Critique of ‘Against Naive Uploadism’
May 10, 2011 by Randal A. Koene
(credit: iStockPhoto)
In “Against Naive Uploadism: Memory, Consciousness and Synaptic Homeostasis,” neuroscientist Seth Weisberg challenges the comparison of a neuron to a digital computer and the idea that an action potential (spike) fired by one neuron equals one calculation at each synapse. He also challenges the assumption that we are approaching computing power comparable to the human brain.
Overall, Seth Weisberg’s article is refreshing and clearly spoken, in that it lays out the differences between some of what we know is going on in the human brain and what we commonly think of in terms of computer programming.
Those explanations are enlightening, not only to transhumanists with a desire to see the accomplishment of substrate-independent minds, but also to those designing artificial intelligence with the aim to achieve or exceed human mental abilities in silico.
In fact, it is the frequently omitted, rather shallow consideration normally given to human brain processing and the potential importance of details and modular collaboration that prompted the inclusion of a neuro track at the AGI-11 conference in Mountain View this August, a track that I am chairing.
I wish Seth were right, that many indeed see the process of uploading to a substrate-independent existence as an “inevitable” step in human evolution. In reality, there is probably a small minority who even actively contemplate the possibility. The understanding of the mind and human experience as comprising phenomena elicited by purely mechanistic and computable processes is still a marginal one.
And the notion of possible steps beyond current human ability and state of existence exceeds the prevailing perception of the human species as the pinnacle and ultimate goal of evolution and nature. The unguided process of evolution is frequently misunderstood, even by those who accept its reality.
Neural vs. digital processing
Weisberg is correct to dispute the old computing power calculations that compare brain and digital processing. It is incorrect to attempt such comparisons without noting the significant differences in how processing takes place on the two platforms. One can note that in effect, the architecture of a single neuron can carry out processing that requires a large number of operations to model digitally, but at the same time, that it may be necessary to involve a large number of neurons to achieve any robust processing at all. It is difficult to compare the two platforms directly.
A correct comparison of processing power might instead attempt to take into account all of the input information arriving at the human brain, all of the behavioral output being generated, the knowledge being stored, skills learned, and thoughts contemplated, and then to estimate in some manner what processing capacity a digital platform would need to have to accomplish the same.
That is not an easy calculation to make. An alternative is to consider the digital processing power needed to carry out a highly accurate, real-time emulation of all of the biophysical processes of a brain that play a significant role in the processing that is of interest.
I should note though, that Ray Kurzweil has recently been explaining his statements more carefully, as evidenced for example during his presentation at the 2010 Singularity Summit, at which he was surrounded by neuroscientists (e.g., Terry Sejnowski) presenting at the same event.
Building substrate-independent minds
Seth Weisberg is entirely correct to note the many details that need to be taken into account and updated if a computer is to carry out a careful emulation of the processing activity in a brain. Those details are precisely the sort of details that are considered necessary for the conservative implementation of a substrate-independent mind (SIM), which we call whole brain emulation (WBE). It is possible that one may in future be able to upload to alternative implementations, but a whole brain emulation is presently the implementation that allows us to most confidently carry out research and development toward uploading and SIM.
I think it is important to read Seth Weisberg’s conclusions very carefully. For example, Seth states that we must understand which biophysical processes are “correlated with, lead to or actually are cognition, memory and consciousness” in order to make uploading to a SIM possible. I agree that we need to know which ones are involved. But I do not think that this equates to needing to have a full understanding of how those processes achieve our phenomenological experience, any more than it was necessary to fully understand aerodynamics to build the first airplanes.
If we are conservative about our decisions, and we err on the side of caution in the inclusion of biophysical processes, then it is possible to emulate something with partial understanding. There are probabilistic (Bayesian) methods by which one can construct successive models and can calculate whether the addition of a specific detail continues to improve the system’s overall ability to compute recorded activity and behavior or not.
Those methods are concrete examples of ways in which emulation to a specified precision may be accomplished without necessitating a full abstract understanding of all the actions and interactions modeled.
Comments (8)
by AeaeaActual
@Khannea: In response to your three perceived problems:
1. The technologies will not arrive in peractio. In peraction arguments are almost always wrong when it comes to technology, and always wrong when it comes to complex technology. These technologies will, and already are, coming bit by bit, met by ever larger and more distributed demand.
2. See response number one.
3. See response number one. Your shark outcome is, from both a scientific and economic outlook, vastly less likely than a distributed emergence. Of course, you could be right – I’m all for responsible and well-rounded fear-mongering – but I’d rather be an optimist looking at the statistically more likely future than a pessimist shouting that the sky is falling.
You and Hugo de Garis are likely in for some pleasant surprises in the long term. Hang in there.
Ta.
@AeaeaActual
by Brian H
Gradual replacement is persuasive. But, once 100% is achieved, you’re in the world of hardware and software, and THOSE are readily replicated. So you can now be “cloned” readily, innumerable times. Which is the “original”? All would subjectively believe themselves to be. All would be wrong.
by Logic
I always find this this kind of thinking to be so limited in its perceptiveness, both the original article and the Suntzu commenter. Why must everyone persist in the naive notion that the leap will be a moment we face? Few today, as this article’s author suggests contemplate uploading. But as I write this it’s 2011. Technology in its infancy. Only those who imagine, extrapolate and predict would be thinking in these terms. But this (like the singularity) is not a moment that will suddenly happen. We won’t wake up one morning to the news that uploading is now available. It happens gradually. Slowly. Mistakes and inefficient technologies at first, during which people are introduced (and likely summarily reject) the concept. Then as the technology evolves and matures, so does public thinking.
Yes, it is in fact inevitable.
And let’s dismiss the alternative conspiracy theorist notion that it will all be shrouded in corporate culture, corruption and capitalism, too. To imagine these scenarios is to accept the same flawed thinking that imagines that the technologies will appear overnight. We are quickly migrating away from the ominous social aspects that fuel both fears, and that transition will become increasingly rapid. The “democratization” of these technologies will happen as swiftly as the technologies themselves appear, as is the case with almost all new technologies.
I was watching an interview with Kurzweil where he describes gradually adding technologies to the brain, such that, say, 25% of the brain was now technologically augmented. Are you still you? Yes. Then it moves to 50%. Still you? Yes. 75%. Still you? Yes. At some point if you switch to 100% (i.e., “uploaded”) you would still be you without ever really knowing when the transition occurred. This is why the idea of man blending with machine is so profound.
Yes, it’s inevitable, folks, and it’s a good thing.
by xxdanbrowne
I don’t disagree but I’d like to clarify your point because you are NOT spelling out one of the assumptions.
Yes it’s true that with a gradual augmentation you are still you all the way up to the 100% mark.
That’s fundamentally different, however than going straight from 0% augmentation to a copy of you in some other substrate.
Here’s why: as you gradually augment, you are effectively a hybrid orgaqnism. There’s only one organism, part of the processing is done in the meat brain the rest in the synthetic part. Adding further synthetic parts allows greater processing but each time creates a new SINGLE organism that is an upgrade of the previous one. It’s analogous to adding a new graphics card to your old PC.
Conversely, you are sitting in a chair and you are non-destructively scanned and a copy of you wakes up in a synthetic substrate. Yes it’s a COPY of you but it’s not you.
The only way to get around that would be by scanning your brain, uploading the copy and at the same time putting some kind of wifi link and a backfeed mechanism in so that as the synthetic version of you thinks, what it thinks is fed back to you, so that you are in effect a hybrid.
Just doing an upload, isn’t you however, it’s a synthetic twin or clone.
An analogy of this might be backing up all the data on your PC onto a virtual image and booting that image up on a virtual machine on a Mac. Yes it’s an identical copy of the original data but it’s NOT the original PC. If both machines were *NETWORKED* however….
So to conclude, don’t delude yourselves into thinking that you’ll get immortality by uploading. You’ll still be dead if you do it that way. The only way to get a valid continuation is by gradual augmentation and upgrading and I personally look forward to that. A straight upload is a poor substitute to upgrading and I personally wouldn’t pay for one to be done.
by examachine
Hi Randal,
How are you doing? I hope you are making some progress at your lab.
I don’t understand at all what Seth means by the assumption that 1 spike equals 1 operation. Surely, spike trains are not that simple as Bialek’s work shows. However, there have been excellent attempts to characterize the computation performed by a single neuron, and deciphering the neural code. Which is nowhere near as simple as representing each spike by an operation, since individual spikes are not independent. We have known this since 80′s. However, I think it is obvious that Dr. Weisberg has no clue about the basics of computational neuroscience therefore I couldn’t read past the first paragraph!!! I don’t think it’s an article to be taken seriously.
by platon
Maybe reading past the first paragraph would be wise before criticizing? Dr. Koene and I both feel the points made are important. The article does not assume that one spike is one computation, it actually is arguing against that – those are the assumptions attacked in the article.
by Khannea Suntzu
2042
by Khannea Suntzu
The whole idea of uploading poses three massive problems.
(first) it is demand-oriented. At some time this century someone somewhere with sufficient resources will get it in her or his head to want it, and demand it. This is just a matter of time before demand meets supply.
(second) the second problem is that the party making the demand may be so dead set on getting what she or he has in mind (not dying by and large) that this person will take anything. In essence The research may be done outside legal oversight. It may be done in relative privacy. It may be done by immoral or ruthless parties. I could name a few non-democratic countries. I could name a category of very rich, much unaccountable, ostentatiously rich people. The same people who have already been selected to not care much about ‘lesser humans’, but who have been groomed by a culture of corporate Darwinism.
(three) A human neurology is something that operates naturally in extremely narrow and arguably fragile constraints. The tiniest change in human brains induces pathology, in either collapse of thinking or at best insanity. One doesn’t even need to resort to actual damage to resort to making a brain experience failure states – just a tiny bit of stress will reduce a brain to extreme states of despair or antisocial tendencies. Taking the human neurological architecture and then being able to successfully port this to a substrate that works far exceeds anything any lab of scientists can model, anticipate the results of, extrapolate or envision, especially if the person paying for the project exerts power over them. If the procedure involves making a transitional copy (the client is dying and the moment of death is engineered to coincide with transference) then the scientists will at some point me relinquishing control of the new substrate to a emerging mind that has created all conditions to exert power over itself. If they implemented sufficiently robust diagnosing devices to keep the synthetic cortex from collapsing, or going effectively inssane, then the vector will be aimed straight towards recursive self-improvement.
In other words, the moment some super-rich client somewhere this century gets it in his or head to compel a team of neuro-scientists and AI-scientists to transition her or his brain in such a substrate, they will almost certainly be bootstrapping the emergence of an artillect.
And in essence an artillect groomed by years of corporate culture, competition, hostile take-overs, capitalism, globalism, finances, speculation and corruption.
It’ll be like putting a shark in charge.