Today, I was joking to one of my co-workers that at first, we were going to replace him with a computer, but then we decided instead to replace a computer with him.

Alas, I think "mechanical turk" schemes don't have much life left in them, while the age-old fear of replacing more people with computers (more or less) is realistic.

The short version of this argument is that Software as a Service, as a model for pushing certain activities outside of a company, is becoming more and more compelling. Sure, it's just another step in the whole push to use computers to automate and streamline business activities, but it's a pretty big step.

Moreover, more and more of these services are being sent out to a very few class-leading players in these services. Lots of organizations (including my employer) have dumped their web search needs on Google, while Yahoo! picks up scraps, and Altavista (I'm probably dating myself by even knowing of their existence) gets left by the wayside. In-house search engines? Let's put it this way: I routinely use Google to search websites with built in search functions (IMDB, Drupal, and even Wikipedia) because Google has faster, more accurate search results (useful hint: I typically do this by typing the name of the website, followed by the query (eg imdb joseph cotten).

That's a few examples among many. Locally, Darren Barefoot outsources his email provision to Google, thus solving his spam problem. Everybody in the world uses YouTube as their video host of choice, and it just works.

The common thread here isn't even the outsourcing. That's fairly old hat. The interesting thing is how quickly momentum seems to gravitate towards single (owing to a lack of greater precision, I'll call them "best") solutions. I think that before ubiquitous webification (and heaven knows, remembering even that is starting to date my existence; note that first-year college students this September were typically born in the same year I had my first Internet email account; they never knew a world without the Web, more or less), these best solutions were not as easy to distribute, as easily compared, nor was their superiority as easily communicated. In short, even when stuff got outsourced, it often got outsourced to stinky inferior solutions. This still happens a lot.

At any rate, something (and I may well have the essential mechanism wrong) means that it is much easier for class-leading solutions to a problem (or implementations of a new concept) to come in and just win all the marbles. And it's also easier to install this stuff. It's no longer a case of buying the server that runs the system. Indeed, it's no longer a case of renting the maintained server share and putting the application on it. Now, it's often a case of renting the service, and nobody knows what the server is. Or even what software revision you're on. Or where the IT department is.

I think in the long term, this means that a lot more job functions will simply disappear from businesses. You'll just see bigger and bigger companies decide not to bother with in-house email systems, because hey, email (and especially spam) is hard. Some of this is just greater computer literacy, to take only my narrow field: just as computers made functional typing ubiquitous, and ended the need for typists who could take dictation, SaaS and general network literacy is making the remaining necessary IT knowledge ubiquitous, and we don't need IT departments that can manage servers.

I suppose this is where I should admit I fear that the entire field I work in will disappear before I get a chance to retire. That may well be true. But I am sanguine. The heroes of IT will always have jobs, because those server farms, however concentrated, aggregated, and outsourced, haven't gone away. But just like the erstwhile auto workers, the rest of us will somehow find jobs, because it turns out when the burden of managing IT infrastructure becomes a lot lighter, the companies that shed their IT burdens will now be more productive.

The medium-term effect is that we'll all get jobs working hard on whatever thing our organization actually is best-of-class at. For locally-oriented organizations, this may largely reflect a lack of global competition (at least until a smart global competitor figures out how to encroach on that service...), and for global organizations, well, take heart from the lesson of IBM: death sometimes comes slowly. For everyone else: get really good at something, anything, and make a go of it.

And at the risk of showing myself a fool in short order, I see this as a slight thematic re-emphasis here at Wired Cola. As part of our mission to be more Cybermorphic every day, I'm hoping to figure out how technological trends are going to affect ordinary people in the future (and I'll take that as meaning anything from "it's already happened, you just didn't notice" to the end of living memory), and explain them for a lay audience. I often joke that my day job is "nerd-human relations," so the theory is that might actually be what I'm good at. Let me know how I'm doing.