Advice From Seymour Cray

They were giving away DVDs full of historical footage from past Supercomputing conventions last year.  If you’ve ever been to one of these things, you know that half of the experience is grabbing up as much free stuff as possible — so I did my duty and took a few.

After the conference, I popped one into my computer, not expecting to find anything of interest except to longtime veterans and insiders.  But there was a real gem on there, the keynote address from (I think) the first convention, back around 1988, by Seymour Cray, “the father of supercomputing”.  It was a real delight, even if he did meander off into old man stories more than a few times.

I bring this up because he shared an anecdote that really resonated with me — following a somewhat similar path as he had in his early career, I had a similar experience.  So my retelling will probably be a mix of his words and my thoughts, as they’re hard for me to separate.

He had just graduated from college, and had to find a job.  His adviser told him to check out the glider factory down the street, so he did and managed to land a job there.  It turned out that the glider factory was trying to develop computers for the Navy, and they started assigning him tasks that he really didn’t understand, but was just working through anyway, half-blind but managing.  His first realization was that he had no freakin’ idea what he was doing.  This came pretty quickly.  And boy can I ever relate, being assigned all sorts of crazy problems fresh out of undergrad with no clearly defined sets of tools to use, or even a consistent theoretical framework within which to understand them.  It was, and is, tough-going.

Some time later, a few days or a few weeks but no more, Seymour Cray had a second realization: nobody else at the glider factory knew what they were doing either.