Accelerating Change 2005

AC05 happened this last weekend.  Although last year’s conference was specifically focused on digital worlds, this year’s topic of Artificial Intelligence and Intelligence Augmentation provided plenty of fodder for TN readers.  First of all, Vernor Vinge – who’s book “True Names” described the reality of immersive, graphical spaces as well as anyone and predates “Snow Crash” by over a decade – talked about the singularity and the issues that could determine whether it is a “hard” or “soft” takeoff.  He focused on the development of “creativity and intellect that surpasses present day humans” as the next radical change and compared it favorably to previous transitions related to fire, agriculture, and the printing press. 

Read on for more . . .

[EDIT: Added data in body in support of the ASF folks being on top of things!]

The evening before I got to switch into complete fanboy mode and help demo Second Life to Vernor.  It was so cool because he kept asking questions in the form of “Could you  . . .” or “Are people . . .” and we’d go somewhere in world to see it.  It was about the coolest thing ever.

Sorry, end of fanboy interlude.

Ray Kurzweil followed and also focused on the singularity, with particular emphasis and defense of his concept that progress itself is an exponential curve.  He focused on how technological transitions generate overlapping S-curves of progress, netting out at exponential growth.  He uses these trends to even predict a date for the singularity of 2045!  The comment that leapt out at me was that “every form of communications technology is doubling price-performance, bandwidth, capacity every 12 months” with the graphs to defend it.  Fascinating stuff as we consider how online worlds expand, move onto mobile devices, etc.

Among other speakers, one who really leapt out at me was David Fogel.  His discussion of evolving neural nets to play checkers and chess was very cool.  Much like human learning, they did not embed winning conditions or strategy, simply embedded how the moves worked and then started evolving.  “Blondie24”, as the final net was called, is rated 2045, or “expert” at checkers.  They then built a chess simulation using similar methods, resulting in an evolved neural net with a rating of 2650.  Profoundly interesting, since – like our learning – the neural nets are building concepts like “how to win” from scratch.  Cool.

Two different speakers from Numenta talked about the new company’s attempts to leverage Jeff Hawkins’ ideas about the neocortex from “On Intelligence.”  Their demo looked a lot like typical neural net driven recognition demos and generated some serious pushback from the audience.  However, if you’ve read the book you already know what they’d cover in their talks.

Another favorite of mine, Tom Malone, spoke about the ideas from his book “The Future of Work.”  Like “On Intelligence”, if you haven’t read it, you should.  It focuses on the impact of decreasing communications costs on both human societies and on business.  He is a strong advocate of the power of decentralized decision making, management, and production, and his work is nicely paired with Surowiecki’s “Wisdom of Crowds.”  On an immediate level, one can see how Malone’s ideas have to apply to group formation and mechanics within digital worlds, where communication costs are even lower than the real world.  Looking forward, his ideas support the contention that many forms of real work will migrate into digital worlds, driven by the ability to form more effective, efficient, and innovative businesses within them.

I gave a late night talk entitled “How Games Will Save the World” that defends the idea assertion that “games will save the world by teaching the majority of people how to critically filter the increasing information they are exposed to.”  The talk will apparently be up on IT Conversations – along with all of the AC05 presentations – in a few weeks, so I’ll post a more complete discussion when that happens.

So, AC05 was another strong conference from the Accelerating Studies Foundation.  My one area of concern is that anyone who wants to talk about the future ends must be able to both have an open mind about new technologies and ideas, but as Feynman said, “not so open that your brain falls out.”  And this is clearly a difficult balance to maintain.  Dr. Daniel Amen’s talk stood out as an example.

Dr. Amen is an incredibly dynamic speaker and proponent of using singly photon emission computer tomography (SPECT) imaging as a diagnostic tool for a wide variety of brain disorders.  His presentation is amazing, with areas of reduced blood flow showing up as vivid holes in some of the scans, while normal people show healthy, even flow.  He acknowledges that many feel that he is a quack, but continues on with his presentation.  He clearly believes in what he talking about and is passionate about brain health, especially in children where is a strong opponent of sports like football and soccer.  However, the use of SPECT as a diagnostic tool is quite controversial and Dr. Amen has his detractors.  To a skeptical observer, aspects of his pitch – his online ADD and Brain System tests or his failure to discuss the mechanisms that link holes in the brain scan to actual medical problems – read like quackery.  But this isn’t the point.

The point is that it wouldn’t be hard for Dr. Amen to establish scientific credibility for SPECT if it actually works.  He would be able to cite independent, double-blind studies that look at the efficacy, examine both the false positive and false negative rates, be able to talk about the repeatability of the scans, etc.  If those studies haven’t happened yet, how about sharing his own statistics.  Sure, they’ll need to be taken with a grain of salt, but seeing his claims, for example, about the relationship between the online tests would be very interesting.  But none of that happened.

WoW players would never accept data about quests or loot this uncritically.  I know what questions I would have wanted answered.  The fact that they weren’t in his presentation makes me think that the ASF didn’t ask them either and I think that is unfortunate.  The ASF is working hard to legitimize an important and easily misunderstood profession – that of the professional futurist.  IMHO, if they are going to succeed, it is critically important for them to avoid even the appearance of fortunate tellers or soothsayers.  They wouldn’t turn to astrology to predict the future.

I’m not saying that they shouldn’t invite someone like Dr. Amen to speak.  I think that they should, but they should strive for the same level of rigor that Ray Kurzweil brings to the table.  Ray’s talk was, if anything, even more preposterous than Dr. Amen’s, but extraordinary claims must be supported by extraordinary evidence.  Ray went to great pains to provide it, Dr. Amen did not.

[EDIT: ASF actually had the same quackwatch link I used in Dr. Amen's bio. So credit to them, although it was several clicks deep so I hadn't found it.]

Thanks to all of the ASF folks who put on a great conference – I can’t wait for next year!

Comments

[monkeysan]

1.

Cory>I think that they should, but they should strive for the same level of rigor that Ray Kurzweil brings to the table. Ray’s talk was, if anything, even more preposterous than Dr. Amen’s, but extraordinary claims must be supported by extraordinary evidence. Ray went to great pains to provide it, Dr. Amen did not.

I almost have to laugh at this--it is only my respect for you, Cory, that supplies the necessary restraint. Kurzweil, to my mind, is a perfect example of someone who is fascinating and brilliant but completely lacking in philosophical accuity. He does attempt to supply a level of rigor to his work, I suppose, but his most important claims--the 'inevitability' of the 'singularity', uploading our consciousnesses, his 'laws' about evolutionary 'progress', etc.--are as shining examples of unsupported handwaving as one could ever hope to read and equally lacking in empirical support.

I guess I'm just rattling on to say that I agree with your sentiment about futurism--in order for it to be taken seriously and become legitimized, it could do with a bit more empirical grounding and a lot more philosophical clarity.

Aaron

[Brian 'Psychochild' Green]

2.

WoW players would never accept data about quests or loot this uncritically.

Huh? They do all the time. Most people I know treat the online databases as pretty much accurate, allowing for exceptions arising from patch changes or the occasional oversight. I've even had a few times where someone asked for quest help but had completely wrong information and refused to believe me when I told them different. But, overall, nobody I knows requires external third-party verification of data found in the online databases before they accept it as fact. I've certainly never heard of double-blind studies to verify loot information. ;)

Not to say that we shouldn't require more proof from people when they make claims like you discussed, but you picked the wrong contrast group, IMHO.

[Cory Ondrejka]

3.

monkeysan, I generally find Ray somewhat kookie as well, but the amount of data that he has generated supporting power laws in evolution and human development is pretty impressive. There is an entirely separate debate about whether or not that creates the singularity or not.

Brian, the amount of work that MMO players will go through to find optimum paths is really impressive and they will share that data with each other. Agreed that it's a stretch to call that science, however.

Note edit in body. Jerry P pointed me toward the same Quackwatch link I had used that the ASF had placed in Dr. Amen's bio. Kudos to ASF for adding it, although negative kudos for burying it several links down.

[Torley]

4.

OMG... Ondrejka and Vinge! What a duo! That was my favoritest part of reading this. So what'd he enjoy most? Is he gonna sign up and come inworld more? ^_^

[csven]

5.

I look forward to hearing the talks on ITC. Thanks for the overview.

[Brian 'Psychochild' Green]

6.

Brian, the amount of work that MMO players will go through to find optimum paths is really impressive and they will share that data with each other. Agreed that it's a stretch to call that science, however.

Right, but not everyone goes through the calculations to do so. Generally it's a few inventive souls (aka Explorer's in Bartle's original grouping) that go through the calculations, then lots of other people cribbing the information. Many times people stick with that information even when things change. I've seen many times where people stick to the "old ways" without critical thought even though the game has changed because that's the way they've always done it.

Anyway, cool writeup, Cory. I was thinking about heading to the conference, but I've been up to my eyeballs in consulting work right now.

Have fun,

[Rob Sperry]

7.

Aaron,

I have been looking for good criticism of Ray's work but have been unable to find any. By good I mean criticisms that responds to his arguments and with counter arguments and alternative explanation for the data he presents or refutation of the data .

So do you have anything here? Why do you think think increasing computational power combined with better AI theory and brain reverse engineering wont lead to greater than human level intelligence? (ie the singularity)

P.S AC2005 was great fun :)

Rob Sperry