It feels like you switch around how you use snobbery. At the beginning you give the definition "with this belief, I eliminate a class of problems other people have."
To me, that usage doesn't seem like it has much to do with the standard notion of snobbery, but that's not a huge deal. Though with your tiping and hacker news example, it seems like you go back to using snobbery to mean holding "elitist beliefs ".
Also, it didn't like you actually made a case for the general utility of holding snobish beliefs (if that's what you were trying to argue for). The tipping example seemed to be making a particular case for the act of tipping, that didn't seem related to the fact that tipping well could be considered a classic snobish behaviour. In the hacker news example, it seems you were pointing out that if you needed to win over people who had free-floating snobbish beliefs, that it can sometimes benefit you to shout the same slogan as them.
Epistemic effort: I wrote this on a plane flight. I'm often interested in Ribbonfarmian "consider a bad thing. What if it's good? (Here's my favorite example of this.)
As regards updating my beliefs, I'm drawn to motivated snobbery. "Motivated" means "this belief improves my experiences;" "snobbery" means "with this belief, I eliminate a class of problems other people have."
An example of motivated snobbery is "tipping well." Here's my sales pitch: Tipping is an iterated prisoner's dilemma, not an evaluation of their performance! I want servers and bartenders to be happy to see me. A friend said this well: If I'm getting rich we're all getting rich.
That this makes bartenders happy and signals to my friends that I'm atypically generous[0] makes this "motivated." That I'm implying, "The extra money I spend doesn't hurt me" signals abundance; I can't be pressured by what pressures others. This is snobbish.
Some nonexamples for me are the slogans-as-beliefs you see on hacker news: "Machine learning is just multiplying matrices." "Meta-Language is the real ML" "Once we know how to do something with computers, it ceases to be AI, and becomes boring." Though snobs say these things, I don't because I'm not (right now) trying to hire an engineer who agrees with those things[1].
A reasonable objection: "So zlrth, you'll change your beliefs depending on your circumstances?" In a trivial sense, yes. To be a motivated snob you must keep your identity small and avoid mind-killing. Those things suggest not being an ideologue[2]. I don't think avoiding mind-killing is controversial (at lesswrong, anyway), but if you do, you should let me know. What I think is more up for debate is what heuristics you use to decide your rent-paying beliefs, and their end goals.
Another objection: Things HN'ers say are mostly phatic; not beliefs-that-are-supposed-to-pay-rent. Asked, "What kind of thing is, 'Meta-Language is the real ML'?" more people would say that it's an opinion than a belief. I'm not sure what opinions are; I'd rather have rationality-style beliefs. You should elevate your opinions into snobbish beliefs that eliminate for you classes of problems other people have, or dismiss them.
[0] There's a cynical read: I, an alleged sociopath, get off on being perceived as formidibly rich or risk-tolerant. I don't have a ready response. You should try out motivated snobbery; see how you feel : ). I don't feel sociopathic when I do it.
[1] Perhaps there's value in virtue signaling elitist beliefs about machine learning. I think there's more, better virtue-signaling in declining to have a belief in the first place.
[2] I don't think there's anything wrong with ideologues, and I apologize if I sounds like I'm strawmanning them. My point is that having a small identity is a reasonable, uncontroversial thing to do, and a prerequisite for motivated snobbery.