Project Halo
Paul Allen's Vulcan Ventures is funding Project Halo, a competition among 3 teams to try and synthesize all modern textbook scientific knowledge into a computerized tutor. The goal of the three teams is to create in the next 30 months "a computerized tutor that's smart enough to pass college-level Advanced Placement (AP) tests in chemistry, biology and physics."
Sci-fi fiction fans will immediately think of A Young Lady's Illustrated Primer, the world's coolest Choose Your Own Adventure book/supercomputer/encyclopedia, from Neal Stephenson's classic novel The Diamond Age. While the technology is clearly quite raw, the end vision is compelling. If human knowledge could be synthesized by a computer that could then interpret requests and respond with reasoned answers based on that knowledge, education could capitalize directly on Moore's Law and Gilder's Law. Instead of trying to solve the world's education needs by educating and then distributing human teachers to every child on Earth, one could take the knowledge from the world's top educators and distribute it over a computer network to clients like computers that children could use.
The cost of training enough great teachers and motivating them to fan out to the ends of the earth is massively expensive and suffers from few economies of scale. The cost of doing so hasn't decreased with time, and it doesn't decrease with volume either. On the other hand, if some combination of software and hardware could simulate a teacher, the cost of replicating that software/hardware combo and distributing it all over the world would decrease over time. Such a program might never be a substitute for a real human teacher, but it would certainly be an improvement over no teacher at all.
It's still a pipe dream. Cracking the artifical intelligence challenge is, in my estimation, vastly more difficult than synthesizing the knowledge (we already have a crude prototype of the latter in operation today: Google + WWW). But I believe a solid prototype can be built in my lifetime, and that would be a society-changer.
Sci-fi fiction fans will immediately think of A Young Lady's Illustrated Primer, the world's coolest Choose Your Own Adventure book/supercomputer/encyclopedia, from Neal Stephenson's classic novel The Diamond Age. While the technology is clearly quite raw, the end vision is compelling. If human knowledge could be synthesized by a computer that could then interpret requests and respond with reasoned answers based on that knowledge, education could capitalize directly on Moore's Law and Gilder's Law. Instead of trying to solve the world's education needs by educating and then distributing human teachers to every child on Earth, one could take the knowledge from the world's top educators and distribute it over a computer network to clients like computers that children could use.
The cost of training enough great teachers and motivating them to fan out to the ends of the earth is massively expensive and suffers from few economies of scale. The cost of doing so hasn't decreased with time, and it doesn't decrease with volume either. On the other hand, if some combination of software and hardware could simulate a teacher, the cost of replicating that software/hardware combo and distributing it all over the world would decrease over time. Such a program might never be a substitute for a real human teacher, but it would certainly be an improvement over no teacher at all.
It's still a pipe dream. Cracking the artifical intelligence challenge is, in my estimation, vastly more difficult than synthesizing the knowledge (we already have a crude prototype of the latter in operation today: Google + WWW). But I believe a solid prototype can be built in my lifetime, and that would be a society-changer.