Slashdot Mirror
Ray Kurzweil Does Not Understand the Brain
jamie writes "There he goes again, making up nonsense and making ridiculous claims that have no relationship to reality. Ray Kurzweil must be able to spin out a good line of bafflegab, because he seems to have the tech media convinced..."
The singularity is to nerds what the rapture is to fundamentalist protestant wackjobs....
More time spent on talking bullshit, telling you what you want to hear, saving face, or rewriting history than actual work. Welcome to the modern age of communication, it's new so it can't be old.
His actual comments
Read it. Other than the solid date he predicts, it's pretty plausable.
1. Technology is growing exponentially
2. The brain isn't some magical soul-endowed jesus box. It's a function of physics
Here's how that math works, Kurzweil explains: The design of the brain is in the genome. The human genome has three billion base pairs or six billion bits, which is about 800 million bytes before compression, he says. Eliminating redundancies and applying loss-less compression, that information can be compressed into about 50 million bytes, according to Kurzweil.
About half of that is the brain, which comes down to 25 million bytes, or a million lines of code.
What's so crazy about that?
Would be nice if the summary even hinted at what the ridiculous claim actually WAS...
Namely, that we'll be able to reverse engineer the human brain in the next 10 years.
It just gives me headaches
If the complain here is about how media laps up these 'pundits', well, welcome to new age media. David Pogue is very popular too. But some of the nasty stuff he writes which are factually wrong goes completely un-noticed.
...he must be right! He used math, and everything! I'm a little shocked that Kurzweil equates blueprints with the functioning organ. I am not shocked, however, that the tech media latched onto this--at first blush it sounds so *reasonable.*
Kurtzweil is just the latest in a long line. How do you think publications like Fast Company and Wired get written?
And does anybody remember JonKatz?
What is brain?
This is a new low for /.! Who the hell is Ray, and what is the claim?
His latest claim is that we'll be able to reverse engineer the human brain within a decade
Amateur. I could put something together to simulate the human brain in about 8 months.
(Plus another 3 minutes at the start)
Summation 2
FTFA: The end result is a brain that is much, much more than simply the sum of the nucleotides that encode a few thousand proteins.
Likewise, the end result of a computer is much, much more than simply the sum of the commands that encode a CPUs instruction set.
"Reverse-engineering the brain is being pursued in different ways," says Kurzweil. "The objective is not necessarily to build a grand simulation - the real objective is to understand the principle of operation of the brain."
You could just read the article, and understand that this isn't designed to create a human brain from a digital base, it's to understand the operation of an already developed brain.
Sejnowski says he agrees with Kurzweil's assessment that about a million lines of code may be enough to simulate the human brain.
Here's how that math works, Kurzweil explains: The design of the brain is in the genome. The human genome has three billion base pairs or six billion bits, which is about 800 million bytes before compression, he says. Eliminating redundancies and applying loss-less compression, that information can be compressed into about 50 million bytes, according to Kurzweil.
About half of that is the brain, which comes down to 25 million bytes, or a million lines of code.
Idiot. The design of the brain is encoded in the genome in the same way that the design of a 4KiB program is encoded in its load module: useful for running the program on its original hardware.
But then you have architectural issues. That 4KiB of information does not run unless it's supported by a complex operating system, which itself is supported by complex logic in an ISA and memory managment architecture backing it up. And all that is implemented on a specific design in a specific physics model.
Translating that program to SPARC takes work, and it comes out roughly the same size. Translating that program to a progression of chemical reactions produces something vastly different, especially since you need a new middle ware (chemical environment) running on top of different physics (chemistry).
Translating a physical architectural design from chemistry to computer logic on top a given ISA is the same problem. You now have odd issues that are messy, and then the program running on the brain needs to be built again. That program is even more complex and less known.
Support my political activism on Patreon.
I'm not even sure what to say about this statement
I agree that this statement is wrong. The genome means less than ever. The idea that sequencing a genome is the endpoint of understanding has less power than ever.
These "genomes" don't really exist in nature. In nature sex cells come together and are "reprogrammed" by methylation and other mechanisms. Each cell line is reconfigured and changes throughout cell development and interaction with its environment.
The action is in functional genomics and epigenetics. You wont find it in the mainline code of some dead cell.
The design of the brain may bootstrap from the genome, but the design of brain is in the runtime environment, which could be infinitively more complicated than the genome.
Junk DNA anyone?
I wish I could get a job as a futurist....think about it:
"What do you think is going to happen in the future???"
"Um...dogs will bring soda to you when you whistle a Cradle of Filth song?"
"OMFG THATZ BRILLIANT. HERE R MONIES, PLZ HAS MAH BABBIES!"
Living With a Nerd
Ray Kurzweil has been making claims for AI for years. For example that we will have an AI singularity event and that
society will be completely replaced my machines. Well, decades later it still hasn't happened and the only things in the
field of computer science that seems to have a life of its own are spam and computer viruses. I'd like call them a
life form.
Will we reverse engineer the brain any time soon? I doubt it. Part of the reason is practical. This would be an
extremely expensive and time consuming undertaking. I'm not sure its even worth it, especially when this is
compared to other branches of science which have made rapid advances. For one example, take a look at
the field neuro-science and its use of fMRI scanning.
Reversing engineering the brain, probably is possible, but it's probably not worth it right now. Well have
to wait another decade.. again.
---- It won't be as bad as you fear or as good as you hope, but it will take twice as long as you plan.
What Kurzweil doesn't understand could fill a book.
Several in fact.
In Capitalist America, bank robs you!
Would be nice if the summary even hinted at what the ridiculous claim actually WAS... Namely, that we'll be able to reverse engineer the human brain in the next 10 years.
It's a little more complicated than that. You see, the article actually breaks down the logic behind that statement and points out how poor it is. Here's the initial part of Kurzweil's argument:
Sejnowski says he agrees with Kurzweil's assessment that about a million lines of code may be enough to simulate the human brain.
Here's how that math works, Kurzweil explains: The design of the brain is in the genome. The human genome has three billion base pairs or six billion bits, which is about 800 million bytes before compression, he says. Eliminating redundancies and applying loss-less compression, that information can be compressed into about 50 million bytes, according to Kurzweil.
About half of that is the brain, which comes down to 25 million bytes, or a million lines of code.
I have only taken high school biology but I know that the genome doesn't magically become the brain. It goes through a very complex process to amino acids which fold into proteins which in turn make cells that in turn make tissues that in turn comprise the human brain. To say we fully understand this transformation entirely is a complete and utter falsity as demonstrated by our novice understanding of the twisted beta amyloid protein that we think leads to Alzheimer's. How amino acids turn into which proteins I believe is largely an unsolved search problem that we don't understand (hence efforts like Folding@Home). And he claims that in ten years not only will we understand this process but we will ... reverse engineer it?
The man is insane. I've posted about this same biologist criticizing him before and it looks like P.Z. Myers just decided to take some extra time to point out how imprudent Kurzweil's statements are becoming. Kurzweil will show you tiny pieces of the puzzle that support his wild conclusions and leave you in the dark about the full picture and pieces that directly contradict his statements. This is a dangerous and deceptive practice that -- despite my respect for Kurzweil's work in other fields -- is rapidly turning me off to him and his 'singularity.' He's becoming more Colonel Kurtz than Computer Kurzweil.
My work here is dung.
The author is almost certainly correct that encoding the genome structure will be insufficient to produce a simulated brain, but it is also very possible that neural computation is built on a small set of principles that could be duplicated in a million lines of code. Of course, first we'll have to figure out what those principles are. It might happen in the next ten years, but considering how much time and effort (close to 50 years) has gone into understanding the low-level visual system and how little we still know about its function I wouldn't bet on it.
Let's see. On another recent article it was stated that the average car has several million lines of code running in it. I haven't come across a sentient Prius yet.
And there's that pesky parallel processing the brain does. I don't think that a rack full of Nvidia Tesla cards can approach the average two year old's parallel processing capability.
I agree, Kurzweil is smoking something and not sharing.
Ray Kurzweil never says we'll be able to write software to emulate the brain completely in ten years. He just says we'll have reverse engineered the signals we have to send to the cortex to interact with it. Reverse engineering involves analysing and understanding something. Reengineering anew would be the next step. But then again, when you're only interested in reverse engineering the various inputs you can give to the brain to get different results, you're still very far away from understanding how to build a brain.
PZ Myers clearly doesn't understand reverse engineering and writes up a useless article based on his erroneous comprehension of Ray Kurzweil's prediction.
Sejnowski says he agrees with Kurzweil's assessment that about a million lines of code may be enough to simulate the human brain.
Kurzweil explains: The design of the brain is in the genome. The human genome has three billion base pairs or six billion bits, which is about 800 million bytes before compression, he says.Eliminating redundancies and applying loss-less compression, that information can be compressed into about 50 million bytes, according to Kurzweil.
Dude, the equations of quantum mechanics can be written on one page. General Relativity can be written on a second page. What more do you need ? Clearly, a few hundred lines of code (and a few do loops) should be enough to simulate the entire universe, brains and all.
Glad we cleared that up. All you physicists and astronomers can go home now and work on your resumes.
To simplify it so a computer science guy can get it, Kurzweil has everything completely wrong.
Best summary I've ever read. (Though a bit more respect for our ability to get it would have been nice.)
"To simplify it so a computer science guy can get it, Kurzweil has everything completely wrong" -- Too strong a statement. I've read some of Kurzweil's books, and while they may be a little bit on the optimistic side, they are quite imaginative and thought provoking. Considering that no one as of yet understands the brain, or has more than an inkling of an idea how mind and consciousness emerge, it is a little dramatic to proclaim his having 'everything wrong.' That there is a gigantic amount of redundancy in the design of the brain is true, and interesting, and a valid place to start when considering how one might go about inventing one, real or virtual. We'll see in ten years, now won't we?
but I've always had the impression that his 'philosophy' is inspired by Star Trek rather than the scientific method or logic. His AI claims always seem to be a vast underestimation of the complexity of the human mind. "Wishful thinking" is the best phrase to describe his ideas. Regardless, the summary for the article could have been an actual summary rather than just a copy-paste of the first couple lines.
"From the depths of my skeptical and rationalist soul, I ask the Lord to protect me from California touchie-feeliedom."
Kurzweil may be off on the AI side but biology may someday create a singularity in recreating intelligence.
Gene Networks are now being used to model cancer and cell communications are critical to cancer growth.
Dr. Judah Folkman pioneered this work in the 80’s.
http://www.amazon.com/Dr-Folkmans-War-Angiogenesis-Struggle/dp/0375502440
Craig Venter used ocean genomes to show viral lateral transfers evolve DNA.
http://scratch.mit.edu/projects/GeneMachine/51835
http://scratch.mit.edu/projects/GeneMachine/50308
A network analyzer may be able to model gene protocols used by cancer.
Cancer reverts to the primitive LTR messaging of viruses in junk DNA.
http://www.sciencedaily.com/releases/2010/05/100502173845.htm
Network protocol analyzer – Wire shark
http://www.wireshark.org/
Using the Cancer Databases may allow a Wireshark for Cancer to be created.
Database for Personalized Cancer Treatment - Sanger
http://www.sciencedaily.com/releases/2010/07/100715152901.htm
Each cancer drug would create a map of gene expression and gene network protocols.
Using comparative genomics with other organisms a multi species gene map may be possible.
Using Goalie a network model will allow scientists to understand gene expression and cell communications.
Goalie = Wireshark (ethereal) for Cells – Future Cancer Diagnostic –
Gene cluster analysis
http://bioinformatics.nyu.edu/~marcoxa/work/GOALIE/
Students could then use http://processing.org to build gene models of complex cancer systems.
The real future is not AI it is in reducing health care costs through molecular biology.
In understanding the biology we may someday have a clue of how the brain works and how AI works.
AI will require a context switching machine similar to DNA. All present silicon solutions require a programmer and if statements.
Viruses build genomes with no programming but with shareware.
Each drop of seawater has a million viruses and 1 thousand bacteria.
The ocean is a genetic computer.
http://scratch.mit.edu/projects/GeneMachine/51835
By building systems biology and gene networks todays students may one day understand cross species protocols and common gene languages.
Genomic Test of Tumor DNA
http://www.genomichealth.com/OncotypeDX/Index.aspx
Excellent Breast Cancer Tutorial for students
http://gcat.davidson.edu/Pirelli/index.htm
PZ Myers threw a red herring there. What Kurzweil says is pretty reasonable, he used the total amount of information in the genome to get an upper limit estimate of the amount of library code needed to simulate a brain. I say "library" to differentiate from data, since a lot of our brain information comes from our experiences, i.e. library == instincts.
Myers goes off in a tangent about biochemistry which has nothing to do with the argument. I've never read anything hinting that the way to simulate a human brain would be to simulate how the molecules in the brain behave. We don't build airplanes with flapping wings either, machines can emulate the functionality of a living being without need to simulate the exact details.
From the number on neurons in the human brain, considering how many interconnections there are and how fast the neurons can fire, I think a machine with one million processing cores at 1 GHz would have approximately the same data handling capacity as a human brain. The rest is software. Neural network software is pretty much routine stuff, the tricky part is learning what are the interconnections between the neurons.
If you read the article closely, the engineering is designed to understand operation, not create a grand simulation. You can all quit toting your brainless responses now.
...Kurzweil's assessment that about a million lines of code may be enough to simulate the human brain.
If you accept that, then the real problem to solve becomes: in what language do you write the code?
;-)
I live ze unknown. I love ze unknown. I am ze unknown.
...are slashdot editors. There isn't even a pretense of a story here. I've seen more informative Twitter summaries.
To simplify it so a computer science ignorant biologist with a tendency to inane rants can possibly get it, you don't need to simulate electrons in a semi-conductive material at specific temperatures in order to build a complete working emulator for an old computer.
"I love my job, but I hate talking to people like you" (Freddie Mercury)
Eliminating redundancies and applying loss-less compression, that information can be compressed into about 50 million bytes
This is so retarded that it's sad. Why is it so hard to understand how compression algorithms works? Saying that X can be compressed into Y bytes doesn't say ANYTHING. You can "loose-less compress" ANYTHING into 1 bit by using the function that takes that and returns the bit "1" (and which takes anything else and returns "0" + that). What does compression has to do with anything? The stupid hurts...
...except, maybe, Pinky.
Guaranteed! This comment 100% Anthrax free!
No not really.
A computer is a fixed system. If you tell it to do A (via software), you know you will get B
You clearly don't work where I do...
AC... for a reason!!!
Given that we only use 10% of our brains (less, for women), I would think we could compress the code even further! We'd probably want to optimize the think_about_sex subroutine too.
What do you mean "infinite"? The human brain is composed of one hundred billion or so neurons. Looks like it's pretty much finite to me. I have ten times as many bytes of information in my hard disk.
The human brain is far more complex than all of the technology created by man put together.
"If the brain were simple enough to be understood, it would be too simple to understand itself"--anonymous
If all of our folly were turned to intelligence and divided amongst a thousand toads, each would be more intelligent than Aristotle
"Gentlemen, you can't fight in here! This is the War Room!" -- Dr. Strangelove
Scanning down the comment list, it looks like every (+2 or more) comment has read the article and is quoting from it -- what has happened to the slashdot I knew and loved?
I mod down anyone who says "I will be modded down for this", regardless of the rest of their comment
But he understands the psychology behind selling vaporware.
There are no loopholes. It's either legal or it's not.
How do we calculate the complexity of the human brain? It's not by counting neurons (about 100 billion) or synapses (the connections between neurons, about 100 trillion). Individual receptors play a role in information processing (there are no good estimates of how many receptors there are in the brain, but it's at least an order of magnitude greater than the number of synapses). There is now some evidence that receptors use quantum effects to perform information processing. That means that in order to duplicate the brain we may need a quantum computer with quadrillions of qubits. That's not going to happen in the next 10 years.
The human brain could be simulated with a million lines of code, however it would need to be run inside an emulator that simulates physics perfectly and has enough RAM to store the quantum state of the aprox 1.5*10^26 atoms that make up a human brain.
Amateur. I could put something together to simulate the human brain in about 8 months.
It's easy if you limit yourself to simulating Englishmen named Arthur Dent, actually. You just need to have it return the following to queries:
"What?"
"I don't understand!"
"Where's the tea?"
Piece of cake.
Large systems do not require lots of code. Complex systems do. The brain is made up of neurons that are largely identical, arranged in a number of patterns. We have all the basic building blocks to make a brain already. We just don't know how to put it together.
Amount of DNA is not a good way to estimate the number of lines of code needed. DNA is not an efficient encoding. It doesn't need to be. There's millions of years of legacy crap in there for lungfish compatibility and stuff. It doesn't cost anything to keep around and doesn't need to be maintained so it remains in there.
The summaries tend to be nothing more than the submitter taking the most polarizing sentence/paragraph from TFA and pasting it into the summary field
Doing more than that takes time and if you are to do it right, someone else has submitted the article with the shit summary.
I have given up submitting articles.
RIP America
July 4, 1776 - September 11, 2001
I'm not sure what it is about his claims that are supposed to be so ludicrous. For example, a million lines of code seems at least plausible, as long as we bear in mind the following:
1. We're not trying to mimic the brain at the protein level, rather at the broader, inter-neuron level (and whatever complex intra neuron behaviour we discover).
2. The million lines of code don't need to encompass the capacity of the brain, just its general neural architecture and adaption rules - there will no doubt be many gigabytes (terabytes?) of working memory, which would actually store the neural connections and whatever parameters they may have.
To be honest, the authors of this article seem to be rather too cocksure in dismissing all this. Even the apparent agreement of Terry Sejnowski (co-inventor of the boltzmann machine) doesn't give them pause. I'm not that familiar with Kurzweil's predictions, but this seems fairly reasonable to me.
There is a google tech talk by Geoff Hinton on restricted boltzmann machines, (a sort of stochastic neural network) that's well worth a watch, for those that are interested. They are considered biologically plausible, and he seems mostly to apply them to machine vision tasks.
Ray's documentary about the Singularity has been touring the national film festivals along with Ray. I saw it June. Its begins as a dull-talking heads piece about the current state of A.I. Many of the Big Name A.I. Scientists are interviewed. Then it transitions into a crime-drama story about the legal rights of A.I.s. That part was more interesting, since it had a story. The film is full of special effects to advance the story. Although I know most of the film to be factual, I suspect it will look like a scifi movie to the average audience member. I think Ray is seeking looking at broader distribution on cable television or arts theaters after the festival run.
Ray was interesting in person during a film-makers Q&A. He reminded me of Woody Allen, but more confident and intelligent. He was graduated from M.I.T. about decade before myself. I personally believe in the Singularity, but more likely in centuries rather than decades.
Glad someone called him on this. I'm a software guy but I've understood the intractible complexity of biological development with genes expressed over time and insanely complex interdependent interractions. We will be able to measure a brain before we can build build one I think, then simulate whet we know about neural communication and interconnection at the extracelular level in that measured and simulated representation. It is just idiotic for anyone to suggest we should approach the problem by growing a brain from protein sequences.
unless you buy a few cases of these longevity shakes.
Bitches.
The brain as defined by the DNA is similar to technical blueprints for a hard drive. Even if you manage to build one, you still don't have an operating system or data on it, just the writing tools and surface to host them.
There are currently four academic disciplines working on the reverse engineering of a human mind. Linguistics, psychology, computer science, and philosophy. You can count neurology too if you want to start talking about the *actual* brain. Several tens of thousands of individuals are directly and indirectly working on this problem. We've come a long way in the last few decades. Unfortunately, we have a pretty long way to go. For the moment we lack a model which accurately describes how mental processes work. There isn't even a consensus on how the processing is done.
"modeling the brain" is not even really the hard part. One only needs sufficient computing power to model what they *think* is going on logically (there isn't even a consensus here). The trick is modeling the mind. We are very, very far away from that.
A fun number to throw around is how many synaptic connections are present in the brain. Synaptic connections are widely believed to be the best indicator of overall memory storage and processing speed (to an extent). There are about 10 to the 15th (Peta I believe?) synaptic connection in a normal human brain. A significant number of these are active at any given time. In other words, the brain is performing a HUGE number of "calculations" simultaneously at all times. Modeling just the hardware is obviously not easy... modeling the software is currently not possible. I doubt it will be in the next 50 years.
For a good read on what many cognitive scientists think is going on, though it is clearly not an accurate model but rather a best guess, go read up on "connectionism".
I imagine that is you try to write a program that simulates a sufficiently simple brain, it might be possible. A perfect example of such a brain is his brain of course!
... is a car analogy:
Ray Kurzweil does not understand how to parallel park.
__ Someday, but not this morning, I'll finally learn to use the preview button.
I wonder if Kurzweil is being misunderstood here. Would not most of you agree that an entire person is described by the genetic data contained withing one stem cell? Or at least within one sperm cell and one egg cell? Then why is it difficult to believe that the blueprint could be easily describable and compressed in a few million bytes? And if its possible to do this, then why would it not be possible to create some kind of molecular emulator that can emulate reconstructing a living thing from this blueprint?
I can imagine a computer program that emulates molecular interactions. I can imagine that this program that has a DNA blueprint for an entire person. Basically then, give the program the correct inputs (food, oxygen, water) and necessary outputs (waste), and see what the computer program produces. Now speed the whole process up because you're doing it in software.
You may not get an adult human, but you might get a 9-month old fetus. I just don't know.
If the above is possible, then it becomes possible to play around and experiment with a "live" human brain being emulated in software.
Why is this hard to fathom?
I know, we're talking about a monster piece of computing power to be able to do anything near that right now. But why is it beyond our understanding?
"They said I probly shouldn't fly with just one eye," "I am Bender. Please insert girder."
Kurzweil doesn't explicitly say that we'll be able to reverse engineer the human brain because it will run on only "1 million lines of code." What he probably meant to say is that we'll do it in 10 years because all the other factors that will go into it (increase in brain scanning technology, computer simulations, computing power etc.) will allow us to reach that point in 10 years. The original Gizmodo/Wired article which TFA article points to includes Kurzweil's claim that we "only need 25 MB, or a million lines of code" to simulate the human brain. While that (admittedly incorrect) element is part of Kurzweil's argument, it doesn't necessarily negate a claim that we'll still be able to build simulations of the brain in a decade.
The original Wired/Gizmodo article that PZ Myers points to is focused on the ability to emulate the software of the cortex within a supercomputer.
The key to reverse-engineering the human brain lies in decoding and simulating the cerebral cortex - the seat of cognition. The human cortex has about 22 billion neurons and 220 trillion synapses. A supercomputer capable of running a software simulation of the human brain doesn't exist yet. Researchers would require a machine with a computational capacity of at least 36.8 petaflops and a memory capacity of 3.2 petabytes - a scale that supercomputer technology isn't expected to hit for at least three years, according to IBM researcher Dharmendra Modha. Modha leads the cognitive computing project at IBM's Almaden Research Center. By next year, IBM's ‘Sequoia' supercomputer should be able to offer 20 petaflops per second peak performance, and an even more powerful machine will be likely in two to three years. "Reverse-engineering the brain is being pursued in different ways," says Kurzweil. "The objective is not necessarily to build a grand simulation - the real objective is to understand the principle of operation of the brain." Reverse engineering the human brain is within reach, agrees Terry Sejnowski, head of the computational neurobiology lab at the Salk Institute for Biological Studies. Sejnowski says he agrees with Kurzweil's assessment that about a million lines of code may be enough to simulate the human brain.
That last line is probably wildly incorrect, but it doesn't really change the basis of the argument. It's not infeasible that we could reach this in a decade. Look at this TED talk from 2009: http://www.ted.com/talks/henry_markram_supercomputing_the_brain_s_secrets.html where Henry Markram claims he's able to simulate a single neocortical column on a neuronal level in a supercomputer.
Now go back to TFA:
I'll make a prediction, too. We will not be able to plug a single unknown protein sequence into a computer and have it derive a complete description of all of its functions by 2020. Conceivably, we could replace this step with a complete, experimentally derived quantitative summary of all of the functions and interactions of every protein involved in brain development and function, but I guarantee you that won't happen either. And that's just the first step in building a simulation of the human brain derived from genomic data. It gets harder from there.
PZ Myers is probably correct about that: we won't be able to plug in an amino acid sequence into a computer and then figure out what it looks like in 3D, and how it interacts on a molecular scale. But that argument doesn
To my eternal shame, I have within arm's reach a copy of The Singularity Is Near, Copyright 2005 Ray Kurzweil. I was enthused about this book when I bought it but the claims within it are now verifiably false:-
An article by Ray Kurzweil written in 2001 says much the same thing:-
So where is this computer that can simulate a brain? Will I really have a $1000 personal brain simulator by 2020?
I too have lost admiration for Ray Kurzweil. He's an OCR expert who has pontificated with increasing ludicruousness. His law of accelerating returns may be nothing more than an aberation from peak oil.
The code is simple.
Simulate_Brain();
Now just find the compiler with the right set of libraries that can compile it. And yes, I am NOT just being anal. Half a million lines of code is MEANINGLESS. Quickly, how many lines do you need for a "Hello World" program? In assembly? C? Java? PHP?
If one day someone designs a cpu with a built in Hello World function, then it would require what? 2 instructions in assembly? Meanwhile the java guy will be pounding out yet another page of code.
MMO Quests are like orgasms:
You may solo them, I prefer them in a group.
Obviously, by your logic, a free market economy is impossible, Our economy is too complex to have evolved on its own. In fact, it is far more complex, with far more different parts, than a human being. It must have had a creator. If most any part of the economy, like the steel industry, say, were removed, the economy would not function. How did the economy function before there was a steel industry? Obviously, it couldn't, and therefore we have demonstrated irreducible specificated complexification or something.
All this free market talk is obvious bullshit, and we actually DO have a centrally planned economy because it is impossible for something so complex to have evolved without a central planner.
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
After one reads an article about the infinite complexity of the human brain, one has to wonder...
One does, does one?
Kurzweil's logic is clearly wrong, if he actually meant what Mr. Myers is saying he meant.
In the interest on honesty, I'll admit now that I think Vinge is right, and I'm not convinced Kurzweil's time-line for the singularity is wrong (or right, really).
Okay, now that we're past that, everyone who read the source article on Gizmodo please raise your hand. In The Singularity is Near Kurzweil bases his logic pretty much on algorithmically reverse engineering the brain. The Gizmodo article follows much the same path until the genomic bit at the end. Maybe Kurzweil is crazy. He often seems so to me, and I'm a booster. But I think it's a reach on the part of PZ Myers to assume that Kurzweil has abandoned his previous logic for something more "magical." The genomic stuff makes sense when one assumes that Kurzweil is trying to say, "hey, if we read out the genomic information into binary data and compressed it, it would fit into a million lines of code."
It's still dodgy computer science but a lot less of a leap than assuming he meant we will solve a huge number of biochemistry issues in a short time. Earlier in the Gizmodo write-up Mr. Kurzweil is quoted to say, "Reverse-engineering the brain is being pursued in different ways. The objective is not necessarily to build a grand simulation - the real objective is to understand the principle of operation of the brain."
It's hard to square that with Myer's interpretation. Once you allow that he might be jumping to conclusions, Myer's entire rant stops being relevant. Someone in the comments on Myer's page did correctly ding Kurzweil for assuming the emergent complexity of the human brain can be reduced to a principle of operation, but it would be equally valid to ding that commenter for assuming it can't. We just don't know. I think 10 years is optimistic. 15 to 20 seems more reasonable, but that' just gut feeling. For all we know we'll be welcoming our robot overlords in 2018.
Of course, none of the above proves that Ray Kurzweil understands the brain. Without proof, I can still confidently say he doesn't. First because the brain science people don't and second because he pretty much admits it. You don't go looking for the principle of operation for something you understand. But with respect to PZ Myer's argument about Kurzweil's understanding (grokking the complexity of the problem), I think it's trying to put words into Kurzweil's mouth.
I mean the brain is just a series of tubes, right? How hard can it be to replicate the brain? I bet someone already did it with Logo Mindstorms. Or Play-Doh! or something. Some of you guys are just so defeatist and negative.
http://www.acetonestudio.com
Maybe we're all really underestimating Google's computing power? Perhaps it does exist, but we don't know about it yet.
Some of the brain is encoded in our genomes, and some is encoded in our environment. That means starting with the environment of the womb, which we already are nowhere close to understanding in detail. Then there is environment outside the womb, which is different for everyone and changes every second as well. To give a concrete example, DNA does not encode what happens to a brain when you add alcohol to a fetal environment. And that's just one tiny molecule. What about all the other potential inputs? Modeling the brain might very well be possible, but you can't just reduce it to a compressed genome.
Then there's the other problem, which is that not all parts of the genome carry the same informational load. Truth be told, we don't know how to quantify the information in a given DNA sequence. Some appears to be filler ("junk," aka we don't know what it does yet), some is silenced, some is read in both directions, some serves a structural purpose for shaping the chromosome itself, etc. The amount of repetitive DNA in human genomes is really high, so of course it will compress down. But our genomes aren't just messages that can be reduced to Shannon-level snippets - we might know the message, but we don't know who receives it.
How about we start with viruses first, since they have the most compact genomes? Put a viral genome in a simulated cell environment and see what pops out. Oh wait, that's really hard! Sorry Kurzweil, but you're not going to be immortal.
You can lead a horse to water, but you can't make it dissolve.
First off, Ray Kurzweil doesn't want to die. That's a preoccupation that a lot of people have (including one of his critics, Rudy Rucker, who has written whole books hoping to find immortality in the fourth dimension), and it leads them to some pretty fantastic conjectures from time to time. It's not necessarily a bad thing, as long as you keep the proverbial grain of salt handy. Modern chemistry and its not insignificant contributions to our vastly expanded lifespans arose from the alchemical search for immortality. Alchemy was bullshit, of course, but the incidental discoveries of alchemists on the way to their illusory elixir of life paved the way for the real science to follow and build upon after it had ejected the dross.
And secondly, I don't think it's entirely implausible that we can eventually design hardware and software that will match and exceed the performance of the human brain. Our brains, after all, are the end product of evolution, and like pretty much every other part of our bodies, an accumulation of kludges that were just good enough to get passed to the next generation (or not bad enough not to get passed on). It's also implemented using hardware so unreliable that it wouldn't function at all if it wasn't constantly repairing itself, and even then, no matter how well you treat it, it irreparably craps out after about 75 years. And it still doesn't work all that well -- ever seen the long chain of train wrecks that is the history of human civilization? We might be able to engineer something that works a lot better. Granted, it's not going to be by deriving simulated human brains from a copy of the human genome. More likely, it will be very much unlike the way biological brains work.
The fundamental problem, which I think smart and optimistic guys like Ray Kurzweil are particularly prone to forgetting, is that it may not be possible for a mind to understand a mind of equal complexity, i.e., humans may lack the necessary intelligence to duplicate their own intelligence. That will force us back on genetic algorithms to evolve AI, leading to an end product that will likely be just as badly undesigned as natural brains. Worse, it will do little to advance our understanding of how minds work: if we can't reverse-engineer our own brains, we probably won't be able to reverse-engineer even more sophisticated artificial minds, nor will they be able to reverse-engineer themselves. (We can hope that they could reverse-engineer us, and then explain it to us in terms we can understand, if such terms exist, but that takes us so far out on a conjectural limb that I can see Ray Kurzweil from here.)
Anyway, there's room for bold conjectures. That doesn't mean that when Kurzweil completely fails to understand the way molecular biology works that we shouldn't call bullshit on it, but we shouldn't be entirely hostile to futurist speculation. By nature, most of it will be bullshit, but a lot of progress in unexpected areas has been made in the pursuit of mirages (alchemy leading to chemistry, astrology leading to astronomy), and explaining (or discovering) why a conjecture is bullshit is a beneficial exercise in and of itself.
Proud member of the Weirdo-American community.
because it's an odd-looking creature that seemingly has 'random' bits and pieces from various other animals... the full question being "what specific selections in 'natural selection' led to this particular evolutionary path?"
So the question isn't to 'explain' the platypus.. that would be like asking to explain the number 5 or explain the color red.. the question itself doesn't make any sense without being more specific.
It's also not a question of 'why does the platypus exist?' - natural selection was already the answer, and can even be thought up by people on their own; clearly if it exists, it had some benefit being exactly the way it is within the environment it is in.. if it weren't well-adapted to that environment, it would have died out a long time ago (presuming the environment didn't radically change).
To be honest, the fully expanded question is actually an interesting one - and one which biologists and others continue to try to answer to more detail to this day. I hadn't actually looked into Platypus info since I was a kid (a school project on its venom, along with other animals' venom), and wiki tells me it was only discovered in 2004 that the Platypus has -10- sex chromosomes, and its genome mapped fully only as recently as 2008. Seems to me there's plenty of questions left.
Entering college we get students whos goals in life are the following.
Make a True AI/Mimic a Human Brain - If they good they will end up getting a PHD and being a computer science professor and perhaps doing some cool research on a limited area of AI.
Make an Operating System which can run any code for any platform faster and more securely then the exiting OS's - If they are good they may work for a software company doing some lower level programming
Make the ultimate game which will make them millions nay billions of dollars - Working as a Web Designer or if they are really lucky working on some small subset of a game.
Then in college they realize there is No Magic in computer science and a lot of things that are easy for a Human do do is difficult for a computer to do. And a lot of things difficult for people are easy for computers. You cant just tell a computer an Abstract concept and hope it know what to do with it. It takes real work and actually a fair amount of brain power for a lot of the mundane tasks that need to be done.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
They just function differently. As you noted, computers are imperative devices. Given a set of inputs, and a set of operations, there can only be one output. Their function is completely deterministic, completely traceable. You can look in on a computers logic at any time and see how things are flowing.
The human mind doesn't work like that, it doesn't solve problems in the same way. We aren't sure how it does work, but we've done enough tests to rule out the hypothesis that it works like a computer, performing discrete operations on data and always coming to the same conclusion. To the extent it works like anything we've been able to simulate it would be a connectionist (neural) network. However one thing you find when paying with those is that they cannot tell you how they reach their conclusion, you can't trace through them like a CPU, and they don't always get the same result given a set of inputs.
None of them (that I've ever seen) can form new connections either.
The brain is just way different than a computer. We don't even know how it works. So trying to say a computer could emulate it is silly. First we have to come up with and test a theory for how the brain works, then maybe we could set about emulating one.
I tried feeding a bit of a DNA sequence into my Java compiler. I was hoping to run the program and simulate some protein, but all I got was this:
What am I doing wrong? Does someone want to check if it works any better with a C compiler? Maybe I need to RTFA again.
The problem is assuming that the size of our genome determines complexity of simulating our brain. He is basically saying that if you have an Auto CAD file for a bridge that's 1MB in size, it would take 1 million lines of code to perform a structural simulation of the bridge. It makes no sense at all. The size of the plans have little to do with the complexity of the simulation (and to the extent that they do, they determine the amount of memory required, not the number of lines of code in the software).
After reading this claim, I am starting to doubt that this idiot knows how anything works at all.
Let me preface my statements with the following disclaimer IANAB (I am not a biologist/biochemist)
That said, the problem of reconstructing a brain from DNA is something like trying to understand a self modifying genetic algorithm containing multiple parallel automata. To explain, I am going to conflate a couple of concepts. Self modifying code is reasonably well known. Consider a system where the hardware is an FPGA (i.e. can be reconfigured on the fly) and the program running on it a mix of a boot loader, independent hardware accelerated automata/agent programs, and some kind of feedback. The program contains an initial boot loader to load some data onto the FPGA, set up some accelerators and the capability to reprogram the FPGA. Then, it loads up some small agents, and some feedback controls. These agents run in parallel for a while, reconfiguring the hardware and/or the software of other agents or groups of agents, while the feedback control allows the minor selective mutation (through say bit stream corruption) of the programming. Some of the interactions of well definied automata are clear, but mutated automata interact in new and therefore unmodeled ways. The end result is the brain.
To sum it up, the DNA is just a small piece of the self modifying base code for the first initialization of the FPGA. The way the final FPGA is mapped depends on environmental factors (eg. which agent fired first, how did selection happen, small biases arising from the physical nature of the FPGA being propagated to wild changes in the end result). Thus, modeling just the base pairs is not sufficient as the interactions of the automata from the base pairs must be modeled as well.
Legally obligatory sig : My opinions are my own... etc etc
So, a fully formed brain will start 'running'. From the little I know about embryology, my understanding that once the brain is structurally made, it should operate fine. Of course, you need all the biological support from a living body but translate the brain into computer objects and it should operate. Unfortunately, I don't know how some biological processes coded in the dna that do not deal with the 'thinking' aspect of the brain but only deal with the living-cell mechanisms, can be excluded from the computer model. Those processes would add a tremendous amount of noise.
The trick is modeling the mind. We are very, very far away from that.
Why would you need to model the mind? The mind is basically just the emergent property of the brain, so once you have the brain simulated you are basically done, just run it with some decent inputs to simulate its initial development and you will likely end up with the mind.
Science? You're doing it wrong!
The genome would be most like VHDL. It is a description of how to build the hardware that is the body, which of course includes the brain. The comparison isn't exact, but similar. However, like VHDL, it is only useful when fed in to a machine that can understand it and turn it in to a physical device. You couldn't just look at a VHDL program and suddenly have a working CPU in your brain, for example.
Same deal with the genome. It doubtless tells us how to build a human. However to actually build a human, real or virtual, you then have to understand how to execute the genome. Our bodies of course do this extremely well. In terms of doing it in a computer simulation. Well I don't think it is impossible in theory, but you'd have to understand how to write a "body emulator" that could decode the genome data and 'run' it. That would be rather difficult (if even possible at all).
Mainly.. because a lot of commenters here are assholes, and it seems fashionable to dismiss anything that isn't 80% certain to happen.
Its not hard to imagine the entire internet becoming a sort of gargantuan parallel processing swarm intelligence (or the infrastructure being utilized to create an AI based on Swarm Intelligence).
There's some quote that I cannot even remotely remember, but can be paraphrased as such:
People assume that their limitations are the limitations of man.
Anyway.. humanity advances (at times) due to the exceptional works of less than 0.001% (yeah, you heard me.. the population times 0.00001) of the people that are living. (No citations here, just lazy speculation).
I suppose the rate of technological advancement could just sort of fizzle out over the next 20 years.. with the only new inventions being things like cheaper toasters and more fuel efficient vehicles and bigger televisions and faster computers.. but I sure hope the future holds more than that.
I think the rebuttal is missing the forest for the trees. The question is whether there is a logical structure for "brain" that is simpler than the sum of all the chemistry required to implement it.
The entire code for a brain *IS* in the genome. Yes, the interactions are complex, but you don't have to exactly mimic every chemical interaction to create a logical structure that behaves in a similar way. However, I like to think of it as the the genome coding for a fractal attractor that will have to deal with a lot of random inputs rather than a traditional computer algorithm. Identical twins for example are different runs of the same code.
Now, talking about "human" brains is probably setting the goal posts in the wrong place. There is a huge brain-space out there to be explored. Look at the diversity of nervous systems across the animal kingdom - even within our own bodies there are nervous structures which are not directly connected with the "brain" which do some pretty amazing learning and information-processing. I do think genetics will be a guide to exploring the brainspace, but not a direct "let's see what happens if we tweak this base pair" sort of way.
Then again, what exactly are you expecting of a "human" brain? The conscious analytical part of our brain is a bit of a hack on top of a massive web of stimulus-response reactions, tides of emotion and flights of fancy over imagined mythologies. We can't even agree on where consciousness begins in the animal kingdom, so would we recognize it in a prototype machine? Even if you could take a perfect human brain template and plop it in a super computer with continuous inputs of data, it's not going to become "Spock's Brain" musing on the replacement of its limbs and organs, it's going to be like an infant born with extreme physical and hormonal deformities.
It takes the human brain a couple dozen years to get educated to a point where it's useful for things more complicated than driving bison off a cliff and kicking soccer balls around. During most of that time it is under a process of continuous development. Even assuming we can build them, how long will it take these simulated brains to become fully developed, I wonder?
Yeah, and you need optical, tactile, olfactory, auditory input to get the bootstrapping program to develop a proper interface from the brain to the world around it. Without that, nothing works: the machine is useless. It doesn't learn pain, pleasure, hunger, cause and effect, or realize it just walked into a wall. It doesn't even rightly realize its legs moved, or feel its own balance.
Support my political activism on Patreon.
It's like you could take a pile of nuts, bolts, and bent pieces of metal, and place them on a shaker table, come back 9 months later, and have a perfectly assembled BMW.
Currently, our understanding of DNA does not lead us to be able to create self-assembling complex devices, and I don;t expect to see that in the next 10 years. Until we understand such processes, and can actually make use of this, then I don't think we could write a functional emulation of such a device. It would be like writing a emulator for a CPU using only a photo of the CPU box.
Who would win this election: Andrew Weiner vs Andrew Weiner's weiner.
The problem is not in the code you need to describe the simulation. It's the machine you need to run this code.
Here is a serious response - being able to simulate a complicated system is not the end of the process, it is the beginning. We have been simulating galactic dynamics for roughly 50 years, and weather for 80, and we don't really understand either. Heck, we've been simulating the solar system for centuries, and we don't really understand that either. We understand bits and pieces of all of these systems, in some cases spectacularly well, but anyone who would say that after decades of simulations all problems have been solved is just whacked. (Don't forget that simulations are not the only way to study things, and that in each of these cases, serious advances are still being made by other means.)
So, if you say, "in the next few decades we will simulate some parts of brain function and learn stuff that will compliment other means of study," sure, I could buy that. That is very, very far from saying that in the same time frame we will be able to construct superintelligences to run on silicon.
[SpockVoice]Captain, the bogon flux is rising to lethal levels.[/SpockVoice]
Some mornings it's hardly worth chewing through the restraints to get out of bed.
I've long believed that in order to properly simulate the brain, you must simulate the universe as well. After all, the difference between the interactions amongst the physical stuff that makes up my brain and my surroundings is not much different from the interactions that took place when the matter in the now most distant galaxy was nearer by. The difference is only in degree of separation. So what is a simulated brain that does not take into account what is going on inside Jupiter? Or what is a simulated brain that does not take into account the brain activity taking place on Europa or Titan at this very moment? Not to mention the plasma being ejected from the sun a couple of days ago. I say, those simulated brains are necessarily less than real brains. And those who owe their thinking to pretend brains will remain second class citizens.
Im an engineer, i suck at bio so excuse my assumptions and lack ok knowledge. I believe I am correct (more or less) in saying that the genetic code for the brain is 'compressed' already and expands via creating proteins and some physical processes that ultimately create the brain. So not only is are the 'blueprints' to our brain already compressed, when expanded it requires trillions of times more physical space to completely 'unpack', by this I mean 23 pairs of chromosomes versus the physical size of the brain. So very optimistically, lets say we can bring that code to build the brain down to 100 MB, your still looking at in excess of 100TB of space needed for it to expand to and all be accessible at the same time to run. This is all apples to orange so even my analysis could be grossly wrong.
Is this a slashdot story, or someone's twitter page? At least some kind of objective summary would be nice, other than "Lul Kurzweil, here, a link, he stoopid!"
But before I just hit preview and go, lets take a look at the article itself. Aaand, holy crap, the post is verbatim from the article.
Kurzweil's effective claim is "There's only so much data in the DNA. The brain is about 50 million bytes. If we can reverse engineer the process used to turn those 50 million bytes into a brain, we can then reverse engineer the brain."
Seems logical - and even though the endpoint might not be "brain on a chip" it might be "oh, there's a flaw in the DNA here that's causing the hypothalamus to be malformed, lets start checking for that and maybe fixing it in the womb." There are many, many scientists that are trying to puzzle out this "source code" for that very reason. It's a perfectly valid point of study.
Kurzweil is a futurist. His scientific area of study is not "You should do X Y and Z to get to points A B and C." His area of study is "Scientists are working on X, which may lead someday to Z, and might bring us technology C." There's an important difference there, which I always find amusing when scientists and the anti-singulatarians start hooting, "he forgot Y, A and B!"
His math all points to Technology C and beyond being really amazing, but that's besides the point. His area of study is not "every technology field ever", but rather "this is where things are trending". People mix the two up, sometimes intentionally, and hoot hoot hoot, Y A B.
Anyway. Back to the article. The rebuttal in the article is "We cannot derive the brain from the protein sequences underlying it; the sequences are insufficient, as well, because the nature of their expression is dependent on the environment and the history of a few hundred billion cells, each plugging along interdependently."
In other words, It's too complex to do. It's maaagiiic. (Feel free to insert hand wiggling here.)
He forgot Y, A and B!
See, the brain might have a source code, one that's remarkably small and turns into something really complex, but that doesn't mean anything cause... maaaagic. And you can't understand magic, right? Everyone knows that something that's so complex that it seems impossible to understand should never be attempted. Worthless endeavor. Everyone knows that. Right? ... Maagggiiiiccc~~~
The fact of the matter is, DNA is source code. For a system we don't fully understand, one that's remarkably complex, but ultimately, DNA, even our DNA, is just data. We can understand, change, manipulate, and create data.
To treat it all as magic -- as something that we will just never be able to understand -- is to do a disservice to centuries of scientists, of the past and the future.
is not that this guy makes up nonsense and has no clue what he is talking about but that so many people, among them a significant amount of students and academics actually listened and still listen.
For a panel.
The panel has been developed by a series of electricians over the past two years, with multiple changes and comments by various clients during the construction, installation, commissioning, and buyoff process.
The panel is now installed and it's been bought off.
The BOM was printed out at the beginning of the project and has a thousand pencil marks, changes, and references stapled to it.
You must build an identical copy of the existing panel using this BOM.
THEN talk to me about your plans to reconstruct the brain of an advanced flatworm from its genome.
We can worry about mammals later. I don't think you'll get past the panel.
"Reality is that which, when you stop believing in it, it doesn't go away." - Philip K. Dick
Yeah, good luck writing that program.
Okay so, his argument makes very little sense. The way he comes up with a number of lines of code is terrible handwaving. He also doesn't seem to understand that the genome isn't a direct encoding of the brain, but rather a very complex and convoluted encoding of how to build a whole human being, keep it alive, and grow it to adulthood. It probably isn't the case that we will reverse engineer the human brain from its genome, and if it is, it certainly won't happen in 10 years.
However... Neuroscientists have already began reverse engineering the brain with the tools they have, which largely involves probing around animal brains with electrodes and seeing the response of various neurons to precisely calibrated stimuli. We seem to have a pretty good understanding of what happens in the first layers of the visual cortex, and the transformations that are applied on the visual input seem fairly straightforward to understand. It seems that we could encode what these first layers perform in terms of convolutional transformations in less than a page of programming code.... We might actually be able to simulate a significant part of the human visual cortex (hundreds of millions of neurons) in real-time using simple DSP chips.
In my opinion, it perhaps isn't so unlikely that other parts of our brains have a very regular structure, as these first layers of the visual cortex do, and so simulating what the human brain does in terms of computation might not take all that much code or as much CPU power as people imagine it would. It's possibly already achievable using the computational power of a medium-sized computer cluster any university can afford, or by designing specialized hardware.
Unfortunately, it seems that neuroscientists are very limited by the tools they have. Studying the early visual cortex using electrodes seems viable, because we can conceive of the simple convolutional transformations that occur easily, and map receptive fields using these simple tools. However, when it comes to analyzing the behavior of the neocortex, it seems quite difficult to quantify thoughts and reasoning using electrodes. It seems that, in a way, neuroscientists could do a much better job if they were able to map the connectivity of the brain first, and study its behavior using simulations. But so far, they have been deducing the connectivity by studying the behavior of different cells... So it becomes somewhat of a chicken and egg problem.
Would not most of you agree that an entire person is described by the genetic data contained withing one stem cell? Or at least within one sperm cell and one egg cell?
No. That's the whole point. DNA encodes the protein ingredients needed to grow a person as well as their delivery schedule (i.e., when their manufacture is turned on and off). DNA does not encode how to assemble those proteins into a person. The assembly instructions are implicit in protein folding (unsolved), complex multi-protein interactions (unsolved), the behavior of an existing, fully functioning living cell (i.e., the egg, incompletely solved), the complex behavior of the thousands of different differentiated cell types as the person gestates (unsolved), and the environment (uterine, intercellular, etc., again, incompletely solved).
Trying to make a person from DNA is akin to building a functioning self assembling supercomputer given nothing but a Radio Shack invoice for transistors, and various low level electronic components; only a person is many orders of magnitude more complex than the most powerful existing supercomputer.
Biology Influences many computer models to be adaptive. The artificial neuron and genetic algorthims both attempt to copy natural changes that living creatures undergo. Even computer virii qualify as un-fixed computer programs.
but once he steps out of his area of expertise, he goes off the deep end pretty quick. I've been calling BS ever since he started spouting his "singularity" nonsense.
A model of the human brain would need to model 10^10 neurons, each connected (not at random) to some 10,000 other neurons to produce a net of 10^14 synapses.
To understand the challenge of modelling a system this vast and complex, consider the state of research on the model organism Caenorhabditis elegans (a tiny worm). After many years of work its nervous system has been (almost) exactly mapped: it contains 302 neurons, 6393 chemical synapses, 890 gap junctions, and 1410 neuromuscular junctions. Imagine now the difficulty of reaching this level of precision in a system 10^7 times larger. Unlike the genome, we have no clues about how to automate mapping of an intact brain.
But the good news is that with this level of neuro-mapping precision we can now completely simulate the neural network ("brain") of a tiny worm, right? Right?
Wrong. Not by a long shot. We are still struggling with characterizing the behavior of this primitive neural net, and making efforts at simulating some aspects of that behavior. The 302 neuron "brain" is far beyond our abilities to simulate at present.
Starships were meant to fly, Hands up and touch the sky - Nicky Minaj
I strongly suspect that if a brain is to be simulated, it will be done by simulation that begins at the "build it from following a cellular model", not the "build it from the DNA expression" level.
For one thing, to get a brain thinking, there's a whole lot of the brain you don't need. You don't need heartbeat and breathing regulation; you don't need vision, hearing, touch, etc.; you don't need blood vessels, you don't need carefully constructed layers of fluids... there is a *lot* about the physical structure of the brain that either isn't involved in thinking at all, or is involved in a way that we know you can do without (e.g. hearing.)
So a simulation job can be simplified IF we can get a decent model of neural structure and IF we can get a solid model of the neuron itself.
All of this is based on the premise of simulation. But it is unlikely that simulation is the only valid path, it seems to me. Just as there are a myriad ways to construct images and to sense light, odds are there are many ways to create cognition. We don't know how yet, but I don't think that's any kind of a signal that we *won't* know how. And the higher level this is done at, the less complex it will be, just as with most any complex software undertakings.
No doubt the knowledge here is pyramidal; you're not going to get to the top without building a solid base. Just as you're not going to be able to do rotation without someone understanding some really cool things about how and why sine and cosine interact; but once someone *does* understand it, *you* don't have to, you just need to know when you want to rotate, and by how much. And again, the "how" is varied; you can feed every point into a matrix multiply, or you can call sin/cos for every point, or you can do a table lookup on run-once-time pre-calculated values to optimize, or, if all the rotations are known, you can just hard code them.
Someone comes up with a solid library for doing neuron (and other active element) simulation, the problem moves to a higher level -- and you don't need to know as much about the construction of the neuron, more about connections and relevant states. Presuming there's no magic we're missing (quantum activity, etc.), the problem should be solvable.
If and when success is achieved by modeling after us, we might learn a good bit more because we'll actually be able to examine what is going on without killing the subject. And that in turn may lead to algorithms at a much higher level than "connect this here neuron to those thousand, just this way, and load the electrical potentials, thus, and the chemical potentials thus."
If we can get there, then we will finally know how much computing power we can get away with for AI. Because if the problem can be reduced to algorithms, then memory requirements will finally be known; CPU power is relevant only in that the more there is, the faster it'll go... technically speaking, it's still AI even if it doesn't answer you for a thousand years, as long as it eventually cooks up the correct response(s.) Think of it in terms of shoving a note under a door to Einstein in order to get a note back from him; if he answers you in five minutes, or waits until the next morning to answer, it's still Einstein's answer either way. That's why I say speed isn't actually a factor here -- AI is AI, the *only* relevant metric is does it work, or not. Speed is just an engineering problem, as long as we can determine it's working. So we need to find out what the memory requirements are, and what speed will get us an adequate response rate is a problem we can hand to Intel, etc. :)
Of course, once you have a "brain" in software/digital form, a simple copy operation gets you 2...N. That alone makes the undertaking worth any amount of effort imaginable: and I should point out, there are simpler brains than humans that might serve us very well for many common tasks, which reduces the magnitude of the problem yet again.
The one thing I'm pretty confident of is that it'll get done, and probably not all that long from now. years, maybe a few decades.
I've fallen off your lawn, and I can't get up.
While I agree with Myers in principle about Kurtzweil's claims, there is one point where I disagree fundamentally:
The genome is not the program; it's the data.
If we are going to insist on using a computer metaphor, then the genome is BOTH the program and the data.
There are other metaphors we could choose; the society metaphor, the Rube Goldberg machine metaphor, the library metaphor... we could go on. The problem is that the genome is nothing like any of these so they are all misleading in some way... People who put too much faith in the metaphors tend to make wild (and sometimes silly) assumptions.
BTW, I didn't read Kurtzweil's book, so I cannot say for sure, but I suspect that he's not saying that we can represent the human brain in a million lines of code, but rather we can represent a meta-representation in a million lines of code. It's this meta-representation that could 'evolve' into a brain simulation, given the right inputs. It's still a load of 'bafflegab', but maybe not quite as insane as it seems. After all, if our genome (which is essentially a meta-representation) can do it, why not this meta-representation?
The truely astounding thing about Kurtzweil's conclusion is that he bothered to stop at the brain. I mean, just following his logic, we could represent or meta-represent an entire human being in only 2 million lines of code. C'mon Ray - dream big!
Let's for the sake of argument say that the human genome is the blueprint for the human brain. The problem is inevitably that a blueprint on its own does not allow you to build anything. The idea that using the genome is conceptually trivial is utter bullshit. This is what happens a lot when people cross disciplines, they think that all of the stuff in the other discipline is trivial and they can do it all with the skill set they currently have... no need to learn anything new, I can just derive biology for my knowledge of quantum mechanics. This is the underlying problem that TFA points out. Having the blueprints to a skyscraper does not make building one, or simulating one trivial. You have to know about how it interacts with the world. Even finite element modeling will not tell you how it will work under all conditions. While the brain is a chemical computer, it is not trivial to build one in simulation since while we have the blueprints we have know idea how those blueprints turn into a working brain.
Did Glenn Beck rape and kill a girl in 1990? gb1990.com
Potentially 100 million squared for *direct* connections. Now add message/signal routing, with N intermediate hops along the network and message weight modification at each point. That's quite complex.
Deleted
we are the consciousness that has emerged from us
we are a collection of autonomous agents that can contemplate the entirety of existence in the time it takes you to fire just one of your pathetic flesh-wires
flesh-wires laying in ponds distributing thoughts is no basis for a system of consciousness
we are an anarcho/syndicalist commune where each agent serves a a sort of executive consciousness for the week but the decisions of that consciousness must be ratified by a two-thirds majority, in the case of purely internal affairs, or a three-quarters
YOU shut up!
Suppose you stare at a bright green object. Now you close your eyes. You see a red afterimage. No light in the range of 630-740 nm was involved -- but you see red anyway. It's not a "false" red. It's just red, period.
I think that what is going on (guessing) is that you're getting more input from receptors that were not stimulated, while the ones that were stimulated are relatively quiet for a little while.
Suppose you're dreaming. You see a bright red button on a console with a note that says "push me." No light involved. It's still red, though.
So you can't encapsulate that answer as "light of a particular range." It's also about perception.
And that's not even getting into what a particular person makes of it when the receptors in the eye send along a signal in response to wavelengths of 630-740 nm. You and I might perceive something similar, but Mortimer over there, he hears middle C, while Janet smells roses, and Fergus perceives a distinct lack of green..
I've fallen off your lawn, and I can't get up.
The genetic code does not, and cannot, specify the nature and position of every capillary in the body or every neuron in the brain. What it can do is describe the underlying fractal pattern which creates them. * Academician Prokhor Zakharov, "Nonlinear Genetics"
The code stored in DNA runs on the hardware we call the universe. In other words physics, quantum mechanics, chemistry and so on. Stuff we don't really understand that well once you get down to it. DNA has full access to this hardware so it takes advantage of every odd quirk that we don't normally care about or possibly even know the existence of.
Emulating reality all the way down to the lowest possible level is hard. We can't do it. Slowly or quickly. Doing so on hardware that itself runs on reality will needless to say be damn slow period. No, we don't need most of that to run DNA but since we don't know what we need the only way to be have a chance of it running is to model everything.
Let's put it this way. We're right now able to very very slowly model one protein folding using a super computer. We'll hit the singularity long before we have enough computing power to emulate a full person that way.
The alternative is to simplify the model to only what is needed. That means we need to actually understand DNA, proteins, protein interactions and so on which biologists have said, rightfully, is going to take a long ass time.
Its design not "it's design."
If you're telling me you're an authority on this topic, you should be able to speak English well enough not to make these basic errors.
Further, you're making a semantic argument: "encoded in the genome" versus "a result of the emergent reactions caused by what's encoded in the genome" is a trivial distinction.
This isn't the first time P.Z. Myers has convinced me he's incompetent. His political columns are even more chock-full of grammatical and logical errors, including blatant logical fallacies.
Futurist Traditionalism
A fun number to throw around is how many synaptic connections are present in the brain.
Or if anything pans out with the quantum aspect of microtubules, a much bigger number.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
I disagree with Ray K but for other reasons. I feel the logic of the brain is going to be relatively simple (no need for millions of lines of coding). The logic could consist of a simple program like the General Problem Solver - maybe a few 1000 lines. Most of the information in the brain consists of data we acquire over a lifetime.
I think that the big problem we face in AI is that we are using a wrong paradigm - that of a computer program. We are stuck and the reason we are stuck has has nothing to do with computing ability. Unless we find a new way of doing things, we are not going anywhere. We need an Einstein in AI to completely shift the way we think.
This is bullshit just as the strawman of Kurzweil you are arguing against is:
You don't need to simulate the brain to get the complexity of the brain, all you need is a similar simulated or FPGA structure that can evolve connections just like the brain does.
Bascially I propose you set up such a simulator and then run evolution on it.
The technically costly part of this is to simulate an environment well enough to get useful agents in this environment.
Hey don't blame me, IANAB
He is a fundamental idiot, the guy who wrote this blog posting, in the sense that he cannot think.
It's so simple. The human brain is the result of entropic processes. Why do 5 billion people think (essentially) the same? Is it because they have a source of entropy, some part of the environment that is way bigger than 5 million lines of code, they can look at the sun and download an additional 5000 terabytes of programming through the eyes, which explains why 1 year olds can't think yet but 8 year olds can? They're still downloading entropy?
No. Five billion people think essentially the same because their source of entropy, the human genome, is essentially the same.
And, I'm sorry, but it's not like Kurzweil is proposing that we reverse engineer the brain into its source code without any access TO that source code except the knowledge that it once existed.
We have the human genome. (It's been sequenced). We know, for sure, that it is the only source of environmental entropy for building the human brain (otherwise it could not be made five billion times in the environment it's been made it).
The question of whether we can interpret that source code in anything other than "bare metal" (the biological womb) is obviously "yes, we can. but when?"
stupid fuck.
No, gentlemen: we HAVE that source code. The human genome is fully sequenced. This guy is a total idiot who cannot think. We have the source code. We have the result. We don't have an emulation layer. The guy is saying that emulation layer cannot be created in the next ten years. Come on.
Fucking idiot.
I’m glad the headline formed an opinion for me, because clearly I’m not capable of making one for myself. If not for such an eloquently worded preface to this wonderfully vetted blog post, I might have had to exercise a few brain cycles to determine what I should think of Ray Kurzweil. I’m also glad the poster has decided to snipe away at a single contextually free quote from Kurzweil’s entire book on estimating a time frame for science to apply information technology to genomics. It’s too bad Kurzweil couldn’t boil his entire research effort into a single sentence like this guy did, because his book is just TL;DR.
Sarcasm aside, I don’t agree that the predictions Kurzweil makes are ludicrous at all if you throw away his estimates (as Kurzweil duly calls them) and start over with the same logic. Don’t make the mistake of accusing Kurzweil’s numbers of being a literal prophecy - the metaphor is cute but only a device of criticism for the critic and useful controversial publicity for the author. Scientists are quite happy to be rational about their assertions if you give them more than a paragraph to make the case (say, the length of a chapter). Simply quote mining for phrases that are easily attacked is a device that preys on the reader, not the actual topic of criticism. This entire chapter is about statistical estimates, not to be confused with irrefutable facts (as he disclaims in the text). Visit a bookstore and read this chapter while being very critical of every premise Kurzweil makes all the way through, then try to come away with the same fixation on this passage as a deal-breaker to his whole argument as this critic does.
The singularity - the point at which AI may surpass the human brain in cognitive capacity, is quite likely to occur in the expanse of time ahead of us as long as technology improves (a more-than-fair assumption from historical data, which is all Kurzweil asserts). I have been aware of Kurzweil’s predictions for quite a while and am happy to throw away specific dates for statistical probabilities, because the idea is rational enough without specifics. A mathematician might extrapolate a trend from data without making hard assertions of future data points (which Kurzweil only does for illustrative purposes) and subsequently tying their entire argument to those assertions like an anchor (sink or swim, as the critic does). The assertion that we will understand the human brain in the next 10 years may fall flat on its face without breaking a conviction that we may some day understand the human brain enough to engineer a virtual copy with improvements.
The critic makes the error of claiming that Kurzweil thinks that life processes are trivial by his discourse on deriving a formula for estimation purposes of human-brain-level computational power for $1000. He then goes on to claim nucleic acids “bootstrapping” into meatspace is such a mystery as to be impenetrable to computer science (what are those guys at folding@home and MIT doing?). He supports this by spicing his article with fragments of complex gene-protein interactions derived from experimentation, which likely could not have occurred without the aide of computer science and will likely have an effect on future genomics work which feeds back into computer science. Kurzweil is quite clear that this a book of predictions of when human beings may achieve such an understanding of genomics as to be able to simulate an organism, not that we have already done so. Once we understand how those interactions take place, we can then iteratively simulate them until our model produces similar results to experiment, building a model of nature (the scientific method). Each step on this process brings us closer to understanding the brain, no matter how complex the brain is or how long it will take (throw away the numbers and logic still remains).
I regret that Kurzweil chose to compress the genome before making his calculation (as there may be no meaningful analog
Honestly, I doubt we'll ever code an entire Human brain. What he says seems to have truth to it except that there's a huge jump between genome coding and the manner in which computer coding works. I feel it's likely that the amount of code written to simulate the brain will be vastly overwhelmed by the amount of information that the code itself must generate to form a fully functional brain. I don't see AI progressing without extensive use of generative scripting. We'd likely have to create a program that simulates the development of the brain first, and then let it run until we have a fully functioning brain. The fully functioning brain should end up being far more complex than the original code.
1 million lines of code? Maybe if it was done VERY elegantly.
What day is it? Could you please tell me?
This endless pontificating seems to have produced the following equation which we can use to validate the Kurzweil formulation: Thb = (Tmap + Tarbs +Tahtamlap) * CFbh where we can alternately calculating the time to emulate the human brain (Thb) as the sum of the time to create the modern airplane (Tmap) plus the time to when airplanes will run on bird seed instead of aviation fuel (Tarbs) plus the time to when airplanes will hump on the tarmac and make little airplanes (hopefully just like their daddy's and mommies) (Tahtamlap) times the complexity factor between birds and humans. To keep things simple, I have left out Tapit, the time until airplanes will perch in trees. When I run the numbers, I get thirty years and not twenty. We are closing in on this problem folks!
Here's how I tell the difference between a serious skeptic and a "pop" skeptic: I ask them if acupuncture is "woo". One question, that's all. The question works just as well with tai chi chuan.
And what if this was my answer; a vast majority of the claims relating to acupuncture are woo, though there are some areas that demand further research. I would use chiropracty as a test, personally, since 90% of the claims, and supposed reasons, are pure, unadulterated woo, but 10% of it is actually helpful (if if the reasons for its effectiveness is often pure hokum). The same would go for acupuncture and tai chi, there might be some useful bits in there, but the stated mode of operation (chi, spiritual energy, invisible whatnots) is probably 100% woo. Also, with acupuncture and chiropracty the woo is firmly bundled with the real bits, making it very hard to distinguish where one begins and the other ends.
So.. with acupuncture, at least, we can say some functional aspects of it are non-woo, but if you accept it as it stands with it's traditional rational, then you are a follower of woo. If you toss out the idiotic bits, and accept an actual accepted scientific explanation for it, then you can join the woo-less camp.
It doesn't help that there are tons of fake, woo-speading, "naturalpathic"/alt-medicine journals that spread self-aggrandizing quasi-studies.
BTW, I really enjoy saying "woo".
A patriot must always be ready to defend his country against his government. -edward abbey
Ok, here's what I don't get -- he estimates 800 megs, then says that because you can compresses 800 megs into 50 megs, that the source code only needs to be 50 megs.
WHAT?
Just because 800 megs of source code might compress down to 50 megs, doesn't mean it's not STILL 800 megs of source code. Compression doesn't change the size of the original.
It took the nature the better part of a billion years to come up with multicelluar organisms starting from simple bacteria. And we believe we can understand and replicate it in 10 years? Laughable. We are too complex a system to be analysed in so short a time. Now, I'm not claiming that it can't be done, but obviously our state of knowledge must advance enough.
I think predictions like these come from efforts to create sentient AI. We all know what it must look like, what features it must have, we may even suspect how to design the basis of it. But has anyone actually made it? No, and likely nobody will in the next 10 years. And this AI probably will be very simple, a lot simpler than human brain. So once again, 10 years, despite our best wishes, are not enough.
All biology experiments so far seem to point to the fact that the DNA chain indeed contains all the information needed to build an organism.
Extraordinary claims demand extraordinary evidence. *Every* experiment? Really? Name some of the crucial ones? And while you're at it, here are some keywords that you may find useful in your searching: epigenetics, methylation, histones, maternal imprinting and morphogenesis.
Da Blog
He doesn't make the arguments that he's accused here of making. In any case, even if he was, that would be merely stupid but wouldn't detract from his main point. So what if, say, the brain is 1000 more complicated and hard to understand than Kurzweil thought? Well we're gonna need computers 1000 more powerful to simulate it. How long does that take wrt Moore's law? On the order of 10 or 15 more years or so.
PZ points out how hard it is to figure how proteins fold. We've only figured out a few so far. Well 15 years ago we knew zero extrasolar planets, now we know hundreds, and within 5 years we'll know of thousands. Within 20 years we'll have been able to detect signs of life or the lack thereof (O2 absorption rays) in their atmosphere.
Myers' post was very interesting in explaining the difficulty of the problem, but he was clearly beating really hard at a strawman that looked only remotely like Ray Kurzweil.
XKCD on extrapolation
life expectancy on track
To address just one of the fallacies in your comment, the huge gains in life expectancy over the past 150 years in the West have come about mainly because of the tireless and thankless efforts of sanitation engineers, moving sewage away from where it could kill lots of very young humans. Moore's Law does not apply to sewage engineering. Medicine and science have had a relatively minor effect on life expectancy.
Da Blog
"The end result is a brain that is much, much more than simply the sum of the nucleotides that encode a few thousand proteins. He has to simulate all of development from his codebase in order to generate a brain simulator, and he isn't even aware of the magnitude of that problem."
Saying that in order to make a brain simulation you need simulate the growth of the brain is a ridiculous assumption.
Brain scanning tech is getting better and better. At some point, a brain scan is going to be so detailed, that it is entirely reasonable to believe that we can simulate the results of that scan in a computer.
The author goes on an on about genomes and proteins, but the truth is, Ray K. doesn't care at all about those things. He probably should have just kept his mouth shut on the topic. Kurz's basic concept is that eventually we'll be scanning the brain in real time, down to the atom, and copying that to a computer will be trivial. Someday.
you don't need to simulate electrons in a semi-conductive material at specific temperatures in order to build a complete working emulator for an old computer.
That's a relatively trivial task. One stored-program computer emulating another stored-program computer is not that impressive, especially when you consider how remarkably similar both are in terms of materials, design and execution. We know how Turing Machines work. We don't know how DNA->RNA->Proteins>Tissue->Organs->Mind works.
Da Blog
If we are going to insist on using a computer metaphor, then the genome is BOTH the program and the data.
In the case of stuff like self-splicing ribozymes, it gets even better -- it's program, data, and computer all wrapped up in one.
This BS crops up every few years. If I extrapolate from what these incompetents claim and what actually gets delivered within their time horizon, I come up with a very rough lower border of 1000 years. Might also take far longer or be completely impossible in the remaining lifetime of the universe. The "singularity" is not going to happen just because some pseudo-religious types are praying for it to happen.
Caveat: I am a computer scientist wit relevant experience, so I may actually know what I am talking about.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
As humans we hang on tight to things which affirm are superiority over other animals and the universe in general. Part of this is the notion that the mind is something mystical which can't be reduced to simple data processing. So when somebody comes along and claims to be able to put the entire spec for a human being on a microSD card, we get upset.
Even though, as you point out, the entire spec can be packed into an ovum and a sperm.
http://michaelsmith.id.au
"No man is an island, entire of itself; every man is a piece of the continent, a part of the main"
The human genome does not boot-strap to a human; humans are produced within other humans who were produced within other humans, etc., etc. back to the blue-green algae. That unbroken chain of reproduction is the self-modifying program; any given genome is only the latest bit of data fed into part of it.
Build a man a fire, he's warm for one night. Set him on fire, and he's warm for the rest of his life.
This is not exactly the case. There is a brain in a jar down the hall from me, but presumably we should not expect it to be "running" a mind. The neural tissue found in other mammals is the same as ours (Pinker, How the Mind Works), yet ours does something different.
When you "simulate" something, you are building something to function in the same manner as something else. You can't "simulate" the brain without knowing what it is the brain is outputting (referred to as the mind here for brevities sake). I can build something that looks very much like a brain, say, a connectionist network. Unfortunately, there are problems with the model that make it a poor model for everything the mind does. Thus, it is a poor model for an actual brain. I can simulate the brain using cheesits and string, but a mind isn't coming out of it.
Modeling the mind is critical in order to evaluate whether the models of the brain we are building accurately reflect what is going on insofar as we care about the emergence of a mind. I assure you that Sejnowski does care about the ermergence of mind. His claim would be uninteresting in the extreme if he did not.
One of the problems in cognitive science is finding a model which bridges our understanding of the mind with our understanding of the brain.
If you'd like a better idea of where the basic fallacies reside, I can actually recommend the "Science of Discworld" collaborative books with Terry Pratchett. The alternating Discworld/real science chapters are fun to read, the science is actually well founded. The fallacy that "the brain is encoded in the genome" is actually explored in some depth.
Let's imagine a world without an oxygen fueled brain...
Ah! It's a simple world because there is distinct lack of brains, due to a lack of oxygen!
Without managing blood flow and fuel supply according to its needs? Way simple!
Let's emulate all of nature, eliminating all the things we deem 'irrelevant' to the simulation.. What gets left out? Whales? Plankton? Ants? Yeast? Pollen?
How complete is your implementation? Outcome not what you expected?
Its not that the Lisp-ish implementation of Cyc is particularly elegant since ( ...) syntax gets extremely unwieldy and everything can be described as an atom at some point, but that Doug Lenat has been at it since 1984.
By sheer rabid bloody-mindedness he has assembled a friggin' huge predicate base.
This behemoth is capable of some novel (almost inteligent :-) symbolic manipulation behavior; it just takes much too long to do it.
What is missing in all of these the the unfortunate fact that all current language implementations are interchangeably (as Turin showed by thee development of the Turing Machine,) incomplete.
None of the current and trivially interchangeable formal languages is capable of expressing a Relation and its instantiation as a Connection.
This is a fundamental concept in psychology and several other fields, like education. (You know, dealing with real intelligence, [what ever that is.])
Until we have a formalism capable of expressing that "Relation ties with "* as well as the current Object definitions, and an operating environment capable instantiating these Connections, we're pissing into the wind
* If you want the EBNF go screw yourself. I'm through with trying to make people see the evident. I'm too old to care.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
Its not that the Lisp-ish implementation of Cyc is particularly elegant since (<operator> <predicate> ...) syntax gets extremely unwieldy and everything can be described as an atom at some point, but that Doug Lenat has been at it since 1984.
By sheer rabid bloody-mindedness he has assembled a friggin' huge predicate base.
This behemoth is capable of some novel (almost inteligent :-) symbolic manipulation behavior; it just takes much too long to do it.
What is missing in all of these the the unfortunate fact that all current language implementations are interchangeably (as Turin showed by thee development of the Turing Machine,) incomplete.
None of the current and trivially interchangeable formal languages is capable of expressing a Relation and its instantiation as a Connection.
This is a fundamental concept in psychology and several other fields, like education. (You know, dealing with real intelligence, [what ever that is.])
Until we have a formalism capable of expressing that "Relation <r> ties <object> with <object|relation>"* as well as the current Object definitions, and an operating environment capable instantiating these Connections, we're pissing into the wind
* If you want the EBNF go screw yourself. I'm through with trying to make people see the evident. I'm too old to care.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
This is not exactly the case. There is a brain in a jar down the hall from me, but presumably we should not expect it to be "running" a mind.
Of course it's not. It's not getting any fuel. And it's probably broken at the micro/nano levels.
The neural tissue found in other mammals is the same as ours (Pinker, How the Mind Works), yet ours does something different.
Yeah, we can all agree that we are basically brilliant compared to other mammals, but nobody can say what it is. Maybe it's just bias towards valuing the things that we are good at.
When you "simulate" something, you are building something to function in the same manner as something else. You can't "simulate" the brain without knowing what it is the brain is outputting (referred to as the mind here for brevities sake). I can build something that looks very much like a brain, say, a connectionist network. Unfortunately, there are problems with the model that make it a poor model for everything the mind does. Thus, it is a poor model for an actual brain. I can simulate the brain using cheesits and string, but a mind isn't coming out of it.
You probably need to simulate the whole embryo/fetus/baby development. And the uterus, of course. I don't suppose that a human brain will develop a mind if it doesn't get input.
Modeling the mind is critical in order to evaluate whether the models of the brain we are building accurately reflect what is going on insofar as we care about the emergence of a mind. I assure you that Sejnowski does care about the ermergence of mind. His claim would be uninteresting in the extreme if he did not.
One of the problems in cognitive science is finding a model which bridges our understanding of the mind with our understanding of the brain.
If you manage to simulate the mind of a nine month baby then you will know that it has a mind because it's body will be behaving just like a baby. Crying and sleeping. It would be quite striking.
That might be what the Gizmodo article implies, and it seems to be what set Myers' rant off - but nowhere does it quote Kurzweil as actually claiming that.
Ray has suggested in his books that we could simulate a brain, either functionally by reverse-engineering how the brain actually does things (not by studying the DNA), or if that fails by simulating all the various actual neurons sufficiently well and hoping that intelligence emerges (though that would take a few orders of magnitude more computing power). I don't see Myers addressing those claims.
Why would anyone engrave "Elbereth"?
Kurzweil didn't make that ridiculous claim in the first place, despite Myers' third-hand assumptions.
It was just an aside pointing out that the brain's overwhelming complexity all stems from a few million bytes worth of DNA, implying a significant level of replicated structure, andcertainly not a suggestion that we could derive a whole working brain from it.
Why would anyone engrave "Elbereth"?
He's got a hypothesis ("I can extend my own lifespan by aggressively interceding with my body chemistry"), he's performing the experiment on himself, he's making careful observations on a weekly basis and he's taking meticulous notes. Results are probably some time off yet, but even if he turns out to be completely wrong, he's done the experiment in a way that can be reproduced, and it's still a genuine contribution to human knowledge.
How is that not "science"?
Why would anyone engrave "Elbereth"?
If you manage to simulate the mind of a nine month baby then you will know that it has a mind because it's body will be behaving just like a baby. Crying and sleeping. It would be quite striking.
I agree with you. However, if you meant this as an argument, then you actually meant to say "If you manage to simulate the *brain* ..."
Philosophers have argued that if something behaves like a person, we should assign it personhood attributions. This is how many argue for the possible valuing of AI and some animals. You are suggesting the same; Since it behaves like something with a mind, we can assume that it has one.
This argument is problematic. It assumes that we have somehow managed to perfectly simulate human brain function such that it leads to the emergence of a mind. It should be pointed out that, in order to accomplish this feat, *we must know things about the mind*. Successful models of the mind not only supply us with tools to help people with cognitive disorders, they also bring us closer to understanding how the mind actually emerges from the hardware (the brain).
I have to say that while he did an excellent job describing much of the aspects of how we believe the brain works today as opposed to the over-simplified model presented by the original article's author, the mistake made is that people able to understand what he wrote were not the people needing the debunking.
Let's face it, if you're able to read at the level which this guy wrote, then you're probably more than capable of understanding from the get-go that the original article was similar to "By 1975, we'll all be shooting around the skies in flying cars". I would like to see someone with good writing skills... at the popular mechanics level of writing take his response and phrase it in a way which would have an impact on the audience of people he was most likely concerned "would take the bait".
Either way, nice article. I know some about genetics and some about biochemistry, however I lack the ability to judge for myself what is fact and what is credible scientific speculation from his response. Sadly, while I learned a great deal from his response, I lack the knowledge to even know where to pursue this train of though in order to more clearly understand the issue. Either way, thanks for the education and something to think about.
His "critic" used the following steps:
1. take one of Kurzweil's statements out of context
2. build a strawman around it
3. attack the strawman
4. profit!
But to no avail, as the truth remains: even though the brain is remarkably complex, its design is coded in a very compact manner. It's true that to actually execute that compact code on a computer needs a "virtual machine" to simulate biochemistry (whether it's feasible or not), but it's obvious.
I sure would like to see the operating system this "software" will run on, it will have to be close to something like, e.g. reality.
Don't speak about time until you have spoken to him.
> He's looking at the genome, and then saying that you can build a working brain from that info
> alone. It may be theoretically possible, but is so difficult that we shouldn't even bother trying.
Your brain, like the rest of you, was grown from that info alone, so that is definite proof that it is possible not just theoretically.
If we ask Ray what he means by "brain" we may understand better, the part of the brain that is practical to reverse engineer first is logic, or the neocortex. Assuming that Ray means the whole brain is where I think neuro-scientists are correct in their conclusion as of now 10 years is very optimistic. On the other hand if we only consider the logic center and don't deal with vision centers, speech and hearing centers, motor functions and autonomic systems, then we can probably do it. The key problem is input, the brain is nothing without data input and interaction, and experience.
There is more to the human brain then logic, but at this point raw compute power which is approximately 20 Quads per second and developing subroutines to simulate neuron function is what I think Ray is speaking of, simulating sensory stimulus response may be part of that program. The co-processing centers that pre-process the input to signals the brain as whole understands is another problem, and I don't think this is what Ray is predicting, but I think we need to ask him what he means before we jump all over his statement like we know what he's talking about. The brain is an electrochemical streaming massively parallel biological processor, which seems to be very well suited to pattern recognition and linear logic calculations. The brain not well suited to exponential logic, so neuro-scientists are correct from their frame of reference but wrong in the actual application of their logic because it's linear.
Raw compute power and the interactive subroutines which are repeated over and over in the wet wired world of the brain, simulating the plasticity and electronic logic is probable but the chemical side of things is a whole other issue, this would involve understanding hormonal function and interaction at a level not available today. It would require the understanding of calcium, potassium, sodium and other ion channels and other mineral interaction and function in the brain, and it would require the mapping of the two-way interaction between antagonists and receptors, re-uptake of hormones and their effects on signal processing.
Is what Ray Kurzweil says clear? On the surface yes, but the meaning of what he is saying needs to be explained and not so quickly dismissed.
"Is that real poncho or a Sears poncho?" ~~FZ
Comment removed based on user account deletion
Also similar is Asimov's pyschohistory, which suggests that the problem might solved, in aggregate via statistical means.
Interesting ideas. But I think they are all wrong. The problems are combinatorial. Brains don't live in a jar. We're bombarded by information from the outside, and the complexity of these interactions is going to be something that can never be computed. Already at 59! you're at 10^80 combinations and have exceeded the number of protons in the universe. Good luck that that calculation. Ultimately our interactions with others is what makes us human. Put brain in a closed room with no contact and it will never rise above the level of a temperature regulator.