On Oct 8, 2012, at 6:07 PM, Jed Rothwell <jedrothw...@gmail.com> wrote:
> This is quite different from saying that "computers will definitely become > sentient" (that is, self-aware; conscious). We do not yet understand what > sentience is so we cannot predict with confidence that machines can achieve > it. I expect they can. That is an unsupported opinion. I agree, but I doubt that they can be sentient in foreseeable future. And I base my unsupported opinion that sentience is not just making an intelligent computer, but it is more about motivation. And in the case of human's the motivation to be sentient is quite strongly moulded by evolution. Also the difference between humans and most of the other smart animals, such as elephants, dogs and dolphins is that they lack motivation to develop themselves although here it is only a matter of degree, not qualitative difference such as between animals and computers. I would speculate that this strong motivation of humans, is coevolved with the language. Similarly as dogs are coevolved with humans, so that dogs can understand humans better than any other animals expect other humans. Therefore, if we are to make sentient machines, there would need to be preprogrammed huge amounts of irrational motivations, behaviour and sex drives to do, to progress and to thrive. Learning, reading books and thinking (what we are right now doing) is itself non-rational process and machine will not just get it without complex and subtle pre-programming. And the programming task is not easy, but it must probably be done using directed and accelerated artificial evolutionary process. But I doubt that that this is done soon, because while developing sentience we must face huge ethical dilemma that is it right to discard semi-sentient, lets say chicken level immortal and artificial beings just because we are learning to program cat level sentience? And from cat level, there must be created probably billions of sentient beings until we reach primate and dolphin level sentience not to mentioning if we want to go beyond humans. And also there is problem that it might be impossible to direct artificial evolutionary process accurately enough that it will suit our needs. I do not think that there are short cuts for sentience. However artificial intelligence is by itself evolving rapidly and I would expect that we will make a breakthrough in genuinely intelligent algorithms during the next 30 years. Although strong AI is like hot fusion, that it is always looming 30 years ahead in the future. And cybernetics is the way to go near future! > I do not think biological carbon based computers (brains) have any special > properties that cannot be emulated in silicon or other materials, but I could > be wrong. Yes, I indeed thing that you are wrong, because silicon based brains lack the evolutionary process. Behind human brain, there is 200 million years of evolutionary selection. This is not something that can be done overnight. And I really doubt that kurzweilian neurone by neurone simulation of brain will bear fruits. But I could be wrong. > > In the future I expect every phase of house and building construction will be > prefabricated. Only the final assembly will be on site. There will be two options. First is a prefabrication that will come in massive scale that everyone in the face of the Earth will notice it in November 2012. Because Chinese are going to prefabricate and assemble the world's tallest building that is to be completed in March 2013. There has already been prefabricated 17 smaller skyscrapers, but this one will be the biggest building in the world with 800 000 residential square metres. The cost of construction per m² will be just one sixth of that of Bjur Khalifa. http://en.wikipedia.org/wiki/Sky_City_(Changsha) Note that in Wikipedia there is an old completion estimate in January 2013. Second option is just brute force 3D-printing of houses. This process will suit well for two storey bungalows. So, we have serious problem in near future. Do we want a mile high megacities where a single prefabricated skyscraper will actually hold the whole city with full city infrastructure or that everyone has afford to dirt cheap bungalows with huge environmental footprint? As due to virtual reality, distances do not matter that much. Build a custom home in 20 hours using a giant 3D printer http://dvice.com/archives/2012/08/build-a-custom.php Perhaps if we force agriculture to skyscrapers and deserts, then there is enough room for humans to live comfortably in bungalows. So we turn the idea of city and country side upside down. That in the future humans will live in countryside, while food is produced in the cities and skyscrapers! > On the other hand, I see no reason why interplanetary human colonization > cannot be achieved, with travel time between the planets of weeks or months. Contrary to popular belief, Venus is most hospitable planets for humans to live, because at 50 kilometres altitude pressure and temperatures are at goldilocks zone. Solar energy is abundant, day length is just from 20 to 120 hours due to strong and steady circular atmospheric currents. And most importantly, breathable air (either nitrogen+oxygen or helium+oxygen) is strong lifting gas in carbon dioxide atmosphere and both mixtures are readily extractable from atmosphere. Also 50 kilometres is above the sulphuric acid haze that is present in altitudes between 20 and 45 kilometres. So, if we are to find second large scale home for civilisation, Venus will be next in line. Mars will be just third, because it will take some time to terraform it. However permanent Mars base will be started earlier, because in situ resource utilisation will be somewhat easier for small scale base. Also there is more urgent science to be done in Mars than in Venus, that serves only as a second home for human kind, where as Mars base will be mostly scientific, before martian atmospheric pressure is within safe limits for humans. If polar carbon dioxide ice gaps are melted, there should be enough for 30 kPa atmosphere that is enough for humans to live there at larger scale. > This is especially likely if something like a space elevator can be > constructed. It seems likely to me that an elevator is possible, based on the > book "The Space Elevator." > The main problem is that reusable rockets are cheaper than space elevators. Reusable rockets are two orders of magnitude cheaper than current two or three stage launch vehicles. Also there are at least two promising prototypes already built. Those are SpaceX's Grasshopper and it's Merlin 1D engine and Reaction Engines Ltd's Skylon single stage spaceplane and it's air breathing Sabre engine. If either one of those quite different approaches to reusability will succeed, space travelling costs are reduced by two orders of magnitude and deep space flights to Mars or Venus by three orders of magnitude. My prediction is that first reusable spacecraft will reach orbit in early 2020's. —Jouni