Birman: Not in my view. My machine is fast enough. Trustworthy computing is a much more urgent issue. After all, if we move all these critical systems to computer networks and don't solve the trust issue, we'll be cooked, and faster machines are really not going impact that issue one iota.

Chazelle: We're at the tail end of Moore's Law. Because of power-dissipation issues, the party's essentially over. Back to parallel architectures. Huge work ahead.

But this will be the best thing that can happen to CS. Moore's Law has been tremendously beneficial to society. At the same time, it's been so damn powerful that it has set back the development of algorithms. But that's about to change. Any student interested in science and technology needs to learn to think algorithmically. That's the next big thing.

Dally: The current road maps show [Moore's Law] continuing for at least another 15 years. Even after that point, it will take many years for information systems to fully exploit the densest semiconductor devices. One can also expect that other technologies may emerge to continue scaling -- perhaps in a different way.

Canny: We're still stuck in the paradigm of "more cycles, more value." Now is the time to really start looking around. The market for computing has changed radically and irreversibly. What really matters is how useful our artifacts are to the people who are buying them. We don't know these people anymore, and we don't care much what they're doing with IT. That has to change. We're at an economic cusp right now. Sales have slowed, jobs have slowed, enrollments have slowed. It's not a technical problem; it's a problem of creating value. Every mature industry has to face that eventually. We've had a fun childhood without many cares, but now we're grown, and we have to figure out what we can do for the world that really matters.