What jumps out here is CMU as an incubator of robotics startups. The catchy name, Marilyn Monrobots Labs, distracted me from the real story, which is tHeather Knight's company as an interesting and potentially significant robotics company of the type which could put the U.S. in a leadership position as this field gains ground. (I should state that I'm talking about still-young field of humanoid robots, not the more mature arena of industrial robots.) In humanoid robots, most of the stuff I've read about in the past few years has come out of Japan, most notably Honda's Asimo. I'd add iRobots to my argument about important U.S. companies. iRobots spun out of MIT in 1990. It makes the Roomba vacuum-cleaner robot. Not technically humanoid, but I'd certainly put it in that class because of its in-the-home application. Paging Woody Allen's "Sleeper"...
Alex, I agree with what you're saying. It seems like a lot of these breakthroughs in the humanoid area at least are in the area of "fun and games" without any real world application. Now maybe the goal is, as suggested by the headline, to be solely a phsychological tool to get people used to dealing with non-humans, but I think the drawback there is that they won't be taken seriously in real applications.
Funny you mention CMU and robotics incubation. President Obama actually announced a new robotics initiative at CMU last week as part of his unveiling of the Advanced Manufacturing Partnership, an initiative between government, industry, and unversities to invest in technologies designed to jumpstart manufacturing. One leg of the announcement was a joint effort by the National Science Foundation, NASA, NIH and the Department of Argiculture to pool $70 million to fund development around next-generation robots. It will be interesting to see what comes out of that partnership.
This sort of thing has been tried before, albeit perhaps not with real moving parts. I'm thinking of the irrepressible Microsoft paperclip. Well, actually I did repress him - I turned him off. Dangit! "him?" See what I mean about anthropomorphism? We must be getting close to that GPP (Genuine People Personalities) feature Douglas Adams wrote about in The Hitchhiker's Guide to the Galaxy.
Beth, it would be interesting to know more about the Advanced Manufacturing Partnership. I know many engineers who are ideologically opposed to government spending in general, and I personally have qualms about corporate welfare. At the same time, I think that strategic government support of manufacturing innovation is a smart idea which may pay off many times the investment of taxpayer money in terms of benefit to the overall economy. I also think that technological advancement is a worthy goal in itself, at least as much as the exploration of space, scientific investigation, or art. If we are willing to spend taxpayer dollars on art museums, scientific research, and the space program because we recognize that these things, independent of their economic value, are important to the society we want to have, then why shouldn't we be willing to spend taxpayer money to advance the art of manufacturing?
It would be great to have a guide for how companies can get linked up with this type of government programs. I think many companies are not aware of the possibilities of this kind,of partnership with the public sector.
I'm in the process of reporting/writing a piece on the initiative, so stay tuned. In the interim, check out the White House press release at http://www.whitehouse.gov/the-press-office/2011/06/24/president-obama-launches-advanced-manufacturing-partnership. Be sure to let me know what you think once read the Design News article.
I think we're still pretty far from having to worry about humanoid robots affecting the cost of labor. As Alex points out, even the Roomba and the RoboMower type robots don't have human qualities...yet. In the near future, I think the most successful robots will still be those that don't mess with the "uncanny valley."
It describes some fascinating research using artificial intelligence to try to better understand schizophrenia. Essentially, the researchers looked at what parameters they needed to change in an AI program in order to get the program to respond like a schizophrenic patient would. They then tried to relate this to what might happen in an actual schizophrenic brain. At this point it's just a model, but it's a very compelling one.