| Atomic Razor |
|
|
|
In Here or Out There? There are lots of futures emerging at the moment, a glance through the SF of today gives us any number, but the major ones seem to be:
Now, we all know that science fiction has been pretty poor at prediction. The purists will point out that this is for good reason. Science fiction isn’t really about prediction, it’s about how science and technology will affect people. Still, the whole point of Atomic Razor was to investigate the idea of what happened to the future we were expecting. We know what happened to the space faring future – it was too expensive. We went to the Moon but there was nothing there that we didn’t already have on Earth in abundance (I will ignore the helium-3 until we can actually use the stuff). All the other space related things have fallen by the wayside; we don’t need space based solar power satellites, we are no running out of resources here on Earth any time soon and national prestige doesn’t seem to be what it used to be. There is an outside chance of tourism making a difference but space, as we shall consider, has lots of other problems too. Dystopias, like portrayed by Stephenson, Gibson or MacLeod seem slightly unlikely at this stage. The world is still running along pretty much the same lines, if anything things are possibly better for the average citizen of the West than they were a decade ago – the War On Terror notwithstanding. Computer technology continues to become pervasive in ways which Gibson really didn’t grasp so well.Which, rather neatly, brings us to the Singularity, something that MacLeod, Vinge and others like to talk, to a greater or lesser extent about. The problem with the Technological Singularity, to give it Vinge’s [book mark] full name, is it’s not all that easy a concept to get across. The basic thesis is this: when computers have processing power equal to humans, there will be the potential for Artificial Intelligence to be practical. Assuming that Moore’s Law [book mark] keeps working as it has for the last 40 years (and shows no signs of breaking down any time soon) then raw computer power will double approximately every 18-24 months. Therefore 18 months or so after the first computer capable of human level processing, you’ll have a computer capable of twice that, another 18 months and it’ll be 4 times human level. Computing power increases exponentially thus the time taken to carry out any given calculation decreases exponentially. You enter an era of potentially transhuman intelligences. New MRI scanning techniques are giving us new insights into brain processes and teams, like the BT Labs have been working on methods of simulating human minds and uploading. This also means that AI’s need not arise from pure research on making computers “think” but also from human/machine hybrids. The AI’s could just be humans making use of all the extra computer power to run simulations of themselves or something like that. However, Vinge’s idea is that as you enter this realm you pass the point of being able to predict what the beings who emerge the other side will be like. Hence the term singularity. There are lots of forms this could take, and I’ll list a few, the main thing is, of course, that we have simply no way of knowing what this could be like.
There are some important points to note here. We can’t predict which of these is most likely, only that something is likely to start happening when the machines and humans are operating in a similar processing realm. It doesn’t necessarily have to include all of humanity nor will it necessarily be pretty. Many SF fans and scientists alike get annoyed at this point as it “smacks” to them of being like the biblical rapture. Unfortunately there are probably worse analogies. Compared to other futures this one has something of an edge over other predictions. Firstly, we can see Moore’s Law working everyday, processing power is doubling every 18 months and there is a *lot* of money being spent to keep it that way. AI research is happening and can be conducted by small teams with relatively little resources. This last point we will come back to when we consider the alternative future in space. So, just by sitting around for 20 years we can expect to see machines with the abilities to match humans. Now, that is not to say that this will happen. We still don’t understand what intelligence is, nor what constitutes being self aware. If a machine rolls up and says “Cognito Ergo Sum”, you can quite easily reply, well somebody told it to say that. However, the development trend is there and will not go away just because we don’t think it will be likely we can work out how to build AI’s. Something to think about is not what happens when computers have human level processing abilities, but 5 years later when the computers have 8 times more processing power than humans. Through MRI scans and other technologies we are getting very interesting images of how the brain works and processes data – over those 5 years with the ability to run simulations in real time we may well be in the realm where people are able to run simulations of themselves significantly faster than they can think. For raw problem solving, this could make a huge difference. Not to mention algorithms that learn, evolve and adapt all significantly faster (and at an exponentially increasing rate) than a biological entity. As I said, it’s by no means certain, but it’s worth thinking about. So, why do I think this future, where we loose the ability to predict things accurately in about 20-30 years is more likely than one where we head off into space? As people will point out, AI is highly speculative, as is the concept of uploading [book mark] – whereas we have space travel today and history shows that humankind has always expanded. They will argue that space travel is a here and now technology where all we have to do is wait and space travel will continue to happen. Well, will it? Let us compare two technologies. Flight and space travel. The first heavier than air flight took place in 1903; a decade later people were trying to run commercial services and there were dozens of plane builders. Twenty years later there had been a war with fighter aircraft and commercial services were running between cities. Within 50 years there had been a dramatic paradigm shift as we entered the Jet Age. The first orbital flight occurred around 50 years ago using disposable rockets, within a decade we had lots of disposable rockets of varying sizes and man was about to land on the Moon. This allowed people in the 60s and 70s to make excellent predictions about a moonbase, Mars landings, orbital colonies and so forth. But there was something which the authors of those predictions hadn’t noticed. An individual working alone can build a basic aeroplane in a few months for a few thousand dollars. (There is actually a home aircraft design from the 1970’s which you can build for under $1000 using equipment from a DIY store.) An individual working alone after 40 years of trying has yet to launch a rocket into orbit. All the rockets built and launched had been built by governments or large corporations drawing upon huge reserves of technology and talent. Now, there are two ways to look at this. It is either because of a conspiracy to keep space travel in the government realm, or it’s because space travel using rockets is terribly, terribly hard and this very, very expensive. I have not seen one shred of evidence that it is the former and lots and lots that it is the later. The Earth is a big planet, which means you have to accelerate to around 11km/s to reach orbit. This isn’t an easy thing to do at the best of times, and very difficult with current technology and launchers (i.e chemical rockets with exhaust speeds <11 km/s). We need a paradigm shift, like we did in the time of jet travel, where suddenly we could fly routinely more than twice as fast and multiples of the height. However, for space travel I am not even sure we are at the Wright brothers in terms of a space flight paradigm. I think we are still at the ballooning stage, we have yet to build a simple, reusable space vehicle that can be operated by small teams of people. Several have tried, ROTON being a recent attempt, but as yet there are no plans on any drawing board that I know of which will bring this to reality and bring the launch costs down from their current $10,000/kg to something resembling the $1/kg or less which what we pay for conventional flying (i.e four orders of magnitude in cost) Here’s another thought. Current launch costs are in the $10,000/kg to LEO, 30 years ago, the cost of a Saturn V were $1000/kg to LEO (1970 dollars) in other words, pretty much the same, if not slightly less. It’s not necessarily that surprising, the costs, in real terms of aviation haven’t dropped by all that much since the introduction of the 747. The only real recent change was the creation of the bucket shop travel agents and “low” cost airlines. They exist because we have lots of old, cheap reliable planes with many hundreds of thousands of flying hours left which the main carriers have replaced. Until we have the same for space, there isn’t a hope that those costs will come down. So, how about the profit motive? Space launching is a big business – in 2000 there were around 100 commercial and military launches for around 150+ communications, spy and other satellites. Assuming that the average cost of those launches is around the $120M mark (that’s something of an average but in the ball park) that’s a total 2000 market of around $12bn. A pretty good market to have a slice of. However, with the costs of developing new launch technologies running into the billions, if not tens of billions (remember the A380 a conventional albeit large aircraft is going to cost around $12bn to design and build) – the sums simply don’t add up for the launcher market to be able to make any huge cost savings to improve its profit. There is still the tourist market. Recent polls show that lots of people want to go into space. I will wait and see on that, but again, the sums don’t seem to add up. At $20m a ride on the spare seat of a flight which is going to happen anyway you have 3 or 4 flights a year which make sense. Its good money for a cash starved business. But even if you had 100 people a year to launch at $20m, that’s still only $2bn and even $40M (assuming you fill both seats in the Soyuz) which only just covers the costs of the launch. Even thousands paying hundreds of thousands still does not cover the development costs you’d need to build the large reusable launchers to that are required to make tourism something that will really work, but then the development costs seem prohibitively high. The only way I see it is for governments to throw money at the paradigm and hope for a change. I am not holding my breath. So, back to the singularity I am afraid. The future looks to be in here rather than out there. This is a shame, but on the other hand, as we can’t predict what is the other side it could also be far, far more interesting than living in domes on Mars…
Dave O'Neill (c) 2002
|
|
Comments: feedback@atomicrazor.com |