John Holland (laughing): Oh, that's good. (He speaks from a podium and uses overhead slides.)

We've been asked to be provocative and it's usually hard to do that in five minutes without swear words, but I'll try.

Phil Power started this out by saying that the Marshall Scholar plan and indeed the Marshall Plan as a whole was our preeminent example of a far-horizon commitment. That was a hard thing to do then, and it's much harder to do now in an age in which a CEO can be in fact fired with three or four quarters of bad performance. And that holds not only for CEOs; it holds for universities as well.

I want to say first a word about why that happens, and this is all very familiar to you, I'm sure, but let me point out that we have this process where we have resources flowing into a system (shows first transparency). We have usage of those resources, and we have this strange back loop called investment. This kind of thing, by the way, is typical of what we do when we try to study systems. We have a lot of complication. We form some simple flow model of that complication, and that process is much like political cartooning. You drop a lot of detail and you exaggerate a few things that you want to study closely.

This takes us to the following simple sequence (shows second transparency). We start out and we have this reservoir of capital or knowledge. We have the flow back which continues to add to that, and we have this thing we call the bottom line, the net flow, the profit, the output. Now one way to make that bottom line look a lot better is to cut off the flow or the investment in research or capital. And then suddenly we get a nice balloon, things look a lot better, our quarters look good, and then shortly thereafter, we run out of this capital and we wind up with a diminished flow. The problem is that we seldom look beyond this stage (points to an area on an overhead slide that shows the "nice balloon"). This has happened repeatedly. For those of us in this general part of the United States there was a great railroad called the Penn Central. The Penn Central used exactly this technique to show some good years. As a result of that suddenly, much to the surprise of the investors, that railroad went defunct. They ran out of rolling stock and so on.

Now, what can we do about this? And by the way, I consider this the single most important problem facing us, not only as a country, not only in industry, not only in education, not only in government, but as a world right now. This business of how do we get the far horizon into the bottom line. If we can be effective at that, a great many of our other problems will be solved, whether its population or deforestation. The question is how to get it into the bottom line, and I have two quick suggestions.

One is - and this goes back to what Vinton Cerf said yesterday - one is this business of modeling, and in fact video games are a prime example. They hold our attention, they let us explore scenarios or consequences, and if we can somehow implement this in a much more networked way so that many people can explore many scenarios and then we can get the information back to some central group, we have a chance of looking at these alternatives and seeing where they lead. Now that depends on a lot of things. It depends on an ability to model, to draw these cartoons, and it depends on a lot of people being able to do this so that we get a lot of ideas of where things might lead. And in the end it has to be convincing, of course, as a political cartoon is.

The other direction comes from some conversations that I had with John Seely Brown. That is, the notion of options, of somehow making a bet on the future and paying for that bet. Now, the thing that I know that exists that comes closest to that - the problem here is to do what they call commodotize this, to make this something that actually pays into the current bottom line. The New Zealand general lottery actually uses the lottery to help support its research, and my final suggestion here is if in fact we could institute a lottery, a very popular device which gathers a lot of money, in terms of making bets on the future of particular corporations with the payout determined in the long run by how good the bet is, that is how good the corporation really does. If we made this, if we used the Net to implement such a lottery and we in fact made the input to the various lotteries so that Xerox or whatever, IBM, received the money for the lottery, or at least a portion of it, and then the payback in the future over the 20 years or what they usually do on these lotteries is, in fact, related to how good the bet is. Then we would, in fact, have, I suggest, a mechanism for getting these far-horizon or future-oriented events into the current bottom line. Thank you.