Taking the Right Technology Risks

With one very interesting exception, the only real judgment that can be applied to taking risks is, Will the product be commercially successful? If you're in business, you're in business to make money, and the shareholders are owed every effort to make a business economically successful so they can be financially successful. So you have to tune the technology risk correctly to get the right thing done for real customers who will want to buy the products and services you offer.

The exception to this is in the realm of public service. Often we learn a lot about how to make successful products by doing something for free in the public interest - things like the Internet grid that attempts to "Fight AIDS at Home." It is perfectly plausible that hundreds of thousands of PCs working part-time when the screen savers are up can actually find some drugs by using the molecular modeling software that pharmaceutical companies use or by searching through the genome information that's now becoming publicly available. We'll actually find some things just by brute force by exploiting the idle time of machines that are temporarily or intermittently under-employed, and I think that's a useful thing for the industry to support, endorse, and find ways to enable. But principally, companies won't be around to help in those kinds of projects unless they make money with what they do, and you make money only by finding markets that go beyond the early adopter phase. And to do that, you have to create technologies that make non-technical people capable of something economically valuable, in which the enabling technology is so good that it disappears.