time when only a couple guys in the world seemed able to
follow it, until more sane and understandable methods could come in and make it
systematic and accessible.
More recently, what do we have? Engineering of entangled photons and
attosecond laser pulses? Hopfield for neural nets (okay; good one; handy that
he has lived to 110 or whatever it is, so that work he did as a boy could get
some registration before he died).
In the video linked above, SH wants to conclude quickly that, since the
slow-down is across the sciences, the cause is social and institutional. I
think that’s too quick. My confound would be that, where there were rapid
“discoveries” in lots of sciences that were still somewhat “new” in the first
half of the century, the reason they came fast and thick was that many of them
resembled, in what is needed to think about them, ideas that were already in
place or still being discovered in physics: crucially, they were “simple” and
“mechanical” in some way, and didn’t reflect essential things across the
sciences that made them really new horizons. So a combination of the
“low-hanging fruit”, and an argument that the discoveries aren’t conceptually
so independent, is an important part, together with Sabine’s institutional
claim, which is also important (IMO).
Re. the latter: Surely, the things I most want to do, I could never get funded, because I don’t already understand them. And the things I can get funded, I mostly don’t care about any more, because I learned from them what was significant to me. Between that difficulty, which puts my more interesting efforts in the garage, and the fact that I’m not all that promethean to begin with, I would put most, and perhaps all, of what I have ever done on a margin of Sabine’s “bullshit” category. Though I keep hoping that if I can just try one more time, I might do something that matters more…. There doesn’t need to be any game-playing in making a self-assessment like this, and I would still like to argue for a large and active field of scientific effort, even if much of it is applied and incremental. Among things people can spend their time on, this certainly seems absolutely better for human futures than endless contests of attacking neighbors and capturing land. We can admit
both that important ideas are rare, and also that modest ordinary day-work is perfectly fine, if it doesn’t devolve into an unanswerable priesthood.
This is not to say that I think all of Sabine’s arguments are good, even though
I recognize that, technically, she is far stronger than I am and knows far more
hard stuff than I do. When she was on her anti-beauty kick, she used to trot
out renormalization as an example of a solution that was true and ugly. But
she adduced the 1930s subtraction-of-infinities version of renormalization,
which to me was a strawman. It is both ugly and not the real general form.
The general form is the renormalization group, which is utterly elegant. And
surely she understands all this, if I do. So I don’t go for it all.
Instead of conservative, I wonder if some of it is just Aspbergers. I have
Greta Thunberg’s voice in my ears, explaining to people that “we autistics see
the world clearly, in black and white. Not like you normies, who seem to [in
your distorted way] keep saying everything is gray.”
Oddly, in the segment that Glen originally forwarded from Dr.Dave, I was aware watching it that I recognize a thing about people who are primarily science communicators in it, that before now I did not have a way to articulate. But I know a bunch of them, and I think I can say it now. They remind me of lawyers. Like they are hired advocates for some position, and the moral standing of the thing they advocate for is not their remit. Hence Dr.Dave’s (to me) handwringing that “she’s abetting the enemy”. No; the enemy acts in bad faith, One doesn’t want to PC everything down to ditchwater trying to be defensive against them. One could do that, and to some degree it works (incrementally), but at the end of the day that is less valuable than a moral stand taken by somebody on something that mattered, if the stand showed some depth. I see SH as trying to do this (and also doing her own research, which is the “put up or shut up” half of that bargain). Right or wrong in its
different choices, and probably overstated much of the time, it seems okay to me.
Eric
On Nov 18, 2024, at 7:54, glen <[email protected]> wrote:
Yeah, it's kinda sad. Sabine suggests someone's trying to *deduce* the
generators from the phenomena? Is that a straw man? And is she making some kind
of postmodernist argument that hinges on the decoupling of scales? E.g. since
the generator can't be deduced [cough] from the phenomena, nothing means
anything anymore?
What they're actually doing is induction, not deduction. And the end products
of the induction, the generative constraints, depend fundamentally on the
structure of the machine into which the data is fed. That structure is
generative, part of the forward map ... deductive. But it's parameterized by
the data. Even if we've plateaued in parameterizing *this* structure, all it
implies is that we'll find a better structure. As Marcus and Jochen point out,
it's really the same thing we've been doing for decades, if not centuries, in
many disciplines.
So her rhetoric here is much like her rhetoric claiming that "Science if
Failing". It's just a mish-mash of dense semantic concepts arranged to fit her
conservative narrative.
On 11/17/24 08:45, Roger Critchlow wrote:
Sabine is wondering about reported failures of the new generations of LLM's to
scale the way the their developers expected.
https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fbackreaction.blogspot.com%2f2024%2f11%2fai-scaling-hits-wall-rumours-say-how.html&c=E,1,0ohof8uVbmKPhWTHR2-_AGoL4rCwVbnuFeL01ra85Zt9dcttwUuvaSb7yjADkEbgbEWg8aqmWsrWEmljR0Cjz9GKANLoXKGbNu7TsYoRT6sEmMmymq7u&typo=1
<https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fbackreaction.blogspot.com%2f2024%2f11%2fai-scaling-hits-wall-rumours-say-how.html&c=E,1,SCXmdB5gBm48LPX6D5JgdR7uMPXLmAuG5N__SWCkM5Y-kt_WVCJCGAo9qqO2Ejsl-iyijqOfvHcE4ETfsYk7cyBtaqS_Dq95xZetlsd_WYSSihbiWJZ1RIQ,&typo=1>
On one slide she essentially draws the typical picture of an emergent level of
organization arising from an underlying reality and asserts, as every physicist
knows, that you cannot deduce the underlying reality from the emergent level.
Ergo, if you try to deduce physical reality from language, pictures, and videos
you will inevitably hit a wall, because it can not be done.
So she's actually grinding two axes at once: one is AI enthusiasts who expect
LLM's to discover physics, and the other is AI enthusiasts who foresee no end
to the improvement of LLM's as they throw more data and compute effort at them.
But, of course, the usual failure of deduction runs in the opposite direction,
you can't predict the emergent level from the rules of the underlying level.
Do LLM's believe in particle collliders? Or do they think we hallucinated them?
--
ꙮ Mɥǝu ǝlǝdɥɐuʇs ɟᴉƃɥʇ' ʇɥǝ ƃɹɐss snɟɟǝɹs˙ ꙮ
.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ...
--- -- . / .- .-. . / ..- ... . ..-. ..- .-..
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom
https://bit.ly/virtualfriam
to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
FRIAM-COMIC http://friam-comic.blogspot.com/
archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/
1/2003 thru 6/2021 http://friam.383.s1.nabble.com/
