Goertzel Contra Dvorsky on Mind Uploading

· April 21, 2013

Futurist pundit George Dvorsky recently posted an article on io9, labeled as “DEBUNKERY” and aimed at the topic of mind uploading.  According to the good Mr. Dvorsky, “You’ll Probably Never Upload Your Mind into a Computer.”   He briefly lists eight reasons why, in his view, mind uploading will likely never happen.

UPDATE – here is a video interview on this subject:

Note that he’s not merely arguing that mind uploading may come too late for you and me to take advantage of it – he’s arguing that it probably will never happen at all!

The topic of Dvorsky’s skeptical screed is dear to my heart and mind.  As a person frustrated with the limitations of the human body and interested in unlimited lifespan, I find the notion of mind uploading highly appealing.   While my main research area is artificial general intelligence, I’ve also dug fairly deep into the science underlying mind uploading … in 2012 I edited the first journal issue entirely devoted to Mind Uploading (an issue of the Journal of Machine Consciousness).

Every one of Dvorsky’s objections has been aired many times before – which is fine, as his post is a journalistic article, not an original scientific or philosophic work, so it doesn’t necessarily have to break new ground.   More problematic is that the arguments he presents for the objections are rather weak.

In this article I will briefly run through Dvorsky’s eight objections, and give my own, in some cases idiosyncratic, take on each of them.   I will explain why, from my own perspective, none of them is nearly as compelling an objection as he suggests.   Of course, at this stage, we cannot know FOR SURE that mind uploading is a real possibility.  But in my view, the balance of the evidence is strongly in favor of mind uploading’s feasibility.

Whether mind uploading will be developed in time for you and me to enjoy its benefits is quite another question, of course … identifying what is possible is much easier than predicting what will happen when.  But, the timing of what happens when, depends substantially on our attitude and initiative.  If we believe, pace Dvorsky, that mind uploading is probably a pipe dream, then work toward mind uploading is not likely to advance very fast.  If we believe mind uploading is a highly feasible technology and it’s likely just a question of when, then we’re likely to progress toward mind uploading a lot faster.

Reader be aware: My take on mind uploading, and Dvorsky’s objections, may not reflect that of the prototypical mind uploading researcher, or Singularitarian, or whatever.  Like Dvorsky (whom I count as a friend and whose work I respect), I have my own particular point of view.   For more depth on my general perspective on mind, brain, consciousness and reality from a futurist perspective, see my brief book A Cosmist Manifesto.

I use the term “mind uploading” here because Dvorsky does, but actually I prefer Randal Koene’s phrase “substrate independent minds.”   Koene’s term is less sexy than “mind uploading” but also probably less misleading.  The point is the liberation of the mind from any particular substrate.   Once one’s mind is liberated from a particular substrate, then, at that point, one can choose what kind of embodiment one wants.  Some will want human bodies, some monkeys, some cute fluffy bunnies, some flying space robots, some virtual-world avatars — and some will want things we cannot now imagine….  Each chosen embodiment will influence the nature of the mind instantiated in it, often quite dramatically, and that will be part of the wonder!

 And so, without further ado, Goertzel contra Dvorsky

Dvorsky’s Complaint #1: Brain functions are not computable

“Is” and “are” are complicated words, semantically….  Computability, in the sense it’s used in Dvorsky’s complaint, is a mathematical tool used for modeling certain systems.   So his claim would be “brain functions cannot be effectively modeled using computable models.”

This is an interesting hypothesis, but there is certainly no scientific evidence for it.  Furthermore, within the confines of current science, there is no possible way to gather solid scientific evidence for it.  The problem is that all scientific data ever gathered, constitutes one large but finite set of bits (i.e a finite set of finite-precision numbers).   Any finite set of bits can be modeled computationally.   Of course, someone can claim a non-computable model is “better” than any computational model, for a given finite set of bits.   But this then becomes a subjective claim, based on aesthetics, or intuition.

Perhaps some future discipline, going beyond the bounds of science as we know it today, will formulate some new sense in which brains fundamentally cannot be computationally modeled.  But this vague possibility seems a rather threadbare excuse for rejecting mind uploading.

Dvorsky also conveniently ignores the possibility of analog and quantum computing.  If it turns out the brain cannot be modeled using conventional computers, for some reason outside the scope of contemporary science — or if modeling using conventional computers proves overly difficult even though possible — then maybe the brain will still be effectively modelable using analog or quantum computers.  In that case, maybe we could upload into those.  I don’t really care if my substrate is digital, analog, quantum or twistor based, so long as it works!

Dvorsky’s Complaint #2: We’ll never solve the hard problem of consciousness

 I don’t know if we will ever solve the “hard problem of consciousness”, i.e. the problem of rigorously connecting subjective experiences with physical structures and dynamics.   My gut feel is that this will require the invention of some new discipline going beyond contemporary science, synthesizing aspects of empirical/theoretical science with aspects of spiritual traditions and contemplative metaphysics like that of the medieval Buddhist logicians.  I gave a talk on this at the 2011 Asia Consciousness Festival.

But, whatever….   An artist doesn’t need to solve the problem of defining beauty to make a beautiful painting.

What happens when I upload my  mind into a robot body, and then George Dvorsky argues with the robot Ben Goertzel, saying “Robot Ben, you’re not really conscious, because you can’t clearly demonstrate to me philosophically how your consciousness is rooted in your circuitry.”

What happens when Robot Ben replies to Meat George: “OK, my meaty friend, then perhaps you’re not really conscious either. because you can’t clearly demonstrate to me philosophically how your consciousness is rooted in your neural wetware.”

If life were restricted to that which can be rigorously philosophically justified, to the agreement of the majority of philosophers, not much would happen.

Dvorsky’s Complaint #3: We’ll never solve the binding problem

The “binding problem” has to do with how percepts and concepts are represented in the brain, in such a way as to have a unified feeling and dynamic in spite of their involving physically disparate parts of the brain.

Why Dvorsky thinks we’ll never solve this, I really don’t know.   It’s a non-trivial science problem, but seems likely to be straightforwardly resolvable once we have more spatiotemporally accurate brain imaging hardware.   Of course, I can’t PROVE that, but, really….   Yeah, it’s possible that interpreting future highly detailed brain imaging data will prove somehow beyond human capability, and neuroscience will grind to a standstill.  But… pshaw….

Dvorsky’s Complaint #4: Panpsychism is true

This is a weird one.   I personally believe panpsychism is true – I think everything in the universe has its own little spark of consciousness.   Why this would be taken to mean we can’t upload our minds into computers, I have no idea.   If everything in the universe has its own spark of consciousness, then so do brains, and so do computers.  So what?

The tricky question becomes figuring out what kinds of consciousness correlate with what kinds of physical systems.  What properties of the human brain need to be emulated in another substrate, in order to cause that substrate to correlate with consciousness that is human-like, instead of having some other flavor.

Why Dvorsky considers this problem unsolvable, I don’t know.  Maybe he just got up on the wrong side of bed and fell out of his treehouse?

Dvorsky’s Complaint #5: Mind-body dualism is true

Philosopher Galen Strawson has argued pretty persuasively that dualism is logically incoherent.

But giving Dvorsky probably more benefit of the doubt than he deserves here – OK, what if Strawson and common sense are wrong, and mind-body dualism is true?

Let’s suppose that mind contains some aspect that is separate from physical reality, yet connected to it.  So that the human brain is, in some regards, an “antenna receiving the immaterial mind”, rather than something that produces the mind.  (Though please note, it must be an awfully complex antenna, given all that’s known about the specific cognitive and experiential consequences of messing with particular parts of the brain.)

In this case, if we upload a human brain into a different substrate, why wouldn’t the new version of the brain act as an antenna for the immaterial portion of the mind equally well?

Of course this is not guaranteed, since IF mind-body dualism is somehow true, we have no idea how it would work.  It’s possible to make up various dualist theories under which mind uploading is impossible.  The Christian Church has made up one, for example.   But why should we believe these theories, instead of looking in the direction neuroscience is pointing us?

Dvorsky’s Complaint #6: It would be unethical to develop

Some consider it unethical to develop mind uploading.   Many of us consider it unethical not to – because we feel that extending life and mind throughout time and space is an ethical imperative.

Some consider it unethical to use electrical power, as well.  So what?  That doesn’t stop the rest of us….

Dvorsky’s Complaint #7: We can never be sure it works

As an old man of 46 years old, one thing I have realized more and more in the course of my life is: We can never be sure of anything.

Even death and taxes are not really inevitable, as we now have a society of immortality and abundance within reach.

We may well be living within some sort of computer simulation.  How can we know?

I certainly can’t know that George Dvorsky is really conscious, nor that the world outside the house I’m now sitting in hasn’t recently been replaced by some sort of Truman Show.

So what?

While Meat George is pointing at him saying “I can’t be sure you’re conscious”, Robot Ben will be flying around in space, solving eight billion dimensional nonlinear equations, communing with the trans-galactic superorganism and having sixty seven orgasms at once, cackling “That, my friend, is because you’re limited to a legacy human brain!”

Dvorsky’s Complaint #8: Uploading would be susceptible to hacking and abuse

True enough.

This reminds me of someone, back in 1990, saying that the Web will never be created because it would be susceptible to hacking and abuse.

Well, yeah.  Hacking and abuse happen.  Defense mechanisms are created.  The struggle between organisms continues.  Life goes on.

And, if we’re reasonably lucky, life will go and and on – beyond the realm of meat humans with painfully short lifespans, into the domain of uploads and AGIs exploring aspects of the multi-multi-…-multiverse that meat humans can’t even imagine.

In summary…

Can we prove for sure that mind uploading is possible?   Of course not.

But, red herrings about ethics and hacking aside, the arguments Dvorsky raises against mind uploading basically boil down to:  Many people are confused about the philosophy of mind, and they argue about it a lot….

Many people are confused about the philosophy of matter and the nature of space and time also, but that doesn’t stop us from manipulating matter in pretty cool ways.   Actually, time is just as confusing to philosophers as consciousness — yet few go around claiming that we can’t build machines operating in time until we have an adequate, detailed, broadly accepted philosophy of time!

Mind uploading is coming.  I hope it’s coming in time for us.  Time — even though we don’t understand it that well yet — will nevertheless tell…

For more info on the growing science of mind uploading, see Randal Koene’s excellent Carbon Copies website.

 – This article brought to you by Meat Ben.  Enjoy him while he lasts! –



For more videos of lectures and interviews with thought leaders please
Subscribe to Adam Ford’s YouTube Channel


You may also like...

14 Responses

  1. Sally Morem says:

    On the biological versus machine controversy. Sounds like someone is not taking accelerating tech and its implications very seriously.

    What if accelerating tech makes it impossible to tell the difference between biological brains that had been improved and computers that had been improved? Exactly what would be uploaded from what and why would anyone care enough to start a jihad over the question?

    • Singularity Utopia says:

      This is my point, there is in essence NO difference between machines and biology. The human brain is a machine, we are already in essence uploaded, thus the uploading issue will will soon be irrelevant when people realise we are already uploaded, that there is no difference between biology and machines. Uploaders are merely engaging in escapism, they suffer IMO from a grass is greener on the other side malady.

  2. I totally agree with Ben. Some comments on George’s article, posted also at io9:

    George, this is an interesting and well-written article, but I think all your points are rather weak.

    Quick comments:

    1: Every real physical phenomenon is the result of result of unpredictable, nonlinear interactions among billions of components, but we can replicate many real physical phenomena in silicon well enough. The mind is more complex than the physical phenomena that we are able to replicate in silicon at this moment, but science will advance.

    2 and 3: I am not really able to understand why these are considered as “problems,” it seems to me that they have an ontological status similar to the “aether” of 19th century’s physics (a non-problem caused by non-physical non-notions that were not really needed by science, and have faded out since). I guess all metaphysical “problems” like these will disappear like snow in the sun once uploading becomes a commonplace practice.

    4 and 5: As you say, perhaps consciousness is not substrate-neutral, but rather depends on specific physical features that are only found in some substrates, and not in others. Fair enough, but I don’t see why this is a problem. You seem to assume that uploading means uploading to something similar to today’s computers, perhaps powered by Intel 20ium and Nvidia JupiterForce, and running Windows 30. But this is a restrictive and unnecessary assumption. If a conscious mind can run only on a substrate with certain specific properties, well, then we will just have to use substrates with the same specific properties to upload minds. For example, if consciousness depends critically on subtle quantum effects in our neural circuitry, we will have to develop alternative substrates that exhibit the same quantum properties.

    6: I am a living, rights-bearing human being, and I would love having the chance of being the first human to be uploaded – please don’t protect me from things that I don’t want to be protected from.

    7: Re David’s quote – If what I want to do is replicate the chess game, the textures of the pieces are irrelevant. The game is the same with any other textures, for all practical purposes (FAPP) related to my objective of replicating the gameplay. Of course, the moves are not sufficient if the objective is the artistic appreciation of a fine chessboard and its pieces. As far as identity preservation is concerned, my choice (it is basically an aesthetic choice, you cannot prove one position or another) is to consider the computational pattern good enough FAPP. The texture of my skin, I can do without if I have to. I don’t actually like it that much (more in an exchange with David at http://ieet.org/index.php/IEET…)

    8: The same has been said of the Internet and mobile phones, thousands of times. I am still using both, and so are you. Once mind uploading is developed, I am sure savvy users will develop mental crypto tools, the mental equivalents of PGP and disk encryption. Also, I prefer an interesting life with some risks to a dull and boring life without risk.

  3. Jack says:

    The philosophy of mind objections are a distraction. Once the input/output function is roughly equivalent to a human brain receiving the same input, the algorithms will act and communicate as if they were humans.

    This will suffice to convince society to treat them as real people, not just philosophically, but out of pragmatic necessity – you cannot ignore thousands or millions of entities who insist they are conscious persons.

    Here are some more probable objections:

    1) The problem may be too hard to solve economically. Maybe civilization is disrupted and can’t achieve the progress needed. Maybe Moore’s Law falls short. Maybe scanning technology is never precise enough. Maybe the implementation complexity is just too much to handle.

    2) Generic AGI comes first.

    3) It could be banned globally. The ban could be enforced by surveillance technology. (I don’t think this is very plausible)

    If mind uploading happens, it will not consitute the utopia some people seem to assume. Remember the 99%? Well, in a future with mind uploading, most resources will eventually be owned by copies of a small number of rich and resourceful individuals. And if they are sadistic, they may well use their resources on torturing copies of their enemies. If they allow you to exist at all, they may make you kiss their digital feet all day.

  4. George says:

    I posted this comment on facebook but I’ll post it here again.
    I’m inclined towards Dvorsky’s position on this. The most compelling arguments being those of John Searle: the chinese room argument, and especially that computation is ascribed to physical systems, not inherent to them. (see- https://mywebspace.wisc.edu/lshapiro/web/Phil554_files/SEARLE-BDC.HTM) Also, if consciousness is purely a result of causal computational structure then we’d have to accept multiple realizability. On that account, an (analog) computational implementation of my consciousness could be constructed out of 100 billion very attentive greyhounds acting as neurons in my neural network who stick their tongues out to represent neural excitation and vomit to represent the neuron firing. Or, as in Ned Block’s example, out of ‘cats, mice and cheese’. Or glaciers, or whatever you like. Intuitively (to me at least) that makes it untenable that computation is sufficient for consciousness. That doesn’t mean we couldn’t upload the structure of our minds to computer programs that could then behave exactly like human beings ! Which is scary as hell – inert computations pumping out human-like behaviour could make us very casual about dispensing with our physical bodies when actual conscious experience (the only thing with any value at all) is not being preserved at all.

  5. Randal Koene says:

    Thank you, Ben, for this response! :-) It is exactly the response I wished for George’s article.

    As an aside, to the comment about bio vs machines: I don’t know who you’ve been talking to but the people I know involved with whole brain emulation are not waging any such controversy (see for example George Church’s talk included in the gf2045.com line-up).

    • Singularity Utopia says:

      Randal. The controversy, or the two camps, regarding immortality, is generally the so-called “biological” versus the so-called “machine” future. Of course there will be a few people in between but generally there is a clear split.

      I surprised you have not heard about the controversy, it is reasonably common in futurist circles. The viewpoint of Dvorsky is not new, I commonly encounter his view, similar to how I also encounter the converse view where people are utterly insistent the human body (DNA) is obsolete thus in the future anyone who wants to will exist entirely in “silicon.”

      Charlie Stross wrote, in 2011, “Three arguments against the singularity,” in which he stated his mild scepticism about uploading: “We may eventually see mind uploading, but there’ll be a holy war to end holy wars before it becomes widespread: it will literally overturn religions.” Cory Doctorow, more skeptical, has compared mind-uploading to Thigh-Masters (LOL), thus he stated regarding the option of people leaving their bodies behind: “The question of whether such an option will likely be available to us is something I’m not at all certain about, but in the presence of such an option, I’m very confident that large numbers of people would opt for it. We like get-evolved-quick schemes. If you can sell Thighmasters, you can sell mind uploading.”

      People often wrongly think the Singularity is mind-uploading.

      PZ Myers often criticises the uploading viewpoint and many others think uploading is impossible. It is a common theme of controversy I encounter reasonably often in comments.

  6. Singularity Utopia says:

    I am an uploading skeptic but not because I think it is impossible. I think it is merely a 1980′s type of clunking metallic cyborg fashion statement, which will be hopelessly outdated soon. Uploading seems to be based in a bio-phobic mentality, almost misanthropy. Amusingly I often mention DNA-computing (or other bio-computing) to uploaders, I state I want to upload to my own DNA, which invalidates their escapist bio-phobic uploading purpose. Human bodies are currently limited but biology can easily be improved.

    The whole machines versus biology thing is extremely misguided because biological organisms are machines albeit with different parts. So-called “machines” (artificially created ones) of the future, they will not be clunking metallic things, they will be indistinguishable from any natural living being because natural living beings are actually machines.

    • bengoertzel says:

      Hmm… my own interest in uploading does not result from any sort of bio-phobia…. Randal Koene’s phrase “substrate independent minds”, while not as wizzy as “mind uploading” probably captures the notion better. I just want to liberate my mind from any particular physical substrate. So that it can live in a DNA-based bio lifeform, a computer, a nano or femto tech post-Singularity construct, or whatever…. Of course each of these vehicles will nudge the mind in a different direction via body-mind interdefinition/interdependence, and that will be funky to experience ;-)

      • Singularity Utopia says:

        Hi Ben, the bio-phobic notion is a generalisation regarding a desire to upload or “liberate” yourself from a “particular physical substrate,” thus perhaps not applicable to everyone, although I do think all uploaders (substrate liberators) suffer from an irrational need to escape, in varying degrees.

        Discarding the human body is very escapist. I suspect it’s a phenomenon of thinking the grass in greener on the other side. People talk about uploading their minds to computers or machines but they fail to realise our brains are already machines, our brains are computers albeit with a few kinks due to the randomness of unguided (not deliberate) evolution.

        The escapism of uploading from one computer (the brain) to another computer (computation not based on DNA) means the escapism is an illusion identical to all forms of escapism. This means nothing will really change for uploaders no matter where they put their minds. I think expanding intelligence will soon allow people to realise this therefore the uploading meme will become obsolete, it will be irrelevant, in the not too distant future.

        It is wrong to assume the only way to become immortal is to upload your mind to a “computer.” We can easily see immortality will be achieved with not too much effort via regenerative medicine, gene-therapy, general upgrades-patches to DNA, and if you are overly cautious a nano-shield (or femto-shield) to protect you from any accidents.

        The usage of the word “liberate” is illuminating regarding a psychological bias of feeling trapped. Yes I do realise we are all somewhat trapped by the limitations of being human but uploaders seem to have a disproportionate reaction, an overreaction, to the limits of the human machine. It reminds me of the recent news report about a Russian who killed six people merely because someone scratched his car, allegedly. Analogously our human limitations are perhaps more akin to driving on a flat tyre but my point is valid, namely people are making a mountain out of a molehill, they are having an extreme reaction to a relatively minor problem thus they want to throw everything away. The need for uploaders to escape the limits of the human machine is actually about wanting to escape their minds, thus instead of repairing the scratch or the flat tyre, or upgrading the car with better parts, they want to scrap the entire car. Unfortunately for them they will find their problems will come with them to their new vehicle because, limitations aside, the real problem is the driver.

        Thankfully due to expanding intelligence we will never arrive at a situation where uploading actually becomes popular.

        • ecology19 says:

          There may be problems with the new car, but the new car will be much better than the old.

          • Singularity Utopia says:

            Unless the new car is Windows 8.

            Perhaps you are generally right “ecology19″ and a new car may be better IF your “old” car is sufficiently old BUT consider the cell regeneration of human bodies. When biological technology progresses to the point where all cells can easily be repaired would it be easier for a 60-year-old to have their body regenerated in situ, age-rewind to 18, or would it be more efficient to regrow an entire new body then transplant the old mind into the new body? If a car or body can be regenerated to factory condition, via nanobots, combined with modernisation upgrades, in situ, then I think this would be the most efficient method, more efficient and sensible than uploading to a new car or new body. So there would be no need to buy a new car, your old car could become a transformer seamlessly morphing into a new pristine car without actually scraping the old car.

            The only circumstance where an entirely new car would be applicable is regarding head only cryo-preservation.

  1. April 24, 2013

    [...] Ben Goertzel is getting to the vast unknown of infinite consciousness in his recent post on Mind Uploading: [...]

  2. April 25, 2013

    [...] These questions have been hotly debated of late with prominent figures taking a stand but against and for. [...]

Share Your Thoughts