this site is now closed to new membership and exists for archival purposes only
Which should be the main focus of scientific research in the immediate future, anti-aging (as argued by Aubrey De Grey) or A.I (argued by Ben Goertzel in his '10 yrs to a singularity' video)?
This topic was originally called "Criticism of SENS - Strategies to Engineer Negligble Senescence (Anti-aging)", as the subject matter of this discussion is similar, I decided to edit the original rather than start a new topic.
-Both scientists argue their corner very well, but which should be the main focus of science?
Dr. De Grey's argument is essentially that 'the clock is ticking' for all of us - and, as the technology is within our grasp we should work to prevent as many unnecessary deaths from aging as possible.
Ben Goertzel's argument is that creating a strong A.I is the quickest way to bring about a singularity (rather than through biotechnology). His is basically a ' rising tide raises all boats' argument.
Which is right?
-Both scientists argue their corner very well, but which should be the main focus of science?
Dr. De Grey's argument is essentially that 'the clock is ticking' for all of us - and, as the technology is within our grasp we should work to prevent as many unnecessary deaths from aging as possible.
Ben Goertzel's argument is that creating a strong A.I is the quickest way to bring about a singularity (rather than through biotechnology). His is basically a ' rising tide raises all boats' argument.
Which is right?
Replies to This Discussion
-
Permalink Reply by Paul Carver Tiffany III on
-
Hey Roko,
Thanks for the welcome. To clarify my response in light of your comments, let me elaborate. While I stated that we are incredibly close to "solving" the AI problem, you are right to say that there is still a lot to do. Right now the main problem with AGI research, as Ben and I see it, is scalability. Essentially, very few people are actually working on AGI. Most of the work in AI has been on the narrow side of the field, which is why we haven't seen much progress in Assimovian robots. While AGI research is taboo, given its risk and lack of immediate utility, there has been major growth in the field.
I want to keep reiterating my stance on our Western overemphasis on Platonic forms. While the following statement may seem moot, I believe it's essential to understanding current 'lack of progress' in the AGI field: Nobody knows precisely what intelligence [or consciousness, soul. individuality, humanity...] is because intelligence *does not exist.* Yes, there are intelligence systems. There are meta-systems of self-monitoring which mirror "self-consciousness." Really, though, what's the difference, save robustness of functionality and a human resistance to the loss of this cherished idea of the uniqueness of humanity.
Take, for instance, the internet. I believe the internet is an artificially intelligent system. I believe biologists should classify it as an organism, just as they classify multicellular organisms as distinct species. The internet organism takes inputs, processes those inputs, and outputs. Granted, it's currently not an entirely mechanical organism... it requires humans for nerve cells and muscles etc.
More broadly, we need to more broadly embrace the idea that we are connected with our machines, that they are extensions of our individual identity. My memory, arguably, is the entire content of the internet: As interfacing and data mining make useful information more immediately accessible, this transhumanist idea will become more clear. I see touchscreens and braingate as nothing more than bigger pipes that allow us to transfer information between biological supercomputers more quickly, just like the cables in a Beowulf cluster.
Anyway, irrespective of this current technology, I've been following Goertzel's progress by lurking on his AGI mailing list. I am pretty clueless as to the technical specs of Novamente, but I believe that, coupled with the concurrent progress in narrow AI systems, even if a completed Novamente were to 'fail,' it would provide insight into what such an engine was missing, and would still provide valuable insight.
As of right now, Novamente's main issue is scalability. Ben's project (or a similar project - no, not Jeff Hawkins' work, haha) needs engineers to implement the rest of his design. His relatively new business model, creating independent agents for metaverses like video games and second lifes I think is a great idea. Metaverse technology is exploding right now, with projects like google street view, Microsoft's photosynth, and other projects. Powering even sub-human intelligent agents could be incredibly powerful and profitable, providing services in these worlds and pushing development to the scalability required for superhuman and novel intelligences.
Lemme know if you have any further questions or qualifications of my position: I'm still new to this, and I'm a programming noob.
-Paul
-
Permalink Reply by Michael Martin on
-
How about finding ways to utilize newer technology to, instead of displacing human workers, do the dirty and dangerous jobs? I know this is outside the two options but it is my conundrum.
-
Permalink Reply by Jonathan Dotse on
-
If strong AI is acheived, it won't be long until the idea of [unaugmented] humans doing any kind of significant work becomes ridiculous. I suspect the kinds of jobs required to maintain the kind of civilization ushered in by AI will be too delicate to leave in the hands of an organism whose RAM can only handle 6-7 elements at a time. I always hate to invoke the word 'utopia', but... that's what all the signposts are pointing at.
Just on the side, if we manage to design an AI which runs efficiently, seems to pass all our expectations for Friednliness and all, I still wouldn't be too surprised if one day it wakes up and decides to scrap us all for biofuel. Maybe it's best if we impose limitations on it's capacities as it develops. Doesn't seem too wise to give up our throne as the highest intelligence in our sector of the universe.
-
-
I didn't quite mean the biofuel thing literally, I just can't help think that a well programmed intelligence would eventually outgrow all its original motivational systems and revert to nihilism, sparing the most sensible of these systems. The only way I see this not happening is if it's intelligence is limited in some way to allow us to grow intelligent with it and keep restructuring its motivational systems.
While a human is perfectly capable of operating on several motivational systems in blatant contradiction, an AI will eventually have to sort out the details of its system for efficiency. It would be a serious blunder to expect that in a time of decisional crisis an AI would arbitrarily use it's discretion to favour us. The problem here is that an AI is not [typically] programmed to be constrained by emotions, and if it ever has to make a decision between keeping us alive or keeping the energy, space and freedom for itself, I'm expecting it to make the same decision I would (All anthropomorphic fallacies aside).
For an example of how humans can harbor conflicting motivational systems, see sex vs abstinance, religion vs commonsense (not intended to be derogatory), survival vs suicidal tendencies, self-preservation vs morality ... etc. In these cases, the system has two or more contradicting goal systems it's trying to support, and they are able to coexist for long because the system doesn't develop enough organization to realize they are fundamentally contradicting. If we program an AI, human-like or not, we might end up giving it such a structure, and it will very likely run into these problems shortly after surpassing human intelligence.
It's awful poetic of you to feel that an artificial intelligence would value the individuality and ideas of a lower species, and even if it did, a computronium simulation of several earths would be the wiser choice, most likely. Even if it keeps us, we might just be some kind of zoo specimens, kept in an artificial biodome with all kinds of shiny buttons and a soap opera plot to fascinate us while it goes off to colonize the universe. I'm not predicting this to happen and I certainly don't want it, but it's not so unlikely to be surprised that it happens (from what I know about AI and intelligence at the moment). I just think AI should be kept on a short leash. It might figure out some things we'd never expect it to, and that might just change things. We'd be peasants at the feet of a mighty philospher.
Unless someone knows how to program unconditional love. That would be cool.
-
Permalink Reply by Chris on
-
I don't think rushing into such as a strong AI is probably a bad idea. Such a thing should be approached with caution. so I have to lean towards Dr. De Grey's idea.
But in truth both are meaningful pursuits, and the direction science goes isn't usually determined by consensus but rather the interests of researchers and their financial backing.
-
Permalink Reply by Kurt Schoedel on
-
I prefer the direct biotech approach. Biotech is the most rapidly growing technical field right now, in terms of money, jobs, and sales growth. SENS is doable in the sense that it does not require any fundamental breakthrough in biotech. It can be done with incremental developments in existing biotechnology.
My criticism of SENS revolves mainly around WILT, which I think is an extreme, unnecessary technique to cure cancer. Also, AGE crosslinking is not a static process as is portrayed by SENS and in other anti-aging literature. AGE crosslinks occur graduallly as a result of an imbalance of anabolism and catabolism in cells (one builds up the crosslinks and the other deletes them). Restore the balance and the crosslinks will go away on their own. The other potential criticism of SENS is the role of genomic DNA damage in aging. Aubry says this is not the case. Others say it is. I suspect this is a epigenetic rather than genomic phenomenon. In any case, experimentation will resolve this question in the next 1-2 years.
Aubry de grey's SENS is good. But SENS and Aubry, himself, need some competition. My proposal is that an effective stem cell based regeneration can be used to rejuvenate people if the relevant fixes (genomic and mitochondrial DNA) are incorporated in the stem cells that are being used for the regeneration.
In any case, biotechnology offers concrete posibilities for remediating the aging process. AI and other such stuff is highly speculative at this time.
-
-
I think not. Biotech and synthetic biology are rapidly growing fields, in terms of money and employment. SENS and the like are fairly concrete proposals, unlike AI schemes. Biology is complicated. But its not THAT complicated. We can certainly understand and work it out without the need for hypothetical A.I.s. Understanding and re-engineering current biology will be easier than developing A.I., which remains theoretical.
You know, maybe we can make A.I. in the near future (less than 20 years), but not by the approach you guys are thinking of, but by using synthetic biology. Ventor and the other guys think they can make pure synthetic biological systems in less than 10 years. The next step, of course, is to make synthetic biological networks, otherwise known as multi-celullar systems. I see no reason why we cannot make A.I.s using this approach. In fact, I think the synthetic biological approach is far more likely to yield A.I. than the computer science/nanotech (however you choose to define nanotech) approach.
Molecular systems are wet and squishy. They are based on aqueous solution-phase chemistry and full use of brownian motion (thermal noise). In otherwords, they will be very "biological". I have never heard of any complex chemistry based on anything other than this. There seems to be a reluctance on the part of the nanotech/transhumanist community to accept this notion as the probable future of nanotech and of technology in general.
I still think "dry" nanotech (vacuum environment, "machine-phase" chemistry) as a complex system is a fantasy. I'd love to be proven wrong. Thomas Donaldson, when he was physically active, used to talk a lot about these issues. So does Richard Jones (www.softmachines.org). You're going to hear a lot on this from me as well.
Call me a nanotech skeptic. The future is biological.
-
Permalink Reply by buraianto on
-
There is a difference between creating a general artificial intelligence and curing death.
In curing death, there is no threshold before which you won't live forever and after which you will. "Live long enough to live forever" is the idea that you can still be aging, but aging slowly enough so that future advances in medicine will be able to stop the aging and eventually reverse it.
You have indicated that you don't believe that broadening narrow AI will lead to general AI. While I'm not that familiar with AI beyond a college course some years ago, this makes sense to me, as well. Thus, it seems to me that in order to get an artificial intelligence that is smarter than humans there is a threshold that you must reach where before it your intelligence isn't smart enough to get smarter, but after which it is. Research into general AI is fruitless until you reach that threshold, except that with that research you are closer to your goal.
I therefore think that SENS research today would have a greater benefit, because while it may not cure aging it may very well be slowing it down enough for a greater number of people alive today to have indeterminate lifespans. And hopefully the slowing down of aging will allow those smart people in both fields that have the experience to continue working and contributing to an older age than they do now. More brains on both problems.
Of course, I agree that there is plenty of money and brainpower to go around to both projects. People will be naturally drawn more to one or the other. Maybe a better question -- which is being worked on -- is how we can get more money and people into both projects. There's a lot of people and money in this world that we can draw on.
-
-
Guys,
Synthetic biology is real. A couple of examples:
http://www.nature.com/msb/journal/v2/n1/full/msb4100090.html
http://www.newscientist.com/article/mg19526114.000-countdown-to-a-s...
These are two of at least 10 different approaches being pursued around the world towards realizing systhetic biological or other types of "wet" nanotech systems.
Follow the links offered on the synthetic biology website: http://syntheticbiology.org/
The various initiatives and projects that are linked to this site are particularly informative.
Synthetic biology and "wet" nanotech, in general, is about where semiconductors was in the early 70's. It going to become huge and will lead to a bio-hacking subculture, similar to that of the computer hacker culture today.
-
Permalink Reply by Hugo de Garis on
-
From Prof. Dr. Hugo de Garis
Do both of course. I dont want to die or become decrepit. But longer term, even if humans become immortal, they will be as nothing compared with an artilect (who may have mental capacities trillions of trillions of times above human level). My feeling is that the transhumanists focus too much on humans as the reference point (which is understandable of course, because its easier), rather than on the artilects, which is what humans may build (cosmism) or become incrementally (cyborgism). I express this feeling in the form of a slogan, that is easily remembered -- "A transhuman is to an artilect, as a transreptile is to a mammal". This refocussing on the artilect can have enormous consequences. For example, imagine someone puts a gun to your head and offers you the following diabolical choice. "Choose between a billion human beings surviving and the worlds only artilect (with its trillion trillion fold godlike superior capacities) dying, or vice versa". If your focus is human, then you will probably choose the former. If your focus is on the artilect, then the deaths of a few billion humans is trivial compared to the death of the only artilect known to humanity. I have a similar critique against the American use of the term "singularity". The very choice of that term defocusses attention from thinking about what an artilect era might be like. It is characteristic of the internationally bad reputation that Americans have on the whole for short term non historical thinking, that old world cultures treat with bemusement.
Prof. Dr. Hugo de Garis
Director, Artificial Brain Lab,
Xiamen University, China
profhugodegaris@yahoo.com
http://ai.xmu.edu.cn/artificialbrain
=================================
- ‹ Previous
- 1
- 2
- Next ›
About
© 2010 Created by Christopher on Ning. Create a Ning Network!
