Bob Colwell gave a great talk at FCRC on Tuesday night. A large part of it focused on the big challenges facing computer architecture in the next few years: multicores and parallelism, thermal issues, and process constraints. But he also talked about the bigger picture, why fewer and fewer students are going into computer science. He made a great point — we don’t have any grand challenges. A recent panel had come up with “parallel programming” as our grand challenge, but seriously, who aside from computer scientists really cares about that? A grand challenge should spark interest from normal people and should be something that really makes people want to work in our field.
I think a great grand challenge is AI. Not AI in the computer science sense of searching and planning and such, but AI in the Turing test sense — an intelligent machine. Clearly the public loves this idea; it comes up all over movies and literature. Do we know how to do it? Not at all, but maybe that’s the point of a grand challenge. A pie-in-the-sky idea that inspires innovation.
