July 30, 2006

My Billionaire Mentor

I was designing part of a computer chip, a chip containing hundreds of millions of tiny switches. I went to his office to discuss something with him, some detail of the architecture. I was met with questions, so many questions.

Most people seemed to find him frustrating. They couldn't understand what he was talking about; like he spoke another language. I think that people were scared of him, scared that he would expose their ignorance.

I don't think he had any malice. Perhaps he didn't care what the answers were, but he cared about the questions. How does this work? What happens when someone does that to it? They were frustrating questions; he made me angry. But I stayed, I listened, and I learned.

So many people dismissed him, said he was crazy, and paid no attention. Some took his questions personally; I did too; I beat myself up with his questions. I hadn't yet realized the value of a question, whether answered or not.

He thought abstractly about practical problems. He held a map in his mind; a grand-unified theory of computer systems. When something arose which challenged his map, he re-drew and refined it. When others hurried around with a thousand different maps, he could see one in their place. Because he could see the wood when others were examining the trees, he could identify problems very early.

His architecture dealt with concurrency, determinism, ownership, completion, sequencing, ambiguity, repeatability, and compatibility; these are powerful things to consider, things that many do not consider, questions worth a fortune.

He could look deeply and extract the essential underlying simplicity. His solutions seemed complex because they required some understanding. But with a small amount of knowledge anything he produced could be easily understood.

He did not hide behind jargon and complex theories, but an open mind and some effort was required. Eventually I understood enough to continue his work, to refine and to extend what he had done.

He retired with more wealth than could be spent by ten generations and I applied his architecture to its most challenging problem. Now that is done. Now I know the art of computer architecture.