But it didn't do any good to spot a steep slant in his formulations. He was a salesman, after all. Just so he stayed clear of demonstrable falsehoods and "fraudulent tendencies" (as defined by the 1978 Commerce Act), he was within his rights. He was staying clear. Some of his claims a stickler might want to check up on; but I wasn't going to bother any more to watch for things like that. I thought the stickler would find in each case that he'd been wasting his time. This little salesman seemed awfully good at skating just at the edge. He really knew his profession. I didn't let my bafflement show. I just looked at him dully and made noises as if I was about to say something. I was, but I didn't know what. There just had to be something bad about this Statistomat venture. Without (apparently) any new gimmick, a small new company was producing just as good a product as one of General Computers' best-managed divisions. How could Statistomat hope to deliver a normal profit? It wasn't reasonable. There must be badly cut corners, if not in the product then in the sales program or the servicing of customers; or else the investors _weren't_ hoping for a normal return. In that case there was something funny in their motives--a long-range scheme to undermine G.C., or something. That might show up in this salesman's pitch. So I switched to, "How do I know what stocks this thing'll tell me to buy?" "Not _tell_ you to buy," he corrected charmingly, "_buy_ you. The machine can be connected by direct wire to the Exchange's computer." "Yeah-yeah, but how do I know what stocks I'll be getting? I want General Computers preferred!" He smiled. "Quite possibly you'll find yourself the owner of a considerable block of G.C. preferred--provided of course your time-dependent utility function dictates a policy which--" "You mean," I said, with the very suspicious expression my brother always objected to, "you'd let your machine bid for G.C. stock for me?" "Naturally. The Statistomat has often recommended purchase of G.C. stock. Let me explain to you an aspect of modern firm management which may be so specialized as to have escaped your attention. "Each firm draws up what is called a preference function. It is somewhat analogous to the investor's time-dependent utility function. It gives exact expression to the objectives of the firm. For any conceivable economic position the firm might be in, it determines, let us say, the weight the board places on a dividend this year as against a larger dividend a year from now, or ten. And so on. It is the criterion for all the optimization computations which pattern the firm's activities.