I knew him very well and we had many conversations. I was one of the founding directors in the Robotics Institute at CMU and was awed, but not by him, but by his engineering capability.
I once had a new IBM workstation in my lab (one of the first of a new type) and he asked me for a screwdriver, and smiled. I got my lab manager to get him whatever tools he wanted.
A few minutes later I walked into the lab. He was on the floor with the machine apart, and notes on a pad. He was taking part numbers. I asked him to put it back together before he left, and he did. My lab manager confirmed he powered it up and it booted properly.
I was not awed by his reality distortion field. I was awed by his product engineering competence. The comment I made in this article is because I know with great certainty if I asked him if he knew the evidenciary value of his iphone, he would have just smiled knowingly at me.
When I build things I am also rarely surprised about things other people did not think the things did.
My current interest is in human lying out of a computational cognitive neuroscience interest. Here is the description of that product engineering to reveal how human lies can be proven to other people and the scientific basis for it.
That said, like a lot of people I have known, I liked him, despite some very problematic faults he had. But he did not play people for fools, like Trump does. He did not lie (except to certain fools who frankly deserved it).
