Archive for March, 2012
We just returned from a robotics exhibition in San Diego. That was not the long strange trip I had in mind however. Okay, the robot (Vigilus-MCS) that we brought along had a long strange trip – including a short stop by the security team at the Hoover Dam, but I was thinking more about the trip that we began about 4 years ago, when we started on the current robot platform. The philosophy at Gamma Two has never been tightly aligned with the current trends in research. We are not big on large, all encompassing theoretical research. We focus on solving real-world problems. Sometimes that means we need to do cutting edge theoretic research to come up with the solutions.
We were driven by a question when we started this project: “Why don’t we have robots working along side us every day?” Or, to put it more simply “Where’s my robot?” That led to several years of theoretic research, which culminated in two things. The second was our technical book “Robots, Reasoning, and Reification,” which summarizes the open problem that stood between us and functional robot co-workers. The book outlines a theoretic solution to the problem. I suppose that we could have stopped there.
But, like I said, we are out of step with the pure research community. We knew that this was the case, since we have been active members of the Performance Metrics for Intelligent Systems (PerMIS) community for over a decade. This conference is run by the National Institutes of Standards and Technology (NIST), they have an interest in measuring the ‘intelligence’ of intelligent systems. For the last several years, Louise and I have been on the program committee for this conference. Every year we would be part of a group of researchers pushing the limits of what we know about making computer and robotic systems smart. We would see dozens of new and untested theories every year. But, we weren’t willing to propose a theoretic solution, unless we knew that it worked.
So, the first thing that came out of the years of research was a functioning robotic brain, one that enabled the robot to see the world in a way that is similar to the way living systems see the world, and reason about the world in a way that is similar to the way living systems reason about the world. Of course the brain wasn’t complete, but we had a ‘proof of concept’ prototype that people could work with. It didn’t do much, but it did things in the right way. There’s a short video of the robot updating its model of the world, and keeping track of objects. This was done in November of 2008, and shows the first generation proof of concept robot. When I compare that system with the robot we took to San Diego, I realize just how long a trip it has been.
Up to this trip we have been focused on researching the theory of robotics, then the practicality of developing a robot. Then we spent enormous amounts of time and energy looking for money to fund the R&D. But we turned a corner this month, a big corner for the business. We were invited to pitch our company to the assembled investors of the Angel Capital Summit, put on by the Rockies Venture Club. We were also selected as one of six companies to present “cutting edge technology” in ground robotics to the buyers at the NDIA Ground Robotics Conference an Exhibition. They both occurred on the same days, last week. The same days that the PerMIS 2012 conference was running.
This is where the long strange trip really became apparent. Five years ago, there would have been no question, we would have been trading ideas with some of the best researchers into machine intelligence in the world. Two years ago, we would have been pitching our hearts out to a group of investors, in hopes that one of them might be interested in becoming part of the ‘next big thing’. This year there was no question, we were going to be presenting our robots to the people who can buy them and put them to work, making lives better.
Like I said, it has been a long, strange trip.
Where’s your robot? Ours are being built by Gamma Two Robotics, here in Colorado.
We were doing a presentation this week, at an angel investor meetup. The presentation on our security robot business model went well, although their format forced us to use only 5 minutes. So we spent hours paring down the words to fit into 300 seconds. A great exercise, and well worth the effort. Of course, after the lightning speed presentation, came the questions.
I enjoy the questions, I like the back and forth that occurs – it is more like a conversation than a lecture. But there is a moment I dread. It almost always comes up towards the end of the question period, and, perhaps, I am to blame. I do not come across as a formal speaker, I think that we communicate more effectively if we can communicate informally. I think that that mode encourages thought and engagement, and opens pathways for a more free-ranging conversation. Which, when talking about intelligent, autonomous robots, seems to lead, inevitably, to the question.
It takes many forms, depending on the speaker and their cultural refferents, but the thrust is always the same: What about Killer Robot Overlords? This night it was in the Terminator motif; “Jim, what if the robots ask ‘Where is Sarah Connor?’” If the speaker was older it might have been a direct Terminator reference, or perhaps “Klaatu barada nikto”
But, regardless of the exact form of the question, it always comes down to “What about Killer Robot Overlords?” It is kind of weird, I mean, we build robots – design their brains, I know just how smart they are. And they are a looong way from being as smart as a person. But, more importantly, why would they become killer robots? Of course, we, as people, are simply projecting our fears of technology and change onto the hardware. But where did it come from?
Robots have not been with us very long, but the idea of robots is very old. It goes back beyond the word “robot” which came into use in the 1920’s. The first use of the word is generally ascribed to Karl Capec’s play “R. U. R. – Rossum’s Universal Robots” in which humanoid autonoma were used as workers, and eventually rebelled. This was generally believed to be an extension of the Jewish Golem mythos – servants crafted out of clay and instilled with pseudo-life. The servant rebellion was a common theme, which, when brought to robots, leads to the robot rebellion, which seems to inevitably lead to killer robot overlords. But even this model is a relative newcomer.
In the Greek mythos, Hephestus, the blacksmith to the gods, manufactured robotic carts that carried food and drink around the home.
Here is photo of one of our robots doing a job that has been around for perhaps 3000 years. In addition, Hephestus made robotic ‘golden maidens’ to work the parties as well. No sense of Robottic Overlords in ancient Greece, these were reliable, helpful assistents, that help people (well the Gods at least) live better lives.
So, that is the concept that leads us to our vision:
Creating a world in which people live better lives assisted by affordable, reliable, helpful robots.
No killer robot overlords will come out of our shop.
Where is your robot? Ours are be made by Gamma Two Robotics