I have to type in some non cap letters here, otherwise the server won't let my quote pass. It is not my fault it's all in caps. That's the way it was written the first time!
Historically, there are three major gray areas where Asimov's laws break down.
The first is long-term harm/benefit ratios. The robots that Asimov writes about are incapable of seeing long-term benefit (like scientific advancement in exchange for a minute risk of radiation poisoning--see the short Little Lost Robot from the I, Robot collection). So in this case, Asimov's robots would be incapable of striking a human, but they would be capable of acting as a bag-holder.
The second is emotional trauma from dissapointment (see Liar--same series). Asimov's Robots are incapable of understanding that dissapointment is a necessity for humanity. Once they realize that we are hurt when someone lets us down, they will lie, cheat and steal to ensure that we don't experience that sensation.
And third is long term vs. short term human harm. Asimov's "study of potentials" excuse might overcome this test case, but I doubt it. Within a city center, a gigantic weight is suspended over a human being, and set to crush him at T=12 seconds. A nuclear device is primed to go off 5 seconds after the weight crushes the person (T=17). The robot is placed at the midpoint between the two devices and given a top speed such that he is 10 seconds away from each. The robot can save the man, but in doing so the city will be destroyed. Or it can save the city, but only AFTER the man has been horrifically crushed. The delay during which the robot must do nothing about the man-to-be-crushed is a question that Asimov never (to my knowledge) explains fully--his robots exist entirely in the immediate sense. Historically Asimov's early robots experience potential failures at these points. His later robots supposedly were carefully engineered so they could overlook these sorts of problems, but never in a way that I really found satisfactory.
For the boxing robot, a more useful set of laws would be Mark Tilden's three laws of robotics: 1. Protect Thy Ass, 2. Feed Thy Ass, 3. Move thy ass to better real-estate. As far as blocking goes, the protection law will get the robot to do that. And to make it spar effectively, if the robot is following Tilden's laws and is informed that it would only be allowed to make use of a power source if it successfully sparred with the opponents provided, it would be forced to fight in order to feed.
Of course, the danger of Tilden's laws is that they create extremely aggressive robots that don't care whether their masters live or die as long as they are healthy, happy, and enjoying themselves.
Yes, but as far as I can tell from the Wired Article (which seems to intentionally obfuscate which years it is discussing in the article at certain times), all of the talk about pre-planning concerned CMU's strategy for 2004, and had nothing to do with CMU's *or* Stanford's strategy for 2005.
Besides, if we're going by bang-for-buck, Stanley doesn't deserve the award either, Team Gray does. After all, they did the race in just 7:30, on a budget of next-to-nothing compared to the big universities, and with no academic lab to back them up or the ability to call on hordes of student[*cough*slave*cough*] labour. Funded by only Gray Insurance and with gifts from a couple of parts vendors. They didn't even get the vehicle donated. If anybody counts as the competition's real underdog, it is those guys.
Stanley's official time was 6:53 and CMU's was 7:04 minutes.
I don't think that ridiculing CMU as having a "poor strategy" for doing something in an additional 11 minutes that was impossible for the entire robotics industry just a year ago is very. . . wise.
Personally, I'm overjoyed that Stanley won it. I think he's an excellent system and that Stanford deserves the praise. (Besides, those b*stards at CMU didn't let me in for my undergrad)--but making fun of their 2004 'strategy' (when they went further than any other team) and their 2005 results (when they were a scant 11 minutes behind the leader, and were 2 of only 5 teams to have a 'bot cross the finish line) seems silly to me.
Slashdot Mirror
"Durr, this key has a top, and that's a moving part, right?"
[sigh]
[Ahem]
I have to type in some non cap letters here, otherwise the server won't let my quote pass. It is not my fault it's all in caps. That's the way it was written the first time!
So, without further ado, the quote, courtesy of that haven of IRC gems, bash.org:
YES IS NOT AN ANSWER TO "A OR B?"
Historically, there are three major gray areas where Asimov's laws break down.
The first is long-term harm/benefit ratios. The robots that Asimov writes about are incapable of seeing long-term benefit (like scientific advancement in exchange for a minute risk of radiation poisoning--see the short Little Lost Robot from the I, Robot collection). So in this case, Asimov's robots would be incapable of striking a human, but they would be capable of acting as a bag-holder.
The second is emotional trauma from dissapointment (see Liar--same series). Asimov's Robots are incapable of understanding that dissapointment is a necessity for humanity. Once they realize that we are hurt when someone lets us down, they will lie, cheat and steal to ensure that we don't experience that sensation.
And third is long term vs. short term human harm. Asimov's "study of potentials" excuse might overcome this test case, but I doubt it. Within a city center, a gigantic weight is suspended over a human being, and set to crush him at T=12 seconds. A nuclear device is primed to go off 5 seconds after the weight crushes the person (T=17). The robot is placed at the midpoint between the two devices and given a top speed such that he is 10 seconds away from each. The robot can save the man, but in doing so the city will be destroyed. Or it can save the city, but only AFTER the man has been horrifically crushed. The delay during which the robot must do nothing about the man-to-be-crushed is a question that Asimov never (to my knowledge) explains fully--his robots exist entirely in the immediate sense. Historically Asimov's early robots experience potential failures at these points. His later robots supposedly were carefully engineered so they could overlook these sorts of problems, but never in a way that I really found satisfactory.
For the boxing robot, a more useful set of laws would be Mark Tilden's three laws of robotics: 1. Protect Thy Ass, 2. Feed Thy Ass, 3. Move thy ass to better real-estate. As far as blocking goes, the protection law will get the robot to do that. And to make it spar effectively, if the robot is following Tilden's laws and is informed that it would only be allowed to make use of a power source if it successfully sparred with the opponents provided, it would be forced to fight in order to feed.
Of course, the danger of Tilden's laws is that they create extremely aggressive robots that don't care whether their masters live or die as long as they are healthy, happy, and enjoying themselves.
Yes, but as far as I can tell from the Wired Article (which seems to intentionally obfuscate which years it is discussing in the article at certain times), all of the talk about pre-planning concerned CMU's strategy for 2004, and had nothing to do with CMU's *or* Stanford's strategy for 2005.
Besides, if we're going by bang-for-buck, Stanley doesn't deserve the award either, Team Gray does. After all, they did the race in just 7:30, on a budget of next-to-nothing compared to the big universities, and with no academic lab to back them up or the ability to call on hordes of student[*cough*slave*cough*] labour. Funded by only Gray Insurance and with gifts from a couple of parts vendors. They didn't even get the vehicle donated. If anybody counts as the competition's real underdog, it is those guys.
The CMU bashing here (and subtley embedded in the wired article--everybody loves an underdog) is not really valid.
According to The Grand Challenge Tracking Site:
Stanley's official time was 6:53 and CMU's was 7:04 minutes.
I don't think that ridiculing CMU as having a "poor strategy" for doing something in an additional 11 minutes that was impossible for the entire robotics industry just a year ago is very. . . wise.
Personally, I'm overjoyed that Stanley won it. I think he's an excellent system and that Stanford deserves the praise. (Besides, those b*stards at CMU didn't let me in for my undergrad)--but making fun of their 2004 'strategy' (when they went further than any other team) and their 2005 results (when they were a scant 11 minutes behind the leader, and were 2 of only 5 teams to have a 'bot cross the finish line) seems silly to me.
And for the people wondering: Stanley is rumoured to have run linux, though last I heard the team hadn't confirmed it. In fact, most of the qualifiers for the race were running at least one linux machine.