Posted: Thu, April 11, 2013 | By: Indefinite Life Extension
by Franco Cortese
This essay, originally published on Transhumanity, is in response to the Debate Forum “Will ‘meatbag’ bodies ever be immortal? Is ‘cyborgification’ the only logical path?”, hosted by Immortal Life. In it, I explore the distinction between therapies for Longer Life (life-extension) and therapies for Unlimited Life (indefinite longevity). It was adapted from responses given in the Debate Forum, and so makes several related points rather than one elaborated thesis.
If we’re talking far-future, non-biological approaches to life-extension will win out over biological approaches, due mainly to their comparative advantages (e.g. ease of repair and modification - both of which become methodological problems now rather than technological or “physical” problems so to speak, requiring a reorganization or rewriting of information - thus methodological - rather than a means of figuring out what changes to implement and then devising a technology and technique for actually implementing such changes in physicality) and because they will offer experiential and functional modalities categorically unavailable to biological systems (merely due to the fact that such functional and/or experiential modalities are determined by the structural and operational/procedural modalities of the system, and there are a much larger quantity of potential structures and operational/procedural modalities possible using non-biological systems than are available to what are normatively considered biological systems - that is, based upon cellular units forming emergent tissues, and embodying the structural and operational/procedural modalities of biological systems). That being said, I think that the distinction between non-biological and biological systems (especially if Drexlerian nanotech – that is, using mechanosynthesis or precise molecular assembly – is implemented with any ubiquity) will increasingly dissolve. If a system exhibits the structural, functional and operational modalities of a biological cell, tissue, organ or organism, yet consists of wholly inorganic materials, is it not closer to a biological system than to what we would typically consider a non-biological system? Either the distinction between the two will eventually dissolve, or we will use the term “biological” to designate systems exhibiting the structural, functional, and/or operational modalities of biological systems, rather than designating systems made of specific types of material, such as organic or inorganic molecular substrate.
In the 9th installment of my 10-part introductory essay written for Transhumanity’s Certificate in Transhumanism Studies Program, I make a distinction between life-extension therapies and indefinite-longevity therapies, and I’d like to elaborate more on this distinction here. Life-extension therapies extend longevity, but for various reasons fail to make it necessarily indefinite or unlimited. Often this is because such therapies aren’t comprehensive – a given therapy solves one contributing factor of aging, but not all of them. Others, like SENS (which I’m in no way discounting), fix the major causes of damage, but use a different methodology for each respective source of damage or aging; the drawback of this approach is that if previously overshadowed causes of aging now begin to make a non-negligible impact on aging, in the absence of the more predominant causes, then we have no methodology to combat it. Because each strategy is tied intimately to the cause it seeks to ameliorate, the techniques often cannot be applied to the new source of molecular damage.
Indefinite or unlimited longevity therapies, on the other hand, use one comprehensive approach to mitigate all sources of aging. One example is Drexlerian nanotech (and to a shared but somewhat lesser extent Robert Freitas’s nanomedicine - only because it has specifically-tailored strategies not dependent on the feasibility of Drexlerian molecular assembly or “mechanosynthesis”, in addition to the more comprehensive ones). This approach fixes not the source of the damage but the damage itself, iteratively, and can thus be used to combat any source of molecular damage using the same tools, technologies and techniques. With such therapies we wouldn’t need to come up with a second wave of strategies to combat those sources of aging that might crop up in the future, and which remained unnoticed until such a time only because their impact couldn’t be seen (or allowed to take effect) while the first wave of sources was still predominant.
Due to the title of the debate forum this article is in response to, I’d like to explicate the fact that while what is typically called “cyborgification” does (usually) qualify as an approach to unlimited longevity rather than life extension, unlimited longevity therapies are by no means limited to non-biological therapies. Another potential type of biological approach to unlimited longevity (as opposed to life-extension) is certain varieties of genetic engineering, such as the use of viral gene vectors to iteratively replace the genetic material in a given cell with a copy of its own genetic material. The genetic material would be iteratively replaced with a copy of itself, before it has a chance to degrade to the point of functional or operational distortion and degradation. This would work a manner highly similar to the latter Drexlerian approach outlined above, wherein the genetic material is iteratively maintained and repaired, thus vicariously maintaining the cellular components by maintaining the functional and operational stability of its protein transcription and synthesis systems.
Another potential means of indefinite longevity is “Mind-Uploading”, more recently termed “Substrate-Independent-Minds”, and which I’ve referred to as “(gradual) substrate replacement” to signify a broader designation that includes both computational (i.e. mind-uploading proper) approaches as well as physically-embodied (e.g. via electromechanical or molecular systems to gradually replace our current biological substrates) approaches. In the computational approach, all relevant physical processes become informational. Thus replacing potentially-destructive processes with non-destructive analogues, or augmenting such processes so as to mitigate their problematizing aspects, becomes a methodological rather than technological problem. It entails now a reorganization or rewriting of information, rather than a specific technological system or process to enact physical changes meant to negate or replace destructive processes. This applies both to sources of damage heretofore unnoticed, as well as to completely new sources of damage. However, such informational processes are necessarily instantiated by physical hardware, which lacks this advantage (namely, of potential sources of damage becoming a methodological rather than technological problem), and so sources of damage to such physical hardware still necessitate physical and technological systems for their negation/obviation. This state-of-affairs is still more hopeful than the anticipated requirements for negating biological sources of structural damage and procedural distortion, which necessitate a comparatively larger and more capable technological infrastructure (i.e. the technologies and systems required to implement the repairs required to ameliorate such damage are greater). The organizational architecture (i.e. the relative location and connection of components relative to eachother) is in such non-biological cases completely known to us, having been designed, whereas in biological systems there are multiple sources of ambiguity because it wasn’t designed and so isn’t already known – a situation only aggravated by the current lack of precise, non-damaging and non-distortional in-vivo measuring/data-acquisition techniques, which necessitates that our ability to explicate those inner-workings as comprehensively as the inner-workings of the non-biological hardware is limited.
Moreover, non-biological physical hardware (as used in the “cyborgification” approach) would already be designed so as to be iteratively replaced. Thus it would already have systems and procedures in place for detaching a given component, exporting it out of the system, and subsequently importing and installing a corresponding replacement component; such systems would already be in place for us to utilize for the purpose of iterative replacement so as to mitigate functional degradation due to structural and/or procedural damage or distortion. Rather than constructing a new system and importing it into an existing biological system that wasn’t designed with such extra space for external technological systems in mind, or with the optimal conditions (e.g. temperature, pressure, etc.) of those external systems in mind (which increases the requirements of such systems, which now need to navigate inside a biological environment while neither damaging them nor distorting their operation beyond the tolerance ranges allowed by a given emergent functionality, or in order words the operational variation allowed for a given functionality to be maintained, in an environment that isn’t necessarily optimal for such nonbiological systems’ own functionality). In the case of non-biological hardware, such extra space and optimal conditions are already present in the system. This means that in regard to non-biological approaches, physical sources of damage are, for these two reasons, more amenable to maintenance, repair and/or replacement, to the amelioration of physical sources of damage and to workable solutions than biological systems are.
Thus while approaches to indefinite longevity often need to satisfy more requirements (as they need to embody more functional modalities) and thus necessitate a larger technological and methodological infrastructure for their successful development, verification and implementation than do approaches to life-extension (due mainly to being more comprehensive and thereby needing more capabilities) and thus would presumably cause us to lose more lives in the interim (at the cost of 100,000 per day), their greater comprehensiveness means that if a second wave of unacknowledged sources of aging (i.e. molecular damage) does come up, we may save the lives that otherwise would have been lost while we waited for the development, verification and implementation of that second wave of life-extension therapies – which may be necessary if we allocate our focus more toward techniques of life extension rather than unlimited longevity.
On the other hand, certain means of both life-extension and indefinite longevity (like Drexlerian nanotech) have an advantage over certain approaches to indefinite-longevity that replace biological components with non-biological analogues, like Mind-Uploading or gradual substrate replacement more generally, because they don’t enter into concerns of subjective-continuity. The lack of a quantitative measure for identifying “subjectivity” or a subjective sense of self means that a gradual-substrate replacement procedure will always incur a chance of replicating the functional modalities of subjectivity (i.e. all the outward, empirically-verifiable indicators of subjectivity) while failing to replicate subjectivity in actuality, because we have no way of identifying which attributes (if any) facilitate or instantiate (and thus are required for) subjectivity, and thus no way of verifying if non-biological system incorporate or exhibit such attributes.
In the 4th installment of my introductory 10-part essay, I describe what evidence we have to support the claim that replicating all functional modalities of the brain would in some cases fail to maintain subjective continuity. An explicit case would be if we emulated a given brain, ran it’s processing for a given interval of time, and then stopped, going back to disconnect and rearrange all the constituent procedural-steps in an alternate order than the order in which they originally appeared, as in Greg Egan’s Dust Theory, we could replicate all the outward effects of subjectivity, but we would have in effect disassembled and reassembled you every time a procedural step or state-transition was made. While this hypothetical case we knowingly, explicitly attempted to disrupt subjectivity (by changing to a sufficiently high degree the operational modality of the mind’s underlying machinery) while still replicating the functional modalities thereof, it serves to illustrate, using an intuitively obvious case, that it is possible to both a.) replicate the functional modalities of the brain while failing to replicate the operational modalities, and b.) replicate the functional modalities of the brain while failing (presumably) to replicate subjectivity – in this case due to rearranging the components of procedural steps (e.g, the order of successive state transitions) in between each procedural step actually taken or state-transition actually made.
Either we assume that there are certain aspects possessed by the brain that facilitate subjectivity, or that subjectivity is a natural property of all material systems – a philosophical view called panpsychism or sometimes hylozoism. While our minds operate in the same manner (i.e. largely embody the same functional and operational modalities), we don’t form a continuous part of eachother’s experiential continuum. Our sense of self stops at the skin, and does not continue on into the neural machinery of nearby persons. This is because an intermediary gap is created by our respective neurons being separated by material systems that do not possess these supposed subjectivity-facilitating aspects. Thus it appears that the neurons comprising our brain have shared structural or operational/procedural properties that allow them to form a continuous subjectivity, and the material systems separating the neurons of two alternate nervous systems (i.e. the non-neuronal parts of each person’s respective body, and the air – and whatever else – in between, that separates the neurons in two different nervous-systems) do not possess such attributes. If panpsychism were true, and subjectivity were a property of all material systems, we could expect the neurons comprising the nervous systems of two different people, and all material systems in between, to form one continuous experiential continuum. Similarly, if we had a computational emulation of our own brain, which exhibits all the same functional modalities of our own brain, it would still fail to become a continuous part of our experiential or subjective continuum, because they are separated by intermediary material systems which do not possess the attributes presumably facilitating subjectivity or experiential continuity.
This thought experiment goes to show that there must be some structural or operational aspect(s) or property possessed by those neurons that are a.) shared by a given nervous system and b.) are both able to both form a continuous part of that nervous system’s experiential continuum - and the aspect or property allowing them to form such a continuum cannot be an aspect or property of all material systems in general (because then all material systems would be part of the same experiential continuum), Thus we could fail to include such an aspect in either a a computational emulation or physically-embodied “replicatory system” (what could be usefully considered as a physical or physically-embodied emulation) of a given nervous-system unless specific precautions are taken to avoid failing to replicate such aspects or properties - because such properties are not inherent in or possessed by all material systems.
While I’m optimistic about the chances of retaining subjectivity (instantive subjectivity) and subjective-continuity (temporal subjectivity) through such gradual-replacement procedures, the possibility of unwittingly losing our subjectivity is still present and so should be taken into account when considering the best approach to life-extension. Drexlerian nanotech is not necessary exempt from this – though it is comparatively more likely to not disrupt or fail to maintain instantive subjective continuity or “subjectivity” than whole non-biological approaches to gradual or iterative substrate replacement are. Varieties of indefinite longevity through nanotechnology (usually) comprise either the upkeep and maintenance or the iterative replacement of a.) the cellular components themselves, directly, or b.) the genetic material responsible for the correct protein transcription, synthesis and transport of such cellular components (thus maintaining such cellular components vicariously). They are less likely to cause subjective-discontinuity because by replicating the structural organization of the existing biological system using analogous components (as opposed to replacing them with non-biological analogues that embody the same functional modalities but via a different operational modality – which I designate as the specific, procedural method they use to instantiate of facilitate that emergent functional modality), they would presumably preserve whatever properties or attributes form the basis for subjectivity, which we know exist (because we possess subjectivity, and so our own phenomenological experience is proof of the existence of subjectivity as a property of certain material systems) but which we cannot verify for lack of a quantitative measure of subjectivity. In such a Drexlerian nanotechnological case we would be changing neither the existing functional modality (i.e. emergent effects) nor the existing operational modalities (i.e. how specifically those emergent effects are generated), whereas in the gradual-substrate-replacement approach using non-biological replacements we are likely to use a new operational modality to achieve a given functional modality, unless specific precautions are taken to avoid using alternate operational modalities (simply because there are many ways to accomplish a given emergent “effect” or objective, and one possible but alternative way is likely to be more optimal (e.g. efficient, cost-effective, robust to procedural dysfunction or structural degradation), and this makes it comparatively more likely, but in no way assured, to cause issues of subjective-continuity.
More often than not (though not so rigorously as to be considered even a statistical law), systems and procedures for indefinite longevity require a more comprehensive methodological and technological infrastructure; they are comparatively more complex and require more capabilities (which seems to be a result of using one universal technique to fix multiple alternate sources of damage; a useful intuitive analogy is that AGI are more complex to create than narrow-AI, specifically due to their generality). This means that they are likely to be more costly, and to require more R&D time, than source-specific life-extension therapies. Thus they may be likely to come after source-specific life-extension therapies.
However, due to their comprehensivity, indefinite-longevity therapies could work as a tenable solution to multiple alternative sources of damage. We should then consider the possibility that we could ameliorate the sources of aging faster by funneling funding and effort, previously considered best allocated to separate life-extension therapies aimed at alternate sources of molecular damage, into a single indefinite-longevity therapy that can fix all such alternative sources of damage. In this way we might both be able to ameliorate such damage faster and with less cost and effort than we would have by developing each separate life-extension technique, and avoid the possibility that a new wave of source-of-damage lie in wait, and which require a categorically new set of life-extension techniques to combat them. However, this approach does have a drawback. By putting all our eggs in one basket we lessen the evolutionary diversity of candidate solutions. The more potential solutions we have for a given objective or problem, the likelier that one of them will succeed, because if one fails we have multiple other approaches to try. As to whether the potential benefits of this one-size-fits-all approach supersede the drawback of lacking evolutionary diversity remains to be seen, but the potential benefits of this approach warrant our investigating the matter further, and our formulating an analysis comparing such potential advantages to the known drawbacks of lack-of-evolutionary-diversity.
It is also worth noting that, considering how conservative most of the world is (a travesty, considering that consistency and conformity are antithetical to Man), we might be more likely to garner widespread desire and support for life-extension and/or indefinite-longevity by promoting biological approaches, rather than non-biological approaches, due to the aesthetic and pop-culture connotations gravitating around the word and concept of cyborgs, which appear to be dehumanizing, alienating, and dystopic to a lot of people. I for one am enchanted by both aesthetics – Man is at most his mind, and at his most virtuous in the virtuality of thought - so a body of fleshy furrows and carnal canals is just as beautiful as one of light-slivered streamlined silver to me.
Anecdotally, Gennady Stolyarov II (in the debate forum) raises some concerns (i.e. the inability of mind-uploading to preserve what I’ve called “subjectivity” or more precisely “instantive subjective-continuity” or an “experiential continuum”, and what Gennady calls “I-ness”, due to computational embodiment incurring spatiotemporal discontinuity - which is problematic because he holds spaciotemporal continuity to be a prerequisite for subjectivity) that further emphasize the need for clarifying the term “cyborgification”, and for dispelling some common (though admittedly easy to make) misconceptions stemming therefrom. This, to most people I think, broadly denotes both uploading (thus replacing biological components or processes with virtual, 2nd-order-embodied components or processes) and the replacement of biological systems with physically embodied (i.e. 1st-order embodiment) prosthetic systems. There are varieties of cyber-immortalism (that is, using non-biological techniques for unlimited longevity), that gradually replace the brain with physically embodied, non-biological systems and processes (both varieties of electromechanical system, as through the use of MEMS or NEMS, as well as molecular systems), and which thus have the potential to maintain spatio-temporal continuity through the gradual replacement procedure. While this falls, along with uploading, under the broad designation of cyber-immortalism, it should also be considered a separate category from more normative forms of uploading. And indeed, Gennady Stolyarov II’s concern, expressed through his commentary in the debate forum, may be a result of the problematic aspects of current paradigms of computation – namely their seriality. Being serial, they only instantiate one component or procedure-part one at any given time – whereas in the biological brain all parts coexist and are instantiated at the same time. This may be akin to disassembling you once for every procedural step taken or every time a component in your nervous system undergoes a state-transition, because at any given time there is only one component in “existence” – that is, actually operating real-time, with the potential to causally interact and interaffect with other components, other than as say static data in a memory store. In the 2nd installment of my 10-part essay introductory essay at Transhumanity.net, I describe how massive parallelism may solve the problem of spatio-temporal discontinuity in current computational paradigms. Another problem identified in that installment is the fact that while biological neurons still produce a signal when in an “inoperative” state (i.e. when they aren’t firing), called the resting membrane potential. I’ve referred to this as procedural discontinuity. The fact that normative computational elements falling under the purview of contemporary computational paradigms do not produce such a steady-state voltage when in an inoperative state, instead producing no signal whatsoever until they receive more input, may be akin to disconnecting a biological component from its surrounding components and re-inserting it every time such a normative computational element stops producing a signal altogether when in an inoperative state – or akin to externally inducing neuronal cell death, and revival, each time such an element is activated and then goes signal-silent.
Khannea SunTzu’s comment on “cyborgification” automatically implying a hard-take-off intelligence-explosion type of scenario also serves to illustrate another common and easy-to-make misconception or misconnotation resulting from the term “cyborgification”. I don’t think the term “cyborgification” necessarily implies uploading, or a hard-takeoff Singularity – even though connotations the term has accumulated over the years may well imply it to many people. Cyborgification could include prosthesis (including neural prosthesis) completely disassociated from computational hardware, embodying rather electromechanical or molecular systems which replicate the functional modalities of neural regions without strict computational emulation (or if you like, which facilitate computation via 1st ordered embodiment, rather than 2nd order or virtual embodiment). Neither do I think it necessitates augmentation, though it doesn’t preclude it either of course; the gradual replacement of the biological brain with non-biological (prosthetic/physically-embodied or simulatory/emulatory) analogues, while adding no new experiential or functional modalities, no categorically new skills or varieties of perception, experience and thought, can still be called cyborgification. While the term is a bit loaded, and may indeed engender such connotations, they aren’t necessarily inherent in the term.
And in terms of people feeling more uneasy about “invasive” therapies like those usually considered under the purview of the term “cyborgification”, this stems from an irrelevant dichotomy. All technology is invasive – that is, changes the way in which we act. Man changes technology changes man changes technology, in a self-reflexive feedback loop and reciprocally-concentric timelike curve. This was the point of Marshal McLuhan’s whole professional career, who said that electronic communications systems was an extension of the mind long before such thinking was remotely popular – the ways in which the nature of media (as opposed to their specific content) affected and partly determined the ways in which we live and react to such media. His hot and cold media is nothing less than the claim that technology largely determined the way in which we think – and vice versa. For anyone to talk of the dehumanizing aspect of technogies’ invasiveness is ultimately ironic, because they’re using one of the most invasive examples of technology – language – to libel such technology. How much more invasive could a technology get? Try not only describing, but even formulating a conception of the self, without language. It’s not easy, once you’ve been infected with it. Some technologies are so invasive that they end up becoming the newest platform for the self – this is the virtue of the virtual, and not a disadvantage. Technology has been invasive since before it had a name. The self would need to be concrete and static, rather than disparate and fluidic, to be really invaded in the first place. We invade ourselves; what else is consciousness but a turning upon oneself?