The benefits of the hot questions list are supposed to be for the sites that have these questions. Many of our users are quite capable of answering questions on Stack Overflow, Programmers, Seasoned Advice, Parenting and perhaps others. Getting these questions in front of a much larger and capable audience has proven to be beneficial across the network.
The key question here is whether the people arriving to the hot MO questions are:
- Capable of answering them, and if not
- Competent enough to vote on the usefulness and perceived quality of the question.
Think of something like Twitter, but the majority of users that read it know how a Stack Exchange site works, and the level of quality that we like to produce. It's a good thing to have, for the most part.
Once in a while, something gets on that list that is just rather sensational, and is up-voted quite out of proportion. A good example of this on Stack Overflow would be:
Cycles in family tree softwareCycles in family tree software
However, for every post that gets more votes than are probably warranted, 20 or more get a little more deserving attention and your community remains more easy to discover to newer users on the network that would make a great fit, but don't yet know that you exist. I'll concede that there are a limited number of research level mathematicians that might find the site in this manner, but they do exist. And, while not all are skilled enough to answer, a larger number are competent enough to know a darn good question when they see one.
I strongly recommend staying in the mix, despite the occasional (and known) downside. Math Overflow is growing, there is no doubt about that, and you're eventually going to reach a scale where you simply can't enjoy the benefits of being a relatively insular community any longer - it just won't be possible, and that's precisely what happened on Stack Overflow way before it reached its current growth.