I find a lot of "Reddit knowledge" in areas where I need accurate information to do my job (vintage console repair especially) is complete bullshit, but it makes you look smart if you repeat it.

Almost always it requires an understanding of some sort of niche knowledge/art, like metal fatigue, heat transfer or potentiometers. At the same time, it's usually irrelevant to the problem at hand.

In this case, the statement is nothing more than a vehicle for showing off that you understand what integer overflow is, and therefore signalling to others how smart you are.

And that said, I've probably repeated the "Nuclear Gandhi" myth to a dozen other people, complete with a short explanation of unsigned integer overflow.

Well, no. It's not something you could believe if you ever played Civilization with Gandhi. (Which is itself relatively uncommon - he can only appear in half of games that are configured with 6 opponents. With fewer than 6 opponents, he can't appear at all.)

I almost never played 7-player games, but I can tell you from long experience that it's not plausible for any civilization to be more aggressive than the Zulus. For Gandhi to be "25 times more aggressive" than the other civilizations in the game, he'd need to start breaking peace treaties several times per turn as soon as he developed Democracy.

And even then, Democracy arrives long, long before nuclear weapons. Gandhi would never have been noted for an affinity for nukes[1] - he'd have been noted for an affinity for stabbing you in the back with horse-mounted knights.

The nuclear Gandhi thing makes sense as an example of someone who is completely unfamiliar with a system making up behavior that he thinks sounds interesting. But not as an example of people detecting false patterns. The true patterns are too obvious.

[1] Nukes aren't likely to come up in a game of Civilization in any event. They appear too late; you'll win or lose before they're relevant.