"I've been working on logical AI since 1958," he says, "and I've done some I think good work (and other people also), but still we don't have human-level intelligence yet. I can't predict any definite date at which it will be achieved, even though Ray Kurzweil is eager to say it will happen by 2029. If I live to be 102 and am still capable of laughing I expect to laugh at him then."
He sees no evidence, either, for that science fiction staple the Singularity. Even so, he remains optimistic. He points to genetics as an example. Mendel laid the foundations in 1865, but the genetic code wasn't cracked until the 1960s, And even now, "Still we don't know how genetics controls the shape of an animal. So that's taken even longer than I have so far. Hard scientific problems are hard."
McCarthy's less formal interests are wide-ranging: human expansion beyond the earth, the search for extraterrestrial intelligence and, the topic he most wants to talk about, sustainability.
Partly, he attributes this to his background. "I was brought up as a communist," he says. As a child, he read a translation of a Russian's children's primer, 100,0000 Whys, which, he says, "was enormously optimistic about technology". (McCarthy became a liberal in 1952, then, in 1972, a Republican). Attending local colloquia on the search for extraterrestrial intelligence has sparked thoughts about what kind of intelligence would likely exist.
Take, for example, the difference between octopus eyes and vertebrate eyes, which do the same job but are completely unrelated genetically. "It says the world presents the same kinds of problems to animals and plants, and sometimes they solve them, and even things with quite different backgrounds may come up with the same solution." Applying that to SETI, he imagines that the networks of neuron in which intelligence is implemented on earth might exist in other solar systems where intelligence evolved differently – though that intelligence would still be able to do arithmetic and know mathematics. Sustainability occupies many pages of his Web site. ("Why," he asks parenthetically, "won't Who's Who let you list your Web address?") The gist: "Doom-saying is mistaken, and material progress of the kind we've had in the past is sustainable for the foreseeable future, contrary to public opinion."
McCarthy's site assembles evidence to back up his contention on a range of topics: energy, food supply, population, water, forests, wood supply, pollution and biodiversity. "My argument that these menaces will be overcome does not depend on any futuristic technology. It depends on the technology we already have." Although, he adds, "Naturally there will be futuristic technology."
This sort of optimism is at the very least unfashionable at the moment, but McCarthy believes the evidence supports his point of view, noting that the average time frame for disaster predictions is two years. "I would say that the area where I'm weakest is erosion, because the estimates of the rate of loss of topsoil by erosion that I've found differ by a factor of 100 from each other." Overall, he says, "The key thing that gives rise to confidence is that plenty of energy is available, most specifically nuclear energy, but solar probably will also work."
McCarthy, who for a time early in his life read a lot of science fiction, is, however, willing to consider solutions that environmentalists generally would hate. For example, when William Calvin predicted in Atlantic Monthly in 1998 that the Gulf Stream would stop, he asked his students to calculate the cost of covering Western Europe with transparent plastic, making it a giant greenhouse.
"Calvin and the doomsters in general do not write counter-measures to the disasters they are predicting," he says. "There is an actual prejudice against global engineering on the grounds of interfering with nature."
Overall, he says, "I like to think in terms of opportunities rather than inevitabilities. So technology will offer that opportunity. Who will take it is another question." µ
Uses 20 percent less power than traditional systems
It's becoming more prevalent in car research and development
Sign up for INQbot – a weekly roundup of the best from the INQ