23 replies to this topic

[Anthony_Loera]

Ray Kurzweil does not understand the brain:
http://scienceblogs....ot_understa.php


The article: Good

The state it leaves you: Thinking Ray is crazzyyy

The comments below the article: Priceless!!

[okok]

Probably someone mentioned this in the replies, i just scanned them - 474 and counting, are hyou nuts - but i think Kurzweil is first a businessman and second a scientist. How much he believes of what he says i don't know. I try to read up on psychology, amongst many other things.

[aLurker]

Great article.

Speaking of priceless comments, here is the first one from the thread at slashdot:

The singularity is to nerds what the rapture is to fundamentalist protestant wackjobs....

I'll do my best not to get left behind. Praise the singularity!

[Luna]

I would have liked it more if he did understand the brain and what he was saying was true :/

[Elus]

They called him crazy when he predicted the internet and that a computer would beat a man in Chess.

I hope he will have a hearty chuckle at all the skeptics if we do achieve full brain simulation in 10 years.

[JLL]

It seems to me the writer of this article doesn't understand what Ray is talking about.

Like someone said in the comments, the fact is that the brain IS coded by the genome, and the genome only. The "million lines of code" is not too far from the truth. That the developing brain interacts with the environment does not change this.

It's like an AVI movie; all the information about the movie is there, and it's watchable once you open it with the right software. If you write the very same characters by hand to create an exact copy of the AVI file and open it in the same software, the movie will be identical.

Obviously, this does not mean that you can change random stuff in the file and expect the main character to grow a moustache. But I don't think this was what Kurzweil was after.

Besides, if you change random characters in an AVI file, chances are it won't open. But if you change random genetic information about the brain, chances are you'll get strange results, but some results nonetheless. I think it would be a pretty interesting way of reverse engineering some of the parts of the brain.

[e Volution]

I think this has already reached a level of popularity that Ray will personally address this critique of his ideas. I hope he does, as it is an interesting idea that brain simulation is stuck at the protein interactions and complexity. I think to claim he does not understand the brain is a ridiculous statement however; he is writing a book called how the mind works and how to build one. This PZ Myers is focusing on the pennies, while Ray is concerned about the dollars. That or he is just being sensational, a very real possibility as I am sure it brought him some order of magnitude greater spikes in traffic (and maybe add revenue).

[Kolos]

They called him crazy when he predicted the internet and that a computer would beat a man in Chess.

I hope he will have a hearty chuckle at all the skeptics if we do achieve full brain simulation in 10 years.

Internet was there since a long time in one form or another, it wasn't that hard to predict it. I remember internet already played a very important role in Orson Scott Card book the Enders Game from 1985.
Anyway he might be more or less right about the fact but less about the reasons behind it like with his Soviet Union collapse prediction.

[e Volution]

Just as I predicted Ray Kurzweil Responds to “Ray Kurzweil does not understand the brain”

Ray Kurzweil Responds to “Ray Kurzweil does not understand the brain”
August 20, 2010 by Ray Kurzweil

While most of PZ Myers’ comments (in his blog post entitled “Ray Kurzweil does not understand the brain” posted on Pharyngula on August 17, 2010) do not deserve a response, I do want to set the record straight, as he completely mischaracterizes my thesis.

For starters, I said that we would be able to reverse-engineer the brain sufficiently to understand its basic principles of operation within two decades, not one decade, as Myers reports.

Myers, who apparently based his second-hand comments on erroneous press reports (he wasn’t at my talk), goes on to claim that my thesis is that we will reverse-engineer the brain from the genome. This is not at all what I said in my presentation to the Singularity Summit. I explicitly said that our quest to understand the principles of operation of the brain is based on many types of studies — from detailed molecular studies of individual neurons, to scans of neural connection patterns, to studies of the function of neural clusters, and many other approaches. I did not present studying the genome as even part of the strategy for reverse-engineering the brain.

I mentioned the genome in a completely different context. I presented a number of arguments as to why the design of the brain is not as complex as some theorists have advocated. This is to respond to the notion that it would require trillions of lines of code to create a comparable system. The argument from the amount of information in the genome is one of several such arguments. It is not a proposed strategy for accomplishing reverse-engineering. It is an argument from information theory, which Myers obviously does not understand.

The amount of information in the genome (after lossless compression, which is feasible because of the massive redundancy in the genome) is about 50 million bytes (down from 800 million bytes in the uncompressed genome). It is true that the information in the genome goes through a complex route to create a brain, but the information in the genome constrains the amount of information in the brain prior to the brain’s interaction with its environment.

It is true that the brain gains a great deal of information by interacting with its environment – it is an adaptive learning system. But we should not confuse the information that is learned with the innate design of the brain. The question we are trying to address is: what is the complexity of this system (that we call the brain) that makes it capable of self-organizing and learning from its environment? The original source of that design is the genome (plus a small amount of information from the epigenetic machinery), so we can gain an estimate of the amount of information in this way.

But we can take a much more direct route to understanding the amount of information in the brain’s innate design, which I also discussed: to look at the brain itself. There, we also see massive redundancy. Yes there are trillions of connections, but they follow massively repeated patterns.

For example, the cerebellum (which has been modeled, simulated and tested) — the region responsible for part of our skill formation, like catching a fly ball — contains a module of four types of neurons. That module is repeated about ten billion times. The cortex, a region that only mammals have and that is responsible for our ability to think symbolically and in hierarchies of ideas, also has massive redundancy. It has a basic pattern-recognition module that is considerably more complex than the repeated module in the cerebellum, but that cortex module is repeated about a billion times. There is also information in the interconnections, but there is massive redundancy in the connection pattern as well.

Yes, the system learns and adapts to its environment. We have sufficiently high-resolution in-vivo brain scanners now that we can see how our brain creates our thoughts and see our thoughts create our brain. This type of plasticity or learning is an essential part of the paradigm and a capability of the brain’s design. The question is: how complex is the design of the system (the brain) that is capable of this level of self-organization in response to a complex environment?

To summarize, my discussion of the genome was one of several arguments for the information content of the brain prior to learning and adaptation, not a proposed method for reverse-engineering.

The goal of reverse-engineering the brain is the same as for any other biological or nonbiological system – to understand its principles of operation. We can then implement these methods using other substrates other than a biochemical system that sends messages at speeds that are a million times slower than contemporary electronics. The goal of engineering is to leverage and focus the powers of principles of operation that are understood, just as we have leveraged the power of Bernoulli’s principle to create the entire world of aviation.

As for the time frame, some of my critics claim that I underestimate the complexity of the problem. I have studied these issues for over four decades, so I believe I have a good appreciation for the level of challenge. What I would say is that my critics underestimate the power of the exponential growth of information technology.

Halfway through the genome project, the project’s original critics were still going strong, pointing out that we were halfway through the 15 year project and only 1 percent of the genome had been identified. The project was declared a failure by many skeptics at this point. But the project had been doubling in price-performance and capacity every year, and at one percent it was only seven doublings (at one year per doubling) away from completion. It was indeed completed seven years later. Similarly, my projection of a worldwide communication network tying together tens and ultimately hundreds of millions of people, emerging in the mid to late 1990s, was scoffed at in the 1980s, when the entire U.S. Defense Budget could only tie together a few thousand scientists with the ARPANET. But it happened as I predicted, and again this resulted from the power of exponential growth.

Linear thinking about the future is hardwired into our brains. Linear predictions of the future were quite sufficient when our brains were evolving. At that time, our most pressing problem was figuring out where that animal running after us was going to be in 20 seconds. Linear projections worked quite well thousands of years ago and became hardwired. But exponential growth is the reality of information technology.

We’ve seen smooth exponential growth in the price-performance and capacity of computing devices since the 1890 U.S. census, in the capacity of wireless data networks for over 100 years, and in biological technologies since before the genome project. There are dozens of other examples. This exponential progress applies to every aspect of the effort to reverse-engineer the brain.

[e Volution]

And the response from PZ Myers Kurzweil still doesn't understand the brain

Kurzweil still doesn't understand the brain
Posted on: August 21, 2010 11:08 AM, by PZ Myers

Ray Kurzweil has responded to my criticisim of futurist fortune-telling. It really just compounds the problems, though, and gullible people who love Ray will think he's answered me, while skeptical people who see through his hocus-pocus will be unimpressed. It's kind of pointless to reply again, but here goes.

His first point is silly.

For starters, I said that we would be able to reverse-engineer the brain sufficiently to understand its basic principles of operation within two decades, not one decade, as Myers reports.

I don't care.

I didn't make an issue of his timescale in the first place; in fact, I said it made no difference. The problem is that he has provided no reason to specify a date, other than his vague mantra of "exponential growth". Why not say 5 years? Why not 50? The heart of the Kurzweil method is to simply pick a date far enough in the future that we cannot predict what technological advances will occur, and also far enough forward that he isn't likely to be confronted with his failure by people who remember what he said, and all is good. My complaint isn't that he has set a date by which we'll understand the brain, but that he has provided no baseline value for his exponential growth claim, and has no way to measure how much we know now, how much we need to know, and how rapidly we will acquire that knowledge. "Really fast" or "exponentially increasing" are not informative.

I mentioned the genome in a completely different context. I presented a number of arguments as to why the design of the brain is not as complex as some theorists have advocated. This is to respond to the notion that it would require trillions of lines of code to create a comparable system. The argument from the amount of information in the genome is one of several such arguments. It is not a proposed strategy for accomplishing reverse-engineering. It is an argument from information theory, which Myers obviously does not understand.

I think I understand it better than Kurzweil. If we have a seed of information that initiates a process, followed by many activities and interactions that add progressively more information to the process, you can't use information theory to measure the amount of information in the seed and then announce that you've put an upper bound on the amount of complexity in the process.

For instance, you can't measure the number of transistors in an Intel CPU and then announce, "A-ha! We now understand what a small amount of information is actually required to create all those operating systems and computer games and Microsoft Word, and it is much, much smaller than everyone is assuming." Put it in those terms, and the Kurzweil fanboys would laugh at him; put it in terms of something they don't understand at all, like the development and function of the brain, and they're willing to go along with the pretense that the genome tells us that the whole organism is simpler than they thought.

I presume they understand that if you program a perfect Intel emulator, you don't suddenly get Halo: Reach for free, as an emergent property of the system. You can buy the code and add it to the system, sure, but in this case, we can't run down to GameStop and buy a DVD with the human OS in it and install it on our artificial brain. You're going to have to do the hard work of figuring out how that works and reverse engineering it, as well. And understanding how the processor works is necessary to do that, but not sufficient.

Kurzweil does add another piece to his argument, although it doesn't help: the modularity and repetitive organization of the human brain.

For example, the cerebellum (which has been modeled, simulated and tested) -- the region responsible for part of our skill formation, like catching a fly ball -- contains a module of four types of neurons. That module is repeated about ten billion times. The cortex, a region that only mammals have and that is responsible for our ability to think symbolically and in hierarchies of ideas, also has massive redundancy. It has a basic pattern-recognition module that is considerably more complex than the repeated module in the cerebellum, but that cortex module is repeated about a billion times. There is also information in the interconnections, but there is massive redundancy in the connection pattern as well.

This is true — the cortex is a layered structure with similar elements repeated over and over again, in broad arrays. Pyramidal neurons, for instance, are instantly recognizable and and share a whole suite of common morphological elements between each other — but each one is also as unique as a snowflake. Those differences matter, and they are not specified in the genome. (For that matter, you won't find any blueprint in the genome for the dendrite pattern of pyramidal neurons, either). If you want to recreate a generic human brain, it won't work if you just make every pyramidal neuron exactly identical; there have to be spatial differences and differences in connectivity. You especially won't be able to carry out something far more specific, such as emulate Ray Kurzweil's brain, if you decide to simplify and make his cortex a uniform array of identical modules.

In short, here's Kurzweil's claim: the brain is simpler than we think, and thanks to the accelerating rate of technological change, we will understand it's basic principles of operation completely within a few decades. My counterargument, which he hasn't addressed at all, is that 1) his argument for that simplicity is deeply flawed and irrelevant, 2) he has made no quantifiable argument about how much we know about the brain right now, and I argue that we've only scratched the surface in the last several decades of research, 3) "exponential" is not a magic word that solves all problems (if I put a penny in the bank today, it does not mean I will have a million dollars in my retirement fund in 20 years), and 4) Kurzweil has provided no explanation for how we'll be 'reverse engineering' the human brain. He's now at least clearly stating that decoding the genome does not generate the necessary information — it's just an argument that the brain isn't as complex as we thought, which I've already said is bogus — but left dangling is the question of methodology. I suggest that we need to have a combined strategy of digging into the brain from the perspectives of physiology, molecular biology, genetics, and development, and in all of those fields I see a long hard slog ahead. I also don't see that noisemakers like Kurzweil, who know nothing of those fields, will be making any contribution at all.

So what exactly is the basis of Kurzweil's expected magic great leap forward? And no, the miracle of exponential growth is not an answer. If all a futurist has to do is wave his hands and say things will change more rapidly than we expect, then futurists like Kurzweil are nothing but techno-gimmicky Criswells. Utterly useless.

[Luna]

I think it was interesting until it became a insult/mock-the-other ego argument.

If a message isn't worth replying,. don't reply. If you're replying, don't say it's not worth it.. for starters. Goes for both :/

[eternaltraveler]

Like someone said in the comments, the fact is that the brain IS coded by the genome, and the genome only. The "million lines of code" is not too far from the truth. That the developing brain interacts with the environment does not change this.

sure. then you only need a few hundred trillion (or other suitable absurdly large number) lines of code to code for the "environment" (or a robust enough model of physics, and a matrioshka brain to figure out the interactions between the few types of fundamental particles with a few properties).

Ultimately all the complexity of the universe is based on maybe 30 elementary particles (or so) interacting in unbelievably complex ways. The genome encodes for 22,000ish protiens, also interacting in unbelievably complex ways.

Could happen I suppose. I wouldn't count on it being good for the meat beings currently living here.

Edited by eternaltraveler, 23 August 2010 - 02:48 AM.

[okok]

Now PZ Myers't tone in his rebuttal discredited him a bit, notwithstanding his argument vis à vis Kurzweil's exponential growth may have some validity: iirc, he himself (Kurzweil) admitted that bio sciences aren't subject to Moore's Law. And if they do see exponential growth, it's more their quantitative than qualitative aspects. Also the size of the exponent makes a huge difference.
Myers is right in that a bottom-up mathematical approach is too complex, but i give Kurzweil the credit to take into account empirical growth data and recognizing that once we have extracted the main working principles we can discard evolutionary complexity and see emerging properties when combining the best of both worlds.

Edited by okok, 24 August 2010 - 02:10 AM.

[Luna]

Now PZ Myers't tone in his rebuttal discredited him a bit, notwithstanding his argument vis à vis Kurzweil's exponential growth may have some validity: iirc, he himself (Kurzweil) admitted that bio sciences aren't subject to Moore's Law. And if they do see exponential growth, it's more their quantitative than qualitative aspects. Also the size of the exponent makes a huge difference.
Myers is right in that a bottom-up mathematical approach is too complex, but i give Kurzweil the credit to take into account empirical growth data and recognizing that once we have extracted the main working principles we can discard evolutionary complexity and see emerging properties when combining the best of both worlds.

Kurzweil tone wasn't that amazing at the beginning either.

[Reno]

Myers was looking for Legitimation, and Kurswell gave it to him. It would have been better to not give Myers the time of day.

Personally, I think biotechnology and computer technology are right at the point where they are beginning to merge into one field. There are already biological testing chips running hundreds of experiments at one time. There are supercomputers designing unique lab experiments. The closer the two fields merge the more one can expect Moore's law to start to apply to biotechnology. Every 18 months we can expect a double in performance. How long will it take for today's lab novelties, growing in performance exponentially, to become tomorrow's brain designers?

I will say Myers might be right about Kurswell's timeline. I think that exponential growth requires a constant incentive in improvement. AI may bring wonders and shit blunders, but... if the public is satisfied with the status quo then the pace of change will be surprisingly slow. The Ipod had to be created before there could be a huge demand for it.


“You have to crawl before you walk, walk before you run."

Edited by Reno, 24 August 2010 - 06:35 AM.

[platypus]

I hope he will have a hearty chuckle at all the skeptics if we do achieve full brain simulation in 10 years.

I don't think the necessary computing-power will be there. BTW, what came out of the mouse-brain simulations etc.? At least not Skynet..

[Kolos]

For starters, I said that we would be able to reverse-engineer the brain sufficiently to understand its basic principles of operation within two decades, not one decade, as Myers reports.

I don't care.

Too bad because it's actually very important, you just proved that Kurzweil was right when he said that you don't really know what he was talking about Mr Myers.
Kurzweil might be more or less right about the time frame but I don't think that he really understands how will this happen, anyway 20 years is certainly closer than 300 as someone proposed in the comments, brain can't be too complicated or it wouldn't be practical.

[Reno]

I don't think the necessary computing-power will be there. BTW, what came out of the mouse-brain simulations etc.? At least not Skynet..

You mean the top secret network of UK military satelites?

Edited by Reno, 25 August 2010 - 04:44 PM.

[cyborgdreamer]

I doubt you could simulate the brain starting with just DNA and basic transcription mechanisms. At minimum, you'd need to simulate the entire zygote. For example, chemical gradients within the original egg determine which end of the embryo becomes the front and which end becomes the back. This ultimately determines the relative orientation of parts of the brain. You'd also need to simulate the nutrient rich growth medium of the womb and that will only get you a simulation of a newborn's brain (and body). You could probably avoid simulating the entire world if you transfer the information to a robotic body and find loving parents to raise it. Of course, that robot might end up with serious emotional issues and might not consent any further experimentation.

A better idea would be to find the simplest way possible to simulate a neuron and then develop a high-resolution scanner to determine the positions and types of all the neurons in a specific brain. How long will it take to develop that kind of technology? I can't begin to guess.

[Mind]

Remind me to not get into any rational discussions with Myers. What is up with the ad hominems?

[Reno]

Remind me to not get into any rational discussions with Myers. What is up with the ad hominems?

Everyone wants 15 minutes of fame. In order to get that attention he has no qualms about trashing a man like Kurzweil. People who step out on a limb to make predictions are easy pickens.

Edited by Reno, 03 September 2010 - 02:46 AM.

[Cameron]

AI may bring wonders and shit blunders, but... if the public is satisfied with the status quo then the pace of change will be surprisingly slow. The Ipod had to be created before there could be a huge demand for it.

The problem is that all evidence suggest the market seems poised to allow one or two decades more of Moore's law. Enough to bring the required hardware costs down low enough for most any individual to easily acquire. Unlike an ipod or tablet, AI is the ultimate technology it abolishes the need for consumers. In fact, it basically abolishes the need for money, even for society's services themselves as sufficient mastery of the science of mind design allows one to obtain not only everything society can provide but things that far dwarf what any near future society without AI can provide.

[Kolos]

AI is the ultimate technology

As a transhumanist I wouldn't put so much hope in AI, it really sounds kinda religious... Augmenting mental capabilities of people that already are intelligent should be much easier than creating it from scratch and when we finally have the knowledge and technology to think about creating this super AI it might be not so attractive anymore, especially if we master mind uploading by that time and barier between man and computer will be rather symbolic. If that will be the case we will need some AI but for things we see as boring, they might be computer bots, NPCs or robots-soldiers, miners etc. so they probably wouldn't be too intelligent.

[Cameron]

AI is the ultimate technology

As a transhumanist I wouldn't put so much hope in AI, it really sounds kinda religious... Augmenting mental capabilities of people that already are intelligent should be much easier than creating it from scratch and when we finally have the knowledge and technology to think about creating this super AI it might be not so attractive anymore, especially if we master mind uploading by that time and barier between man and computer will be rather symbolic. If that will be the case we will need some AI but for things we see as boring, they might be computer bots, NPCs or robots-soldiers, miners etc. so they probably wouldn't be too intelligent.

The idea is obtaining free labor. Particularly the automation of intellectual labor. A super AI would be good but it is not necessary, nor required to eliminate most forms of labor. An advanced neural interface is likely to require years of work from countless scientists and engineers probably. Same thing for advanced tweaking of the brain or perfecting uploads. Advanced molecular machinery and the elimination of aging again seem to require years of work from countless scientists.

If you can successfully automate intellectual labor, you can get more than the equivalent of countless years of work by 100s of experts in any field, at virtually zero cost and in a fraction of the time. Even if one were to significantly increase their intellect that does not suddenly make it that desirable to perform the work of 100s without the aid of automation.... I doubt any sane individual would want to do the volume of work of 100s or 1000s of individuals. Not that this high level intellectual work is particularly difficult, it's the tediousness of it all. AI systems could be used to outproduce in quality and quantity, say the entire software industry... now imagine if instead of AIs doing everything, you were the one designing every feature, writing every line of code and checking for errors. But that work has to be done, it's like building a pyramid, you want to be the project lead not the one carrying each and every single stone, even if you could and were immortal, such grunt work is better performed by someone else.

This is particularly interesting when it comes to content generation. Suppose someone wants private content production equivalent to that of the current human civilization. Who's going to produce that for their private personal use? Obviously they're not going to be personally writing the books, doing the paintings, doing the cg, making the games, etc. Unless they start some sort of cult, and it is successful, there's not going to be sufficient willing individuals to do this for free, either.

Edited by Cameron, 14 September 2010 - 06:29 AM.