Break up Intel. Put Krzanich in charge of the fabs. Split manufacturing from design and marketing. Open up all of Intel's design resources to customization by customers. Go directly after the big fabless competitors -- like Qualcomm ( QCOM) and Broadcom ( BRCM) in communications, Nvidia ( NVDA) in graphics, ARM Holdings ( ARMH) in low-power design. If Intel's fabs were separate from its design and marketing, the company could go after Apple's iPad chip business with a clear conscience. Apple is looking to get away from Samsung, with which it's fighting a long war over patents and a life-or-death one over device share. Intel's fabs could grab that contract and, at a stroke, that would more than pay for the whole deal. Intel's problem is that it's too integrated. Its marketing team is devoted to selling only Intel designs, and are divorced from the real world of what customers do with its chips. Its designers are producing on a single road map for the foundry, and give customers a take-it-or-leave-it choice. Device makers of all types, facing such a choice, would just as soon leave it. But what if that weren't the choice? What if customers were free to customize Intel designs, with Intel's help, and make them their own? What if Intel's hiring were aimed at finding people who could drive that software train, instead of just hardware guys? Moore's Law is a roadmap, but it's just one road. What the chip business has shown, in this century, is that there are many different roads that can be taken. Intel has missed so many of them. I don't see what, save a break-up, can get it onto those multiple paths of profit. At the time of publication, the author was long INTC but not nearly as long as he was a month ago.This article is commentary by an independent contributor, separate from TheStreet's regular news coverage.