That advice has long been old startup lore even before Michael was at YC.
But wrt literal IQ (I recall the advice being "smarter" than the average person in the room, not IQ specifically), I think the intended meaning is more "cognitive ability" or i.e., ability to learn quickly.
I think this is why most early stage startups can do well with mostly strong junior engineers and a lesser percentage of seniors to oversee the cloud engineering concerns like scale, architecture, infrastructure, security, etc.
I made a decision to just suck it up next year and practice algorithms. I have no problem getting offers from average companies. However, all the big dogs want data structures and algorithms. It doesn't feel right for me, but oh well.