Bloggingheads Community

Bloggingheads Community (http://bloggingheads.tv/forum/index.php)
-   Diavlog comments (http://bloggingheads.tv/forum/forumdisplay.php?f=9)
-   -   Science Saturday: Singularity Edition (http://bloggingheads.tv/forum/showthread.php?t=1867)

Bloggingheads 06-07-2008 10:41 AM

Science Saturday: Singularity Edition
 

osmium 06-07-2008 11:07 AM

Re: Science Saturday: Singularity Edition
 
good masthead on your blog, eliezer. wow, right on.

beren 06-07-2008 11:45 AM

Re: Science Saturday: Singularity Edition
 
I've been reading overcomingbias.com for a long time, more out of interest than because I agree with their world view. It's certainly one of the most pretentious and eliteist blogs on the internet. They need to learn humility.

nojp 06-07-2008 11:58 AM

Horgans point
 
I think the point is that science and humanity progresses more from basic science and Synchronicity rather than a linear aproach of pulling the future to us.

I.e. most progress penicillian, internet, pcr, dna ,amplification

rather than say the cure fore aids, polio, and the like

this breakthrough will happen by basic science not writing about the future if that is not to course.

nojp 06-07-2008 12:16 PM

Re: Science Saturday: Singularity Edition
 
How long until the turing test is past 10 15 50 100 years.... good god einstein

osmium 06-07-2008 12:51 PM

Re: Science Saturday: Singularity Edition
 
Quote:

Originally Posted by beren (Post 79885)
I've been reading overcomingbias.com for a long time, more out of interest than because I agree with their world view. It's certainly one of the most pretentious and eliteist blogs on the internet. They need to learn humility.

interesting. i'll give it a read. for now i just looked at the pictures.

edbarbar 06-07-2008 01:29 PM

Re: Science Saturday: Singularity Edition
 
What a lot of fun. I wish we could have heard more from Eli and less from John.

It's funny to hear John think the object is humanity. I don't see it that way. In some sense Chimpanzees created humans, albeit through an evolutionary process. Is the goal of humans to solve chimp issues?

Bloggin' Noggin 06-07-2008 01:32 PM

Re: Science Saturday: Singularity Edition
 
This was interesting. I enjoyed it and thank John for bringing Eli on as a guest and doing the interview.
I do wish John would have used his skepticism as a means of finding out more and revealing more to the BloggingHeads audience about Eli's views and the evidence for them than treating his own skepticism as a positive position he himself wanted to establish over Eli's objections.
A better way to say that might be that I wish John didn't confuse skepticism -- a general attitude of mind that believes propositions only in proportion to evidence -- and nay-saying (a definite position on a particular issue).
Don't ask Eli for strong evidence against his own views. Ask him for the best evidence FOR his views. Probe his evidence and see whether ultimately everything is resting on faith -- don't START with accusations that it's all "religious". That's about as illuminating as "have you stopped beating your wife?" Gotcha journalism is pretty useless in politics -- it seems even more out of place in science journalism.

I wish John could overcome his ADD (seemingly a disability afflicting most journalists). Why can't he be a bit more careful in defining his terms (like "religious")? If he has to make an accusation of that sort, why can't he make one accusation at a time, and let his poor interviewee respond? It's conceivable that this style of interviewing yields dividends for John when it is the background of a book, but as a real-time interviewing strategy, it isn't very good -- and in fact it seems more like hostility or lack of concentration than a means of helping your audience understand the interviewee's point of view -- and what may be wrong with that point of view.

Bob Wright and Will Wilkinson are the masters of interviewing on this site (and really they're better than any television interviewer I can think of).

I think Eli did a pretty good job of slowing John down and responding to one thing at a time, but unfortunately, as the interviewee, it wasn't in his power to impose coherence on the discussion as a whole.

P.S. To soften the harshness of my critique, I should say that John is pretty good at interviewing people he's more in sympathy with -- the environmental discussion a month or so ago was really good, as I recall.

piscivorous 06-07-2008 01:44 PM

Re: Science Saturday: Singularity Edition
 
I agree that this was one of the most disjointed episodes BHTV I have watched. Mr Horgan let his own skepticism and disbeliefs interfere with the flow and direction of the interview. He is not totally at fault in this as Mr. Yudkowsky seemed unable to address, the questions asked, in in direct manner and preferred word craft and word count as a substitute for substance. I am glad that ever time Mr. Yudkowsky tried to go down the path of "well in 10,000 years could you conceive of this happening" which inevitably leads to the well why not 1,000 then why not 500 then why not 100 mind game of gotcha.

ed fielding 06-07-2008 02:18 PM

Re: Science Saturday: Singularity Edition
 
Great divalog.

Congratulations to John for responding (however circuitously; I’m a circuitous guy myself) and rising to freshness of expression and productive persistence.

Congratulations to Eli for an admirable exposition of a good mind at work, and his encouragement for the rest of us to go and do likewise.

Hope we get to see and hear more of Eli.

Many thanks.

themightypuck 06-07-2008 03:14 PM

Re: Science Saturday: Singularity Edition
 
Horgan blew this one. I wanted to hear what Yudkowsky had to say and he kept getting cut off. Plus Horgan repeated one of the most annoying quotes by a smart person ever "Extraordinary claims require extraordinary evidence." There are claims. There is evidence. Once you throw "extraordinary" in there you are in the weeds.

Edit: this isn't to say there isn't strong evidence or weak evidence but my math and statistics skills blow.

ogieogie 06-07-2008 03:43 PM

Re: Science Saturday: Singularity Edition
 
What a disappointing mess. A fascinating guest--so Horgan interrupts him, insults him, and quibbles around the margins of his ideas, which consequently never get expressed. There was nothing to this diavlog except Horgan making an ass of himself.

So please, I'm begging you, have Elie on again with Bob Wright!

Wonderment 06-07-2008 05:29 PM

Prediction
 
As someone who also grew up in Orthodox Judaism (and escaped!), and as a fifth-rate amateur psychoanalyst, I can predict that a boy raised in Orthodox Judaism and obsessing throughout his childhood about the coming of the Messiah and the Messianic Age is likely to secularize and substitute Singularity for Messiah.

Such a person could even keep the rabbinical beard and continue to spend his adult life surrounded by nerdy autodidacts prone to endlessly but brilliantly debating the minutiae of visionary theories that the rest of the world views as esoteric and beyond the comprehension of ordinary folks.

Or... AI could be for real and he could win a Nobel Prize.

StillmanThomas 06-07-2008 05:56 PM

Re: Science Saturday: Singularity Edition
 
Quote:

Originally Posted by ogieogie (Post 79901)
What a disappointing mess. A fascinating guest--so Horgan interrupts him, insults him, and quibbles around the margins of his ideas, which consequently never get expressed. There was nothing to this diavlog except Horgan making an ass of himself.

I disagree completely. John is an excellent interviewer, but he's a huge skeptic about science in general and brain science in particular. I think Eli's intelligence is ponderous at best. I also think he was filibustering throughout the conversation, and John was laboring heroically to get him to stay on point. The point being that AI has a long history of inflated hopes and claims, and precious little to show in the asset column.

Quote:

Originally Posted by ogieogie (Post 79901)
So please, I'm begging you, have Elie on again with Bob Wright!

I agree that would be interesting.

osmium 06-07-2008 06:21 PM

Re: Science Saturday: Singularity Edition
 
Quote:

Originally Posted by beren (Post 79885)
I've been reading overcomingbias.com for a long time, more out of interest than because I agree with their world view. It's certainly one of the most pretentious and eliteist blogs on the internet. They need to learn humility.

i've watched the diavlog in piecemeal now (and also understand that overcomingbias is not just eli's blog, i.e. do not comment when you've watched one minute, etc).

it seems to me that the most pregnant part of the diavlog is this part, which is between the pre-chopped segments. twice eli alludes to progress in the AI field, but then john doesn't request an overview, nor does eli begin one on his own. i think that would have been a good use of ten minutes or so, especially since it seems like that's what john wanted to talk about re: skepticism.

i understand from the PDF article posted on the right side that eli is forming policy/philosophy on AI before "superintelligent" AI exists, and there's an analogy made with the first nuclear pile at the university of chicago in there that i thought was quite nice. but the second half of the diavlog would have been more useful if it had begun somewhere like an introduction to AI and then proceeded. coz, like, i gots no idea what AI people are doing, and i would be happy to learn.

StillmanThomas 06-07-2008 06:23 PM

Singularity Gets Lost in the Weeds
 
To me, this snippet illustrates the two fatal flaws with the whole singularity concept:

http://bloggingheads.tv/diavlogs/116...2&out=00:05:19

1. "Smarter than human." Human beings have many different types of intelligence. Eli, to me, seems very "intelligent" in a brute force kind of way. He's capable of bringing a lot of gray matter to bear on a problem. But that's not always the best way to solve problems, and indeed, there are many vitally important problems that would be totally opaque to that kind of intelligence. Think of what we face in learning how to cooperate within and across national boundaries to address renewable energy and climate change. With all due respect, I don't think Eli, smart as he is, would have a clue how to address this challenge. It will take all of the different kinds of intelligence we have to fix this: ratiocination, planning, exhorting, compromising, negotiating, and so on.

2. "We won't be able to predict the future." We can't predict the future now. So, predicting the singularity, beyond which we won't be able to predict the future is exactly the kind on inanity to which Eli's "intelligence" leads, IMO. Again, I don't mean to belittle Eli. I'm sure he's absolutely brilliant given a very narrow spectrum of problems.

I do thank both of you for your conversation, and I hope to see Eli back again. John, is a perennial favorite of mine.

EliezerYudkowsky 06-07-2008 06:56 PM

Re: Science Saturday: Singularity Edition
 
I thought I was the one who talked too much.

I also thought that the episode needed to be at least two hours longer to get to the interesting parts.

One question I fumbled was "What's the strongest opposition you've seen to Singularity ideas?" The basic problem is that nearly everyone who attacks the Singularity is either completely unacquainted with the existing thinking, or they just attack Kurzweil's technological determinism. There's no equivalent in Singularity studies of Richard Jones's critique of nanotechnology - which I don't agree with, but at least Jones has read Drexler.

People who don't buy the Singularity don't put in the time and hard work to criticize it properly. This is giving the advocates an unfair advantage!

What I should have done, though, was interpret the question more charitably as "What's the strongest opposition to strong AI or transhumanism?" In which case there's Sir Roger Penrose, Jaron Lanier, Leon Kass, William Hurlbut, and many others. None of these are good arguments - or I would have to accept them! - but at least they are painstakingly crafted arguments, and something like organized opposition.

-- Eliezer Yudkowsky

StillmanThomas 06-07-2008 07:09 PM

Re: Science Saturday: Singularity Edition
 
Eliezer:

I can't think of a single instance when a participant in one of these diavlogs posted to the comments section. I'm probably wrong--I've missed more than a few--but I'm sure it's pretty rare. Thanks for your thoughts and participation. Hope to see you back again soon.

Kevin 06-07-2008 07:57 PM

Re: Science Saturday: Singularity Edition
 
Quote:

Originally Posted by Wonderment (Post 79906)
As someone who also grew up in Orthodox Judaism (and escaped!),...

That is a clever and funny observation, Wonderment.

Generally speaking, it was too bad that they hopped around a lot. I have seen/heard some talks by singularity people, and I have heard some skeptics, (Bruce Sterling's Long Now talk comes to mind, other than JH) but haven't heard a pro and a con get into it with rigor. I wanted to hear more like what Yudkowsky touched on about the rate of change from 1940 to 1970, just as an example-- he has problems with one of the premises, I assume, for talking about the rate of change in the future based on the rate of change in the past, so it would have been nice to hear more with that kind of specificity.

Contrary to Horgan's remark about bloggingheads people having nothing better to do?? :) Hey, watching this episode was very pragmatic for me, that's why I watched.. because I feel as though I am in mortal danger over the next few years of blowing $100 on protein powder from Ray 'n' Terry's. I always gravitate towards the skeptics of Really Big Changes Pretty Soon, because I recognize that the true believers are seductive and compelling, and I can use all the critiques I can find. Including, more attention on whether futurists have a vested interest (like a nutritional supplements sideline) in generating buzz for their ideas.

It would be very interesting to see John Horgan diavlog Ray Kurzweil, "on the merits", more wonky and in the weeds than this one was.

bjkeefe 06-07-2008 10:58 PM

Re: Science Saturday: Singularity Edition
 
Thanks for saving me some typing, Bloggin. I pretty much would have said everything you did.

Just to elaborate a little bit:

John -- It's good to be a skeptic, and I don't want you to stop being one. But when a guy sounds this interesting, and clearly appears to have thought as much as he has about his ideas, I wish you would spend a little more time helping him flesh out these ideas before you get to debating them.

Granted, I could read some of his stuff and then more easily appreciate your quickness to disagree on certain points. Probably I will do some reading, and then come back and watch this one again.

It's always hard, I know, to strike the right balance when guessing how much background knowledge an audience brings to the table. All I can say, speaking for myself, is that you may have assumed a little too much this time around.

One more thought from my own perspective: It seemed at first glance that you were rejecting (possibly) new ideas from a perspective that is overly weighted by analogy with past events. Maybe you were right to do so, or maybe you were one of those people who laughed at the Wright brothers by pointing to the previous centuries' worth of failed attempts at flight. Again, always a hard call in these situations, especially given the amount of woo in the world.

No major complaints -- please take this as constructive criticism.

bjkeefe 06-07-2008 11:00 PM

Re: Science Saturday: Singularity Edition
 
Quote:

Originally Posted by Bokonon (Post 79916)
Eliezer:

I can't think of a single instance when a participant in one of these diavlogs posted to the comments section. I'm probably wrong--I've missed more than a few--but I'm sure it's pretty rare. Thanks for your thoughts and participation. Hope to see you back again soon.

Eli isn't the first diavlogger to jump into the fray, but that doesn't mean he doesn't deserve the shoutout.

Thanks, Eli. What Bokonon said.

AemJeff 06-07-2008 11:37 PM

Re: Science Saturday: Singularity Edition
 
I'll chime in too. I like John - his particular viewpoint is a good counterpoint to my own infatuation with wild, fundamental science. He asks a lot of good questions, and he often rightfully keeps things grounded.

Today, and Eli's protestations to the contrary notwithstanding, I repeatedly wanted to tell him just to let the guy talk. Eli's style of speaking may be somewhat discursive, but he speaks in paragraphs - there's structure to his replies, or so it seems to me. He kept setting up a theme, and before he could get to the point, John would step on his reply. I began to find that frustrating after only a short while.

Wonderment 06-08-2008 12:24 AM

Re: Science Saturday: Singularity Edition
 
Quote:

Probe his evidence and see whether ultimately everything is resting on faith -- don't START with accusations that it's all "religious".
I thought -- contrary to everyone else, apparently -- that John did a great job. He did apologize for the "cheap shot" of comparing Singulatarians to Christians.

John has a lot of experience in the debunking business and the interviewing scientists and philosophers business, so I tend to trust where he's trying to go as a journalist and what he's trying to illuminate in the given time limit.

He was obviously getting impatient that the clock was ticking (they went about 10 minutes over the 60-min. mark), and he wanted to get Eli to be a bit more substantive. John had done a lot of homework on the topic and -- rightly, I think -- wanted to get to the core controversial claims.

As for the "religion" charge, John is not entirely off base in calling attention to the quasi-religious, millennial, apocalyptical, messianic tone of some of the literature produced by Singularity writers. It's certainly true of Kurzweil and it's also comes through in the linked writing by Eli (see his speculative musings on how AI may suddenly reconstruct the entire solar system, treating humans as scrap metal).

That said, it's all fascinating stuff, and I enjoy reading and hearing about singularity speculation. A lot of very smart people are hard at work and it will be interesting to see how it all turns out (when I'm 12,000 years old or so).

Quote:

Bob Wright and Will Wilkinson are the masters of interviewing on this site (and really they're better than any television interviewer I can think of).
Hey, what about Carl Zimmer?

bjkeefe 06-08-2008 12:49 AM

Re: Science Saturday: Singularity Edition
 
Good argument for the other side, Wonderment. And I second your question about Carl Zimmer.

Ocean 06-08-2008 12:56 AM

Re: Science Saturday: Singularity Edition
 
Well, yes, somewhat chaotic... let's look at the bright side: it allows us to do more of the thinking and processing, instead of getting the digested edition.
John was stuck in his skeptical agenda. Eliezer wouldn't get to the point. The bottom line is that experts in AI have to retreat into speculation since they have failed miserably in their attempts to re-create something that doesn't even have a definition. So let's be honest, we still don't know what intelligence is about. Scientists don't agree. There are too many functions that we try to capture with the word "intelligence". And if we decide to throw in concepts such as "wisdom", we'll really get into obscurity. Wisdom includes many processes of synthesis and integration. It includes logic, knowledge, values, appreciation, hierarchies, consensus, common sense... Perhaps we need to consider all the contradictions that are contained in a wise thought. Wisdom is not found in the obvious but in the irony of life. Now go and create software that can imitate that and perhaps we'll have the utopian AI!
And still, I'm hopeful. Just any day...

jhorgan 06-08-2008 01:02 AM

Re: Science Saturday: Singularity Edition
 
I appreciate Elie’s gracious attempt to get me off the hook, but I think my critics have a point (and oh how I hate to say that). My goal in these chats is usually content-directed. In this case, I wanted to introduce people to the Singularity, as represented by Elie, serve up my criticisms, get his responses. But Elie has such an idiosyncratic perspective—he really represents only himself, not any widely held view, and I honestly mean that as a compliment—that I had no chance of achieving my goal. I should have abandoned my script and just let this intelligence-obsessed intelligent person riff. That would have been much more interesting than a half-assed, not-very-informative, quasi-argument about the Singularity. Hey, I’m still trying to get the hang of this Bloggingheads thing, and the biggest challenge for me is finding a balance between eliciting and questioning guests’ views. But I can’t resist making this last obnoxious point to Elie. I think you’re wasting your talents on the Singularity. I predict—I hope--that you move on to other more fruitful obsessions.

Ocean 06-08-2008 01:18 AM

Re: Science Saturday: Singularity Edition
 
I would ask the experts: Is there such a thing as "too much intelligence"?

I tend to think that there is. It manifests when an answer to a question gets entangled with so many irrelevant points, that neither the answer nor the point shine through. It is as if the point of wisdom was lost in chaos. I think that both intelligence and wisdom should be crystal clear, like art, to anyone. Yes, I tend to agree with Einstein on that...

John, don't sweat it. You are just fine...

edbarbar 06-08-2008 01:28 AM

Re: Science Saturday: Singularity Edition
 
I think you missed the point. The 10000 year point is it's hard to imagine it won't happen. That humans will be unable to develop the science to understand the principles and processes behind intelligence. It almost seems to me it takes a leap of faith to believe we can not.

The time scale is really meaningless, and while he put up some youthful optimism about it in his 20 to 50 year projection, I think that was more one of the factors that keeps him personally driven. I suspect if you pressed him he might say, "yeah, maybe it will take 1000 years, but it seems unlikely."

The thing I like about this talk is it brought me to pondering yeah, there will be a time when we are apes among men. This new world might well materialize in my children's lifetime.

edbarbar 06-08-2008 01:45 AM

John Interviews the Wright Brothers
 
The problem with John's arguments is they are non-substantive. They all had to do with so and so failed. "The Japanese failed." "Brilliant Pebbles failed." "So and so failed."

So what. How many light bulbs did Edison try before he got one that worked? How many people tried flight before the Wright brothers succeeded? Thank goodness there are people like Eli who ignore the John Horgans of the world.

That having been said, no one knows how hard the problem is. Maybe it is a 20 or a 50 year problem. Maybe it is longer. Regardless, it is hard to imagine human beings with their intelligence not solving it in a million years, or even 10000, given where we are today.

The hardware will be there. Silicon is much faster than neurons, with switching speed millions of times faster, and beyond that there are new, potential technologies that are cool and three dimensional. Replicating the compute power of the brain seems quite solvable, more a problem of technology than science.

So it will be a matter of software, and understanding the principles behind intelligence. If you believe the majority of that arose recently, the search space shouldn't be too vast, and the problem will be solved.

Happy Hominid 06-08-2008 02:15 AM

Re: Science Saturday: Singularity Edition
 
What Brendan said.

As big of skeptic as I tend to be, I thought Eliezer made some great points, such as why would you listen to someones optimistic prognostications, be excited, then when it doesn't happen and the person you were listening to becomes disillusioned, you become so also? And, that something failed to materialize is not, in itself, proof that it will fail to materialize.

Further, when you look at the advancement of humankind over the past 500 years, and then try to extrapolate out 1,000 or 5,000 years, it should be obvious that things will be incredibly different and that you probably can't begin to grasp how different it will be. Imagine being of average intelligence in 1508 and trying to contemplate nuclear power, human genomes and robots exploring the surface of Mars. Factor in the exponential growth of information and achievements and project that out to the coming 500 years and you see the problem with discussing these things.

But since this is completely beyond rational discourse, we kind of have to stick to 20-30 years and base our projections on where we are right now. I would accept that we are probably moving faster and doing more than we realize in the field of AI and even that rather short period could end up incredibly different than the world we are in at the moment. A Singularity world? hell, I don't know. I'm just along for the ride and fascinated by each new advance.

I just know that Eliezer has a jug of water that should get him through the next 30 years.

bjkeefe 06-08-2008 02:21 AM

Re: Science Saturday: Singularity Edition
 
Ocean:

Quote:

I would ask the experts: Is there such a thing as "too much intelligence"?

I tend to think that there is. It manifests when an answer to a question gets entangled with so many irrelevant points, that neither the answer nor the point shine through.
Being neither an expert nor in possession of too much intelligence, I answer anyway.

This'll quickly devolve into semantics, I expect, but I would say, no -- there is no such thing as too much intelligence. By definition, you would not be hampered by irrelevancies if you had enough intelligence. It's sort of like Eli's point about there being no good arguments against what he believes -- if there were, he'd believe them instead.

Now, in a more colloquial sense, I can go along with the familiar image of a person who overthinks a problem -- Jimmy Carter in the White House is the canonical example.* So in this sense, I guess the answer to your question would be yes. But I think you were asking something deeper than that. And in any case, I'd return to my first answer and say that the idea, in my example, that Carter allowed himself to get bogged down in minutiae indicates that he did not have enough (of a certain kind of) intelligence; e.g., he was not smart enough about delegating and prioritizing.

The one case where I could maybe agree that "too much intelligence" might exist would be in some 6-sigma type of person, who sees things so much more clearly than the rest of society that he or she can't get along.

=====================
*Play along here. I'm not claiming this is absolutely true.

AemJeff 06-08-2008 02:38 AM

Re: Science Saturday: Singularity Edition
 
Quote:

Originally Posted by Happy Hominid (Post 79940)
Further, when you look at the advancement of humankind over the past 500 years, and then try to extrapolate out 1,000 or 5,000 years, it should be obvious that things will be incredibly different and that you probably can't begin to grasp how different it will be. Imagine being of average intelligence in 1508 and trying to contemplate nuclear power, human genomes and robots exploring the surface of Mars. Factor in the exponential growth of information and achievements and project that out to the coming 500 years and you see the problem with discussing these things.

There be dragons. If there's to be a "singularity," extrapolation is meaningless. You can't plot a curve beyond that point, you can't pierce the veil. I haven't read any of Eli's papers (I will soon), but Kurzweil's Utopian fantasy seems a little precious. ("Newcomer to the namespace" - that was delicately put!) The singularity is a scary idea, looming in our possible future, a point at which the rate of change (of what? even that becomes hard to nail down) seemingly becomes infinite - or at least unpredictably nonlinear. -- Our robot overlords may require our body heat to run their civilization; or we all live forever in a virtual paradise, Gods creating new universes in our copious leisure time; or, most likely, something else, about which we don't have the vocabulary or the grammar to even compose descriptive sentences.

bjkeefe 06-08-2008 02:44 AM

Re: Science Saturday: Singularity Edition
 
AemJeff:

Quote:

("Newcomer to the namespace" - that was delicately put!)
Wasn't that outstanding? Totally pinned my geek meter. I love when an insult is constructed so that only those in the know will even be aware that it is one.

Man, I wish I had dingalinked that.

Ocean 06-08-2008 03:02 AM

Re: Science Saturday: Singularity Edition
 
The trick in the question is in the "too much" part. "Too much" is always relative to something else and intrinsically subjective.
The solution for the 6-sigma would be to blur his/her vision by wearing the wrong glasses and then get along. I just don't think it would be much fun.

Brendan, I agree with you it is about semantics after all. We make up these words such as "Intelligence" and then we are stuck for great part of the history of humanity trying to figure out what we meant by it! I think we should just drop it and come up with better ones. Unfortunately humankind loves to ruminate about its own creations as if they were real outside one's head. I guess it is the belly button adoration syndrome... Language has allowed so much progress and yet so much confusion...

EliezerYudkowsky 06-08-2008 07:07 AM

Re: Science Saturday: Singularity Edition
 
John Horgan wrote:
Quote:

I think you're wasting your talents on the Singularity.
That's actually one of the funniest things that anyone has said to me. I think ever.

I reply in all seriousness that the most important thing an intelligent person requires in order to avoid wasting their intelligence, is an important problem.

The important problems are not the ones that everyone immediately thinks of when you say the phrase "important problem", like string theory or global warming.

The important problems are the ones that, frequently, almost no one is working on, because (1) it is not immediately clear what you would do to solve them, and (2) some of the fundamentals are still confusing.

I believe that the terms commonly used to designate properties (1) and (2) are, respectively, "impossible" and "unknowable".

To bring order out of scientific chaos, you must first find a field in chaos. Then, directly confront the chaos instead of sticking to the islands of order. This requires pursuing, for years, an agenda that may not lead you to reliably publish one paper per month.

One of my pet-peeve subjects is how there is no systematic training in handling scientific chaos, probably because it only comes along once a generation or so. So everyone who does confront chaos has to figure out how to do it on their own, all over again.

The first lesson is not to run away. Science has handled difficult problems before.

The field of AI is in a state of chaos, fear, confusion, dismay, and disorder, at least outside a few well-lit islands of sanity.

Anyone who does not recognize this as a promising state of affairs has absolutely no sense of scientific history. An opportunity like that comes along once in a generation.

...which isn't the real reason I'm working on the Singularity, of course, but it is a point that needs to be made.

Incompetence Dodger 06-08-2008 10:38 AM

Re: Science Saturday: Singularity Edition
 
We're up to around 13 diavlogs per week, including a Science Saturday that is consistently worth watching twice. My record of listening to every diavlog since the beginning is in serious danger of being broken.

Meh, I'm not as put off by John's 'tude as some others are. He's been on SciSat long enough that we should all know that he's, to say the least, not a disinterested observer of the AI, neuroscience, etc. scene (much as we all know that Conn Carroll is not an impartial observer of the political world, and yet he has interesting things to say much of the time). Frankly I think everybody is just peeved because John threw the bhTV audience under the bus in the first three minutes (take a tip from Mickey and wait until the end of the diavlog to do that).

I was really impressed by Eli. In particular, there were three things he said that really set off a fire in my mind: 1) the idea that the reverse of stupidity is not necessarily intelligence: 2) the idea that the more brainpower you bring to bear on a misbegotten idea, the worse you can make things; and 3) that you need to bring most skepticism possible to bear precisely on those hypotheses that you want to be true (I've been beating the drum for this one for years).

However, precisely because Eli was so impressive, I'm going to have to call him out for this. Oh come on, Eli, that's a sub-David-Frum-level attempt at slight-of-hand. (The 10,000 years business is a bit of a cop-out, too).

I wish John had brought up what I think is one of the most effective notes of skepticism with regard to the singularity: namely, that it's an example of what I call the reverse Malthusian fallacy. By which I mean that you can take any existing accelerating trend and project it out to come to an absurd (yet compelling) conclusion. I worked on Wall St. at the turn of the century; believe me, I had a front row seat on one of the most notable recent examples of this fallacy. Given that a reverse Malthusian effect has never before in history continued on to the logical conclusion, I think the burden of proof is on the Singularitarians.

Whatfur 06-08-2008 11:26 AM

Re: Science Saturday: Singularity Edition
 
From "Cognitive biases potentially affecting judgment of global risks"
(Yudkowsky)

"...A followup study by Combs and Slovic (1979) tallied reporting of deaths in two newspapers, and found that errors in probability judgments correlated strongly (.85 and .89) with selective reporting in newspapers."

as I hate when the Science Vlog turns political (even by example..cough cough). I will only refer to this for inquiring minds.

themightypuck 06-08-2008 12:58 PM

Re: Science Saturday: Singularity Edition
 
Thanks for responding Eliezer,

If there was one question that I wanted explored (probably due to my non-scientist status) it is: what is the difference between the uncertainty at t=now and t=singularity?

themightypuck 06-08-2008 01:16 PM

Re: Science Saturday: Singularity Edition
 
To be fair. It's been people willing to go where "there be dragons" who have changed the world.

StillmanThomas 06-08-2008 01:21 PM

Re: Science Saturday: Singularity Edition
 
Elieazer:

Beautifully and compellingly stated!


All times are GMT -4. The time now is 06:52 PM.

Powered by vBulletin® Version 3.8.7 Beta 1
Copyright ©2000 - 2021, vBulletin Solutions, Inc.