sajam is a smart dude on broad analysis, though the particular drum he's been beating of "fighting games aren't inherently harder to learn than other genres" isn't entirely true. maybe for RTS games or tactics stuff since they have similar or greater level of mental stack required to play well
even the most complex fps (cs) or moba (dota) doesn't have as much genuinely relevant contextual, abstract information to learn as fighting games. abilities in mobas can be difficult to use (or have weird interplay like in dota), and fps's can require both good aim and interior map knowledge, but these aren't all that abstract. the basic mechanics work like you'd expect in most cases
whereas just the idea of counter-hitting in most major fighting games is a lot to wrap your head around. based on the speed of your attack versus the speed of theirs, the opponent will suffer a longer window of vulnerability than usual, allowing you an opportunity to go in for more damage
that's just one example, and that still asks you to carry a lot of contextual information (what is a frame trap? what attack is an opponent likely to use? what's the frame data of these animations?) in addition to learning to react in the situations this mechanic creates. basically, awkward to describe and tough to learn. I know how this works and yet I still have a lot of trouble reacting to counter-hit opportunities
in any of these other games, it's easier to do well enough that you have a reasonable idea of what's going on. for fighting games, there's a lot of contextual state changes and it takes a long time to get a handle on that
