Nuance and context are not the strong points of a text-only medium. If the underlying context of "works for me" is "you're an idiot," then that's a problem. If the underlying context of "works for me" is another data point that could lead to a solution, that's not a problem.
For example if someone says "I have a problem with Kingston when running plug-in X and didn't have a problem with Jamaica Plain," then it's logical for the person to think there is a problem with Kingston...right? However, if someone else says "I'm running plug-in X with Kingston and it works for me," then there's a very good chance the issue is not with Kingston, so maybe the person having the problem should look elsewhere.
Think of all the times in these forums when someone had a problem with a plug-in that was due to having an older version (or interface issues due to an outdated driver), and not realizing an update was available. Under those conditions "works for me" can lead to a solution just because the person with the problem knows that a solution is possible.
If
everyone experiences a problem (e.g., tempo changes not reacting properly when upsampling was first introduced), it's a no-brainer the problem is with the software because no one can say "it works for me" when it doesn't. However if some people have a problem and some people with (seemingly) identical circumstances don't, then it's important to find out why there's a difference. There's no incentive to look for a difference unless it's clear there is a difference.