Grauniad: I am creating artificial life, declares US gene pioneer
Craig Venter, the controversial DNA researcher involved in the race to decipher the human genetic code, has built a synthetic chromosome out of laboratory chemicals and is poised to announce the creation of the first new artificial life form on Earth.
[…]
Pat Mooney, director of a Canadian bioethics organisation, ETC group, said the move was an enormous challenge to society to debate the risks involved. “Governments, and society in general, is way behind the ball. This is a wake-up call - what does it mean to create new life forms in a test-tube?”He said Mr Venter was creating a “chassis on which you could build almost anything. It could be a contribution to humanity such as new drugs or a huge threat to humanity such as bio-weapons”.
Mr Venter believes designer genomes have enormous positive potential if properly regulated. In the long-term, he hopes they could lead to alternative energy sources previously unthinkable. Bacteria could be created, he speculates, that could help mop up excessive carbon dioxide, thus contributing to the solution to global warming, or produce fuels such as butane or propane made entirely from sugar.
“We are not afraid to take on things that are important just because they stimulate thinking,” he said. “We are dealing in big ideas. We are trying to create a new value system for life. When dealing at this scale, you can’t expect everybody to be happy.”
Big ideas, big dreams, possibly big horrors. It seems that’s always the way with the really big innovations.
It should be an interesting time in the evo-creo forums, as the creationists move the goalposts again on what does or does not count as speciation and try to argue that human intervention means that speciation can’t happen naturally while simultaneously arguing that only God has the power to create distinct “kinds”. The ethics and intellectual property mavens will also be closely examining the propriety of granting a patent to Venter for his new organism (or will the patent be only for the technique to create such organisms? The article is unclear).
Yet another science fiction scenario moves from the fantastic into the mundane. What next?
This is the realm of very-out-there thinking. How about life that consumes nuclear waste or can store and release vast amounts of energy (effectively batteries), or having the ability to process materials into more-useful forms?
I can imagine that people interested in seeding other planets with life may be very interested in artificial designer life (notwithstanding many people’s, e.g. Lovelock’s, pessimism of life being able to exist on Mars).
We have used recombinant bacteria (basically human DNA inserted into bacterial genomes) to produce pharmaceuticals since the eighties (1978 in fact, according to wiki).
Nothing new here. The only controversial thing about Venter is that it was discovered that the ‘randomly selected’ DNA source for the Human Genome Project was actually him. oops.
Carbon will be ditched all together : there will be silicon only life. The idea that playing with DNA is somehow dangerous is to miss the point all together : the real monster of the future is sitting in development on your desktop and there’s nothing you can do about it.
No evidence of it, yet, Bill. “Clanking replicators” are still a very long way off, and artificial intelligence hasn’t had a really big conceptual breakthrough since, well, ever.
Robert, the way that AI is dismissed often amuses me. Maybe a big conceptual breakthrough isn’t required. Someone once came up with a definition of AI which basically went: AI is whatever computers can’t do right now. Once computers can do it, it is just routine computer stuff.
BBB
BBB: Yes it is, if we want the Grand Slam of, say, passing the Turing Test. Natural language understanding hasn’t progressed much in decades (arguably, it hasn’t progressed much since ELIZA).
For me AI is all about volition. As long as computers are only doing what we ask them to (including “random” fluctuations we “let” them do) then they’re not intelligent.
FDB: I’m not sure that’s a terribly useful metric. If you feel like reading Alan Turing’s original 1950 paper describing the “imitation game” he discusses this issue in some detail.
By the way, I’d recommend the paper to everyone if they’re interested; Turing writes clearly and the paper appears in a philosophy journal so the mathematical content is minimal.
I don’t doubt it’s an utterly useless metric in terms of actually setting goals in AI design, but it’s the point at which I’d start taking it seriously. Comprehension of language is one thing - just keep fiddling with the programming and there’s no particular obstacle to an extremely sophisticated “mind” - but the evidence AI (again for me) would be, say, a computer struggling to name something it hadn’t been prepped for. Making up new words, recombining existing ones to synthesise a new meaning, getting genuinely confused by novel experience and exploding in a holocaust of childish rage, that sort of thing.
AI metrics are tricky , but I think that the relevant metric is neurone to transistor ( or other swithcing element) . A bee has 1 million neurones and is capable of complex social behaviour.
A computer equivalent can’t be far away. From a bee to a human is a couple of a hundred million years in evolution but in the evolution of silicon( millions of times faster) can’t be more than a couple of hundred years.
Nanobots replacing our every cell seems a distinct possibility in the next 2 decades so I’m investing in fucking machine futures.
Rat - was “fucking” a verb, an adjective or an expletive?
Bill O’Slatter on 9 October 2007 at 9:07 am
That is most likely true. Certainly going the techno way seems to be the quickest and cleanest form of artificially incarnating intelligent life.
I suggest that artificial life can proceed in three basic bio-engineering directions: recombinant molecular, regenerative cellular and (what I suggest we term) recognitive modular systems. That is to say we can use technology to dramaticly extend and enhance intelligent life by redesigning genes (a la Venter), regrowing cells (a la Thomson) or replicating minds (a la Kurzweil).
I am guessing that symbiotic cross-fertilization will come into play with the advent of technological life. (Just as it did with the advent of biological life.) We have DNA computers and computerised DNA in the genomic map. Likewise we have technological prostheses (artifical legs) we are heading towards biological prostheses (lab grown organs). No doubt neural implants (accelerators and plug-ins) are on the drawing board somewhere.
So we will likely have convergence of diverse evolutionary forms. Carbon-based men will become more like silicon-based machines and vice-versa.
That is why it is absolutely critical that philosophers finally get down to outlining a proper ethical guide to the regulation of anthropomorphic moral agents. It must embody the ontological fundamentals of cognition and volition. One that reconciles diversity of forms with unity of function, construction of the new with conservation of the old.
But I dont see much evidence of that. Just the same old same olds of ideological speculation and moral indignation.
The point is that “volition” is an extremely slippery little sucker, at a fairly deep philosophical level.
Would you mind proving that your posts are of your own violition?
It seems that the media has jumped the gun on this anyway, according to a commentor over at Hoyden.
mmmbuzzz… mmmclick…
Sure, volition is VERY hard to positively identify (without resorting to a reductio ad Cartesium), but we can usually see where it ain’t, and questions about definitively getting a handle on it are usually just mind-games.
When a computer does something that genuinely surprises its designer (other than crash, melt or actually work as intended) then we’re on to something.
I don’t know about Bill’s numerical reductionism - there has to be something qualitatively different (or quant. in a way more complex than simply a neuron/transistor count). We can already study the bee language and I doubt there’s much more than logistics involved in making a bee-bot that would fool real bees with current technology. The problem is making a machine capable of actually building its own language.
“I don’t know about Bill’s numerical reductionism - there has to be something qualitatively different (or quant. in a way more complex than simply a neuron/transistor count).”
Here’s a mathematical analogy FDB. These transistors/switches/neurones create a representational space ( a representation of the environment, whatever that is) . The human representational space could be mapped to a device with a similar number of switches per /second. I don’t think anyone would have a problem with that. At the moment we don’t have a computational device of that size. The second thing is that that the representation space has evolved over the last couple of hundred million years.
Venter and his team are trying to find the minimmum number of genes that will allow a cell to function (or at least some biological proceses of interest). Anything approaching designer function is still a long way off - more than just a few *months* for sure.
I do.
Such a map would be merely that - no more a human mind than a really really hi-res google map would be a real world you could live in.
Bill O’Slatter on 9 October 2007 at 11:08 am
[Caveat Reader: Unlike theory of political culture, which is a piece of cake for even the rankest amateur, this stuff is way above my intellectual pay-grade. So treat the following with caution.]
I am a little bit skeptical of the Moravev-Kurzweil line in the quantification of technological progress. The difference between carbon- and silicon-based intelligence is qualitative as well as quantitative.
Contemporary techno-computational mechanisms have a vastly different architecture to bio-computational metabolisms. Roughly speaking, a computer processes information in series using a digital coding mechanism. A brain processes information in parallel using an analog coding organism.
Brains seem to be able to draw diversified information from objective existence and bind them into a unified formation of subjective experience. Building computers with greater grunt is certainly a start but it will not get us machines which express creative insight or even common-or-garden variety sense. It will just develop a genre of incredibly fast idiot-savants.
There are advantages for the evolution of analog over digital-type intelligence, at least operating on classical scales. According to Dyson, processing information in continuous rather than discrete forms is a better long term survival strategy.
Perhaps, as Penrose-Hameroff suggest, the utilisation of quantum mechanical process is fundamental to human cognition and volition. If so then the most promising way forward for building a computer with anthopomorphic+ intelligence is through the development of quantum computational systems.
Obviously such a machine would not operate according to classical Turing machine principles, it would be much faster and more probabilistic. Like neural networks this type of machine would be capable of more-or-less rather than either/or thinking.
Whether it could match or surpass the performance of the human brain is an open question which everyone with a computer science degree is hereby invited to answer.
PS Congrats to tig-tog and Robert Merkel for at least running this issue. Once upon a time liberal philosphers were at the forefront in analysing both the the practical and moral implications of sci-tech progress eg Kant, Russell. It would be nice if they took a lead in these matters, just for a change. Instead of endlessly fussing, fuming and tut-tutting over whether or not some member of the LN/P ministry may or may not have put a member of a minority groups nose out of joint.
Only if our godlike power is matched with godlike wisdom and benevolence.
FDB : reference http://www.abc.net.au/rn/philosopherszone/stories/2007/1916966.htm
Tho I think he overdoes embodiment . IMHO it is just another computational artifact tied to “sensory” perception.
Cliff : where does Jesus fit in ?
Yes, our species record on responsible usage of power beyond that which our neighbours own is not a reassuring one, is it Cliff?
Jack , there’s no evidence that human thought involves quantum processes despite Penrose’s claims. The idea of cloning complete brain states does give rise to what might be called pseudo paradoxes. Assuming you can and did so where exactly would you be ?
Cloning complete brain states might end up being a necessary process in the event of complete failure of all bodily systems , but in the case of mis diagnosis you would up with two versions of yourself.
Jesus may or may not be proof that Godlike wisdom and justice may be possible in man… depending on whether or not you believe in that stuff (unfortunately I have my grave doubts).
Those who make the analogy between a brain/mind, and a computer, on the basis that neurons are just like switches, are barking up the wrong tree.
The brain is not just a very large number of switches that can be directly reproduced in a conventional computer. The firing of a neuronal synapse is much more complicated than just switching a light on. There are other cells (glial) that interact with neurons, and seperately with each other (independent of synaptic firing), that are involved in this process. Synaptic firing is not just on/off, there are different degrees of pre-firing potential, etc, and we are a long way from a clear understanding of how this all interacts and coheres to produce a functioning brain/mind.
I am sceptical that we will produce genuine AI in my lifetime, if ever. Certainly not with conventional binary computer logic.
To answer your question Hilker would be to say to what extent are neurones and other electronic switching elements interchangeable. Some work towards this goal is in
“Modeling self-developing biological neural networks ” Berry H , Temam O in Neurocomputing 70 (2007) 2723–2734. this work is being done on C.elegans ( a worm) that laboratory workhorse of genetics with a brain of 302 neurones. It seems to me that a lot of progress has been made in this area. I
Questions of organic vs inorganic hardware aside, the operating system must be pretty crucial, and that’s what we have most trouble with in terms of “how minds work”.