Bubbles aren't bad things, at least when they get started. When they are little, itty, bitty starter bubbles, they represent enthusiasm and energy and innovation associated with a genuinely wonderful thing: canals, railroads, electricity, radio, jet engines, automobiles, calculators, personal computers, you name it. Even non-agency mortgage backed securities and derivatives based on them have some genuinely awesome wonderfulness associated with them (the pretend-the-risk-vanishes part, not so much, and the pack-it-full-of-vanished-risk really not at all).
By the time corporations (never mind the investing horde) gets hold of an idea, it's usually been scuffed up a bit. The basic idea of the Cloud is: you can use my compute cycles/memory/data storage from a distant device, if you compensate me. That's really not a new idea -- that's how all kinds of people figured out a way to justify the purchase of a mainframe or several for a corporation when no department could possibly afford it on its own. But there are a bunch of aspects to it that make it a "Cloud": It Just Works, You Don't Have to Think About IT (<-- little joke there, har de har har) and It's way cheaper than running my own IT.
Old forms of shared compute cycles/memory/data storage tended to present in a really user-unfriendly way; the Cloud doesn't make you work through a file system (or worse), which is also genuinely helpful -- and the Cloud (at least if your applications are implemented correctly) also offers reliability characteristics that are really fantastic. For people running businesses on the web, the Cloud is tantalizing, because it's not just expensive to keep everything up and running all the time. It's hard. For someone else to offer a turnkey solution? Sweet Deity, Praise the Whoever.
It's quite fine to get all excited about the Cloud and what companies can do with the Cloud (or a Cloud or several Clouds or whatever). We're going to be seeing a lot of that in the near future. We're also going to see a whole swathe of IT people looking for other work.
I would imagine it'll look a little like all those guys working at printeries for newspapers and/or magazines and/or direct mail advertising and/or books etc.: more jobs lost than created, even if LightningSource is hiring.
The flip side, however, is that whenever it gets easier/cheaper to use a tool, a whole crowd of people show up and start using that tool to do things that weren't possible/economically feasible before -- new and interesting things. And _that_ is where I predict we'll see the Cloud bubble. It's going to look a little like all the apps made for Android and/or iOS and all the self-published books/music/video produced for numerous platforms: a lot of crap, a fair amount of interesting/good stuff that's hard to find, and a few monster smash hits that we can't remember not knowing about after they've been around a couple years. It'll also look a little like name-compute/memory-intensive-algorithm-that-does-something-people-will-pay-a-little-for -- that is, every idea that AI ever had that wasn't actually feasible is probably going to get trotted out again and tried whenever time in a/the/some Cloud is cheap and paused during the Christmas season when cycles are fully utilized. Recommendations engines and social-wtf will likely top this list, but I wouldn't be too surprised to see matchmaking services go really and truly nuts (think airbnb, relayrides, etc., but for the moral equivalent of Pets.com). There will be big promises, like, We'll figure out who those people are in your collection of old photographs, track down their email address and connect you via Facebook! We'll find the perfect playdate for your n-month-old! Those promises will seem stupid cool when they're being advertised with sock puppets during the SuperBowl. And we all know already how that will turn out.
File this under: predictions to laugh about when I turn out to be totally wrong.