Log in

No account? Create an account

Previous Entry | Next Entry

two of the best talks I've ever been to

Whew... I am still catching my breath! These last few days have been incredible, especially last night.

On Thursday, I went to a Philosophy of Quantum Mechanics talk given by Richard Healey. Going in with pretty low expectations (and almost not bothering to attend at all), I was pretty blown away by how well he actually understands math and physics being in another field. Perhaps it's just from arguing too much with complete idiots on livejournal, but I was beginning to lose my faith that there were actually good philosophers out there. And unlike Bas van Frasen, who came two years ago, whose opinions on quantum mechanics (while not unreasonable) I really can't see myself agreeing with, Healey's views are right in line with mine, modulo a few issues that I raised to him after the talk, and emailed him more about after I got back. Actually, after hearing what he had to say on these issues, it made me realize just how much I have to say that I really ought to write down and publish somewhere. There are several specific problems he mentioned he was waivering on that I have strong opinions on (and good arguments to back them up). My email to him ended up being 5 pages single-spaced, and all I did was list the major points I'd like to publish on, and sketch very briefly the arguments that I'd use to back them up. I feel like each paragraph, I could expand to at least a page... so it would be almost no trouble at all for me to crank out an article on this. Which I intend to do, whether or not he emails me back, but it will be nice to get some feedback if he does. Perhaps he can at least tell me which journal would be the best to publish it in.

For an example of how cool Richard Healey is, here is a paper he wrote On The Reality of Gauge Potentials which argues that holonomies should be taken as fundamental to gauge theories, rather than the gauge potentials themselves. Brilliant! It makes me wonder, though, how many other philosophers can actually read an article where he makes use of fiber bundles, holonomies, and Wilson loops to make his arguments (all far above the usually required mathematical literacy level for philosophers, as far as I knew.) As for my email to him, I might make a friends-only post disclosing it soon, I haven't decided yet whether that's a good idea. Or maybe I'll wait and see if I get a reply first.

On Friday, I went to a talk by Eliezer Yudkowsky of the Singularity Institute for AI. I'd met the other two founders (Brian and Sabine Atkins) of the Singularity Institute years ago when I lived in Atlanta, but until yesterday I had never met Eliezer who is really the "brains" behind the operation (Brian and Sabine mostly being the "money" behind it). Eliezer's talk was absolutely mind-blowingly awesome. Although I've thought about things like emergence, evolution, and intelligence before, I've never connected them all together or thought about the full context for them as much as I did while listening to him talk. He is a really bright guy. The only annoying thing was that some of the questions from people in the audience were not exactly what I'd consider "up-to-par" with his level of discussion, and it slowed the whole thing down without adding much to it. Actually, I thought the same thing about the philosophy talk I went to, but I guess that just indicates how much a cut-above the usual lectures I go to they were. After Eliezer was done speaking, we all just hung out and chatted and I got to meet several more people from the Singularity Institute (such as Tyler Emerson, who will be speaking at our Santa Cruz Future Salon in April). I also talked to Catherine M (I forget her last name) who did all the research for Ray Kurweil's latest book The Singularity is Near and reads about 1000 books per year (and 1000 words per minute!); she gave me a lot more useful information about Modafinil which I've been thinking seriously about ordering recently (it's a drug that basically gives you a full night's rest in a pill... partially obviating the need for sleep, if used correctly).

While I still have a passionate dislike for the word "Singularity" because I think it's inappropriate and untrue to its meaning in physics, and I'm not convinced yet that the rise of strong AI will be all that rapid (and I don't see cause for any discontinuities in the predictability of the future), I really admire everything they're doing at the Singularity Institute and, if I weren't doing incredibly interesting stuff already that I'm somewhat committed to, I would really really love to help out. Especially with their "World's Most Important Math Problem" which they've been trying to recruit someone to solve. I also don't really blame them for picking that name, because I'm sure they get a lot more funding with it than if they'd picked something less sensational. It just leads to a few people showing up to their conferences who have no idea what they're talking about. Hats off to them, especially Eliezer.


( 13 comments — Leave a comment )
Feb. 26th, 2006 05:16 am (UTC)
What do you think of Kurzweil? I've always gotten the impression he's an arrogant nut who likes to make extreme statements that aren't justified, but I've never read any of his books....

As for holonomies, if you don't already know about it you might be interested in the Makeenko-Migdal formulation of Yang-Mills theory. They developed a set of equations describing the theory entirely in terms of Wilson loop variables. It's not easy to use these equations for anything, since loop space calculations are a mess, but it is nice to see that theory can be written explicitly in terms of gauge-invariant stuff.
Feb. 26th, 2006 06:01 am (UTC)

What do you think of Kurzweil? I've always gotten the impression he's an arrogant nut who likes to make extreme statements that aren't justified, but I've never read any of his books....

Yeah, I sometimes get that impression too... although likewise, I've never read any of his books so I have to reserve judgement. He might just be on the same nuttiness-level as Stephen Wolfram, which is not all that bad, but it's hard to judge these things without actually reading it.

Yudkowsky was not nutty at all. He is very careful about not making any outlandish claims. Personally, I think maybe the SIAI just chose the name "singularity" in order to up the potential for funding by drawing from Kurzweil's already dedicated fanbase. But I shouldn't really speculate much on this until I read his books. Some of the claims I've heard second-hand from Kurzweil I find totally outlandish, but maybe he has some good arguments for them, I don't know. I plan to read at least one of his books in the near future to find out.
Feb. 26th, 2006 08:16 pm (UTC)
I've read a couple of Kurzweil's books and tried to read the latest -- it's the least bad of those three, but it still irritated the hell out of me and I didn't finish, despite basically agreeing with the conclusion that we're headed for a singularity in the near future. His arguments are really sloppy -- I'd recommend Vinge's essay instead.

Hey, to sort of tie together the Singularity and QM topics from this post, have you seen Robin Hanson's work? Brilliant guy who's written about both.
http://hanson.gmu.edu/mangledworlds.html for his QM papers.
Mar. 4th, 2006 06:13 am (UTC)
I took a look at the mangledworlds essay... interesting. I was under the impression this problem was already solved in a more straightforward way, but I could be wrong. I'll have to take a closer look. I'm not sure I understand exactly what he's getting at with some of it.
Feb. 26th, 2006 08:45 pm (UTC)
Yudkowsky has interesting ideas and he's clearly very smart, though I don't expect him to succeed because that'd amount to getting ahead of an entire R&D community on a community-sized problem. That can be done, by the NSA for crypto for example, but it's not the way to bet.

But my main objection to his program is that I think creating a god to decide all our fates, even a super-ethical god, would be highly unethical.
Feb. 26th, 2006 06:03 am (UTC)
oh, and thanks for letting me know about the Makeenko-Migdal, I'll look into that. Is loop-space what loop quantum gravity is done in by any chance?
Feb. 26th, 2006 09:39 pm (UTC)
Loop space in this context means something really simple: it's just the space of loops over Minkowski space (or whatever you're defining your QFT on), i.e. of maps S^1 -> R^{3,1}. In quantum gravity I imagine it's more complicated since your space is dynamical, so it's not just loops over a fixed background. I don't really know much about LQG and, based on some of the blog comment threads Lee Smolin has spoken up in, I get the impression LQG probably doesn't even exist as a theory, so I try not to pay much attention to it.
Feb. 27th, 2006 05:43 am (UTC)
I thought the whole point of LQG is that it doesn't assume a background dependence.
Feb. 26th, 2006 08:52 pm (UTC)
Oh, what is the world's most important math problem? I couldn't find anything but a talk announcement online.
Mar. 4th, 2006 06:11 am (UTC)
Well, most of his talk was about that. Basically, the problem is to develop a certain type of recursively self-modifying algorithm and prove that it's stable under the property of "friendliness". In other words, the AI has the ability to self-modify its own code in ways which we cannot predict, yet if this friendliness property is defined in just the right way... it will not want to modify itself to be unfriendly and will take care not to. He's already made a lot of progress in finding a robust and rigorous way of defining friendliness, although I think there is still a bit to be done there as well as proving the stability under self-modification for some class of algorithms.
Feb. 26th, 2006 09:01 pm (UTC)
wow that sounds like an amazing weekend. I kinda wish I had been able to do my trip on SF on friday instead of this monday.

I love Tyler. So passionate and humble.

You'll have to post more about Modafinil as I've been contemplateing that as well for a year or so.

Eliezer is bright, but it's always to get him Ben Goertzel and Peter Voss together.

Singularity is useful as it can be controversial and from a particular vantage point relating two points it will seem unfathomable.. Nobody gets inspired by mediociryt.
Mar. 3rd, 2006 05:51 pm (UTC)
you're missing an adjective
"Eliezer is bright, but it's always to get him Ben Goertzel and Peter Voss together."

you're missing an adjective
Feb. 26th, 2006 09:34 pm (UTC)
Damn, I should have tagged along.
( 13 comments — Leave a comment )


domino plural

Latest Month

May 2017


Powered by LiveJournal.com
Designed by Lizzy Enger