Friday, June 08, 2012
I won't try and repeat his whole lecture but these are some notes:
Fields in which Robotics is expanding:
unfortunately he ran out of time so the last 2 weren't discussed, Oh well.
discussion of the DARPA Challenge to build a fully automated self driving car, which I have discussed as an example of an X-Prize working at 3% of conventional funding costs. The designer of this has now built a car for \Google & we saw some film of it driving in San Francisco on the route used in the car chase scene in Bullitt, though not as fast.
Fully automated cars have been legalised under Nevada law and he suggested that they are going ro be considerably safer (and somewhat more fuel efficient) than human driving and thus will, in the not distant future, become a major element of road transport. I have written previously of automating rail transport which is obviously many orders of magnitude simpler.
While a fully automated armoured vehicle does exist, at a cost of $80 million the major area for such craft is in the air. Again something I have said before. Currently these are drones being used in Afghanistan, Pakistan and Somalia, controlled by people sitting in California. However the potential for making them fully automatic exists and indeed if they are ever to be used against a technologically advanced enemy, who will easily be able to jam signals, this becomes a necessity. The ethical problems of drones, making war safe, at least for the attacker, are obvious in the real world. The ethical problems of having robots making the decision to kill are clearly far greater.
This is the HTV-2 hypersonic glider which can be fully automated and at 13,000 mph is virtually a spaceship.
Population ageing is not simply a western problem. Japan is ageing more than us and he told us that 2/3rds of the population of Shanghai is over 65. The major role here seems to be not so much pure robotics as enhancement of failing bodies. For example this HAL suit by Cyberdyne (really) is far more versatile than a wheelchair.
This Hello Kitty robot is abnle to converse with children; look ; warn them of danger and alert paresnts. The children really enjoy them as friends. We also got a quote from a mother who was really pleased that her work wasn't beiong interrupted by her child ctying because he was lonely. The ethical problems of having children thus socialised only with robots are obvious. On the other hand so are the problems of having such a mother.
In the Q&A section I asked about robots potential as a force multiplier in space development. He alsered that at present NASA are uninterested because robots weigh so much. However I am sure that as the market develops the weight will reduce and in any case remotely handled vehicles will not weigh appreciably more than conventional ones. It seems to me that robots which are stronger; able to work in vacuum; under high radiation & temperature; and 24/7 can do far more work than people and that they could turn a space community of 1,000 people into a manufacturing centre equal to 1 million people. While the lightspeed lag of about 3 seconds to the Moon might cause problems roboticisation seems sufficiently advanced that the Earth based handler need not wholly control it. This makes Moon industrialisation, and even thay of the asteroids feasible.
Afterwards I asked him about one of my hobbyhorses - not robotics but something in the direction of AI - a computer programme acting as a judge in international law. He did not think it would work, pointing out that even human judges are not good at consistency and mentioning an international meeting he had been at where there was vigorous argument over the meaning of 1 word in an agreement. However my thoughts are for something simpler than a programme, on its own, interpreting witness evidence and more on using the relatively simple and limited precedents of international law to determine whether or when wars and interventions are legal.. I remain convinced that this involves fewer decisions than driving across Nevada.