All content is by Stephen Ashworth, Oxford, UK,
unless attributed to a different signed author.
(1) Will humans be replaced by machines?
Well-known SETI astronomer Seth Shostak wrote recently in Acta Astronautica that intelligent beings would enjoy a period of only a few centuries after inventing radio before constructing the thinking machines that replace them. “So you’ve invented your successors and only for a few hundred years are you... a ‘biological’ intelligence.”
“Dr Shostak says that artificially intelligent alien life would be likely to migrate to places where both matter and energy – the only things he says would be of interest to the machines – would be in plentiful supply. That means the Seti hunt may need to focus its attentions near hot, young stars or even near the centres of galaxies.”
In his recent book, Paul Davies is saying something very similar.
Are we on the verge of provoking our extinction by inventing our own successors? Some relevant thoughts.
(1) Our species will not last forever in its current form. Three possible futures.
(a) We continue to develop biotech, infotech, nanotech, etc and continue to transform ourselves into what I call a technobiota: a presumably immortal fusion of biology and technology spreading through the Galaxy and developing further as it spreads out. (Shostak: the biology is shuffled off; me: the biology continues to be an essential component.)
(b-c) Alternatively: progress comes to an end and we succumb to change in the same way as other purely biological species, in which case (b) after a few million years we just go extinct; or (c) after a few million years we go extinct but not before engendering one or more successor hominid species, which may or may not redevelop technology and get the chance of option (a) again.
(2) Will we be replaced by our own machines? This is a simplistic question, based on a rigid distinction between the biological and the artificial – a distinction which is already out of date.
A couple of days ago, somebody at the Voltaire Foundation where I work said: “I must go and get my handbag – it’s got my brain in it.” She was joking, of course ... or was she? She meant that in her handbag was the notebook or pocket computer with all the data on it to help her do her job and lead her life. Even a simple pen and paper is an extension of our brains (think of all the things you can do with their help which you could never otherwise achieve – write a novel, do a mathematical calculation, keep precise records, send a letter ...), and the computer and the internet even more so. Of course they’re still not actually physically connected to our biological brains – but that’s a minor detail. For functional purposes, they are connected.
Paul Davies gets this on p.161 of his book: “the distinction between living and non-living, organism and machine, natural and artificial, is set to evaporate soon.” He settles on the term “auto-teleological super-systems” (ATS) for the beings that intelligent organisms such as ourselves are evolving into.
But his discussion of the question still falls back into “sloppy terminology”. The computer-machine concept leads him into supposing that our successors will lack the sense of personal identity that is so characteristic of human experience, and he ends up with the idea that our ultimate destiny is to transform ourselves into quantum computers which will sit around amusing themselves in cyberspace forever, perhaps solving ever more subtle mathematical theorems.
(3) Consider the pattern of evolution to date. When new types of creature appear, they generally do not replace the old ones but add to them. The appearance of eucaryotes did not render bacteria extinct. When multicellular animals and plants appeared, they did not replace single-celled organisms. For all the stresses and strains we humans have loaded onto the prehuman environment, we are not replacing our mammalian cousins outright, and there is even considerable interest in preserving our closest surviving relatives in the primate family.
This pattern does not prove anything, but it does suggest a future in which biological humans and a wide variety of engineered superhuman entities coexist.
The evolution of a new kind of life leads, not to a monoculture, but to the radiation of enormous variety, as in the Cambrian explosion of animal life. Our spreading into space would greatly multiply this radiation of varieties of post-human.
(4) Paul Davies provides a superficial measure of when computers (them) will rival humans (us). The human brain is capable of up to 40 trillion synaptic neuron firings per cc, or 40 teraflops in computer jargon, so theoretically up to about 10,000 teraflops for the whole brain. The fastest supercomputer currently achieves 360 teraflops, but if Moore’s law holds up then machines could surpass the power of the human brain by 2020. Once that line of sheer computing power has been crossed, Davies argues, “in principle, artificial intelligence could rival human intelligence”. But, he adds, “there are huge caveats”, to do with architecture and software (p.156-157).
What commentaries of this kind usually ignore is the fact that people and their machines are enjoying ever-closer integration. In my lifetime computer inputs have progressed from punched cards, to keyboards, to point and click. By the time Moore’s law gets us to the point that our laptops can manage more floating-point operations per second than the human brain, I would expect to find users interacting with them directly by thought. I think we are looking at, not the replacement of biological brains by digital ones, but their gradual merging into a composite brain with both biological and digital lobes. A process that began when an ancient Sumerian scribe first scratched marks onto a clay tablet in order not to forget what they represented.
The point is not whether humans or computers can think faster or more intelligently, but whether the symbiotic human-computer combination can, which it clearly can. So rather than artificial intelligences replacing humans, I think that the man/machine distinction will fade away as we integrate ever closer with our machines.
Listen to Stan Getz playing the tenor saxophone, or Sergei Rachmaninov playing the piano, and try to tell me the perfect fusion of man and machine has not already happened!
(5) Whether the resulting post-human beings will look more biological or mechanical is another question. They will certainly not look like C-3PO from Star Wars (or the very similar Tin Man from The Wizard of Oz). Nor will they be boxes sitting in buildings somewhere and communicating with their human servants on screen. We will not be able to look at them and say: Oh my god, it’s a bloody machine!
Actually, I think the post-humans will look very like us. Why? Because life is more than cerebral thought, more than adding up floating-point operations per second. The idea that when computers surpass the sheer number-crunching power of our brains it’s game over for anyone who began life in the maternity ward is pure science fiction. What we’ll have will be an intimate co-evolution with both biological and manufactured elements, driven by existing global markets. And one in which, while mental computing power may be one desirable factor, quality of life experience will be another.
My iMac by itself has zero quality of life experience – in this dimension, Moore’s law hasn’t even got started. But the combined Stephen-iMac organism certainly has better quality of life than either one alone, and this is the way it will go further.
But even if it could exist on its own, any “intelligent machine” with a grip on reality will still want to experience what we experience. It’ll want to climb mountains, play football and have sex. Especially sex. Yes, maybe there’ll be some super-brains in bubbling glass vats or in humming plastic boxes, dispassionate, humourless and relentlessly logical, intellectually light-years ahead of any human genius. But they must necessarily take part in a broader economy involving more active beings who do everything we do and more. Otherwise they’ll be unable to pay the electricity bill, and eventually become extinct.
They’ll inevitably be plugged into the internet. How much of the internet – how many of its floating-point operations per second – is devoted to solving mathematical theorems, and how much to music, movies, and porn? QED.
(6) Is pure intelligence infallible? How, after all, do intelligent people operate? Do they proceed logically from premises to logical deductions? Then why do smart people on occasion believe and do weird things? (As discussed by well-known skeptic Michael Shermer.)
No. We come to our beliefs and interests in all sorts of ways, and then the cleverer we are the better we are at justifying what we believe and why we are into such-and-such. Assuming we eventually build robots that are clever enough to stand up and demand legal recognition, they will be no different. They will believe and do weird things, and be even better than we are at justifying themselves.
Shostak’s key point is that as computers are involved in designing their successors, so their rate of evolution is vastly faster than biological evolution has been. True, except that the designing will work with both computer chips and genes. While we don’t know what the outcome at any stage will look like, I’m betting that market forces will drive it in the direction of beings that look like us and enjoy the kinds of things we enjoy. If, say, their skin is manufactured plastic, it will be no less soft and sensitive than human skin, if perhaps it can also be transformed into a hard protection on occasion when necessary. They will be enhanced humans, because that is what people will demand and how the economy will grow.
The take-home message is that in my view, we can confidently plan for an expanded human future, on Earth and in space, without worrying too much that the society we’re creating is something dehumanisingly dreadful which we would disown if we could see it in advance.
And the message for Seth Shostak and SETI is: keep looking at sunlike stars and their planets (but you won’t find anyone there, because we’re the first!).
Shostak quotes from the BBC report by Jason Palmer (23 August 2010).
See also: Seth Shostak, Confessions of an Alien Hunter (National Geographic, 2009) – an entertaining and lively read.
Paul Davies, The Eerie Silence: Are We Alone in the Universe? (Allen Lane, 2010).
(2) The ISS: ten years of life in space, but a long way to reach sustainability
On 2 November 2010 the International Space Station passes its tenth year of continuous occupation (not counting the first two-day taxi flight to reach it). The ISS exceeded the previous record, set by the second long-term occupation of Mir at 9 years 354 days, on 22 October 2010 (25 Oct. according to NASA).
But with access costing $30 to 40 million per passenger and limited to half a dozen manned and half a dozen supply flights per year, the achilles heel of the ISS is clearly in the highly inefficient transport systems, used to access it. This must be a classic example of an engineering project which is overdesigned in one aspect (the orbital accommodation) and underdesigned in another (access). An object lesson for students of systems engineering.