In other words, instead of worrying about the kind of hosting you have right now, shouldn't the focus be on figuring out ways to overload it? Isn't it kind of ironic or contrary to this belief, that people want the best host with the most web space and bandwidth? Shouldn't they want to slow as many machines to a crawl as possible?
I've heard stories of how businesses (startups or otherwise) love to switch to $5/mo web hosting with 5000GB of bandwidth which of course they will never reach before being kicked out. Or startups wondering what the best hosting company is. But, isn't that self-defeating to put those thoughts into your mind? Or that you're feeling inferior if you can't outgrow $5/mo web hosting?
I say, pay $10 a month for something sane, like 100gb transfer. Something that is $100 realistic. Then, when you outgrow the server, you know you're doing something right. And you don't have to feel guilty or inferior or that you're not using the space or bandwidth, which in reality you don't even have access to (they would shut you off.) So why do businesses try to disillusion themselves?
Wouldn't it be much better to find any web host with a bandwidth cap (so they simply shut you off and you're not stuck with a huge overages bill) and focus instead on building traffic of visitors to your site every day?
But all in all it's a nice problem to have. I tend to agree with Delicious guys...worry about scaling when it becomes a problem. That shouldn't be an excuse to code bloated SQL queries.
On a related note, I find that user feedback (of any kind) is a terrific boost: since most people are lurkers (or worse, look for 5 sec and never return), people who take the time to email us or post a blog comment provide a huge lift.