The Execution/Fruit Mystery are two games that are minimally interactive and don't have any clear goals. A lot of people would say they're not games, perhaps more would say they're insolent, pretentious wastes of time. But I'm to experimental games as Jesus was to prostitutes, I like them even if I don't buy in.
I think trying to lasso a boundary of "game" and "non-game" is not very useful at all, maybe even harmful. What happens is you get things that are game-like but become classified as "non-games" in the affirmative, like how nothing can be something... man. Except they're not nothing, and they may have value to people outside the conventional audience.
What's ironic is that both of these games use procedural rhetoric that revolves around the meta-game experience, in other words instead of trying to use rules to make a point, they use a constraint about how you have to engage the game. For Execution, you're given one verb, execute, and then sort-of half-guilted for doing the only thing you could possibly do with it. For Fruit, it insults by default while your performance doesn't make a difference. (FTW, Fruit was kinda funny at the end.) It's really easy to say these are cheap cop-outs, one-offs that should be put in the dumpster and ignored after use, like disposable diapers. And maybe they should, I don't want that nasty diaper lying around.
But consider what they can teach us -- there's a dynamic outside of agency that gives meaning to interaction. This is kinda a running theme of mine, lots of games have been wholly focused on flowing agency, now we're seeing a bunch of games that are focused solely on meta-mental tricks, phantasm, whatever you want to call it. In the same way that you can dismiss a game for being generic, you can dismiss stuff like these for being artsy-fartsy, but the generic games provided and refined patterns that are straight smack math, useful to the future, and that principle works both ways.