In 1988, the computer scientist Hans Moravec observed that computers already excelled at tasks that humans tended to think of as complicated or difficult (math, chess, IQ tests) but were unable to match “the skills of a one-year-old when it comes to perception and mobility.” Six years later, the cognitive psychologist Steven Pinker offered a pithier formulation: “The main lesson of thirty-five years of AI research,” he wrote, “is that the hard problems are easy and the easy problems are hard.” This lesson is now known as “Moravec’s paradox.”
The paradox has grown only more apparent in the past few years: AI research races forward; robotics research stumbles. In part that’s because the two disciplines are not equally resourced. Fewer people work on robotics than on AI. There’s also a funding disparity: “The flywheel of capitalism isn’t spinning fast enough yet in robotics,” Jang told me. “There is this perception among investors based mostly on historical data that the payoff of robotics investments is not very high.” And when private companies do put money into building robots, they tend to hoard their knowledge.
What are folks thoughts on this paradox, as well as the hypothesis for supposed development disparity presented?
“computers and language” are fully digitally representable, and works well when tethered to a wall for power. Anyone can use a $2000 laptop to do real development, and that laptop can also be used for regular work, school work, and entertainment.
“mobility and perception” interacts with physics, gravity, inertia, and needs independent power. The cost of experimentation for a useful biped starts at maybe $100,000, and other than serving as a test platform and conversation piece, it can’t really do anything else for you. Fewer people do development with these resources, and thus it progresses much slower.
And, because interacting with the world is much harder than interacting with bits, I think this discrepancy won’t go away.
I mean. I love the mechanical analogy. But generally, it’s definitely true that research in robotics requires more resources than AI, because there is simply more HW involved. If I am not mistaken, the whole motivation behind robocup was “Okay, AI can beat humans in Chess, now let’s beat them in Soccer.”. And that this is far from happening any time soon is in part due to the immense costs in upkeep of one, let alone 11 humanoid robots. I have no good data on whether robotics is underfunded in comparison to AI, but I guess the same money can buy you a lot more GPUs than robots.
On the other hand, I would also argue that a universally capable robot would yield a lot higher productivity than the next big language model, simply because we still live in a physical world.
But,
is probably more a property of private companies than of private companies building robots.