I’ve always felt weird about my interview with Jeff Bezos. Amazon’s PR people set it up as a follow-on to a keynote he gave, in 2006, about the suite of web services then coming online: cloud-based storage, computing, and human labor (i.e. Mechanical Turk). This was for InfoWorld, back when it was a magazine printed on paper. The PR folks were determined that we would put Bezos on the cover. We’d already written about all that stuff, so that wasn’t going to happen unless the keynote yielded something new and important.
Predictably it didn’t. But as they led me to the backstage room for the interview they kept pushing for the cover. And something must have snapped. Because when I pointed my camcorder at Amazon’s CEO and began talking, I became somebody I’d never been before and have never been since: the pushy reporter who is determined to frame a narrative, who steers the interview that way, and who interrupts all efforts on the subject’s part to answer questions not asked.
I’m not linking to the video because I still feel conflicted about it, but it’s on YouTube and I watched part of it recently in the wake of the New York Times story about Amazon’s workplace culture. Mechanical Turk was top of mind for me when I did that interview. I had been a customer of a Turk-based service called CastingWords, to which I had outsourced the transcription of several of my then-weekly podcasts. To prepare for the interview I signed up as a Turker and walked a mile in the shoes of a podcast transcriber. Here’s how I reported the experience at the time:
Podcasts are chopped up into six-minute segments. Last night, the segment I transcribed was worth $1.02, which will be credited to my Amazon.com account (and can then be transferred to a bank account) if my work passes several quality review checks. These checks are themselves implemented as HITs [human intelligence tasks].
…
My encounter with Turk Work wasn’t very satisfying. There was no context, no orderly progression, no sense of collaboration, no awareness of (or pride in) a finished product. If you’re running an MTurk-enabled business, you have to focus on throughput and efficiency. That drives you toward assembly-line tactics. But as Nathan McFarland noted in our conversation, you’d also like to reward excellence:
“We’d be willing to pay certain workers more and other workers less, but that’s not an option right now.”
So for now, that’s a hard constraint. Here’s another. From my perspective as a CastingWords customer, there are clearly some transcribers who have a better feel for my material than others. I don’t know who they are, though, or who’s transcribed which parts of various podcasts. If the goal is maximum throughput, that may be necessary. But that’s not my goal. Transcribing 620 minutes of audio for $260 in six days was darned impressive, but two months later I’m still only halfway through the final polishing, which I’m tackling in fits and starts. If this winds up being a long-term relationship, I’d rather identify the transcribers who do well on my stuff and give them whole tasks. And some transcribers would likely prefer that too. I sure would. Apart from the question of whether it’s in CastingWords’ interest to allow such relationships to form, the current MTurk architecture precludes the possibility.
I pressed Bezos hard on that point. Too hard, some thought at the time. Perhaps they were right. Perhaps it was none of my business to suggest that Mechanical Turk ought to optimize for worker satisfaction as well as for throughput.
My first encounter with a digital sweatshop was in 1993, in Bombay (as it was called then). I was visiting a software industrial park. In one company’s office I saw a room full of desks at which workers were digitizing maps of US telecommunication infrastructure. It seemed like dehumanizing work. Then I looked out the window at laborers carrying piles of bricks on their heads in 100-degree-Fahrenheit heat and realized that those map digitizers, in their air-conditioned office, had some of the best jobs in town.
I have no inside knowledge of Amazon’s blue-collar or white-collar workplace cultures, so I’m in no position to judge. Given that MTurk was a forerunner of the gig economy, though, here’s the follow-up question I wish I’d asked Jeff Bezos then and would still like to ask now. Do you believe that optimizing for customer satisfaction and worker satisfaction is a zero-sum game?
Well…I’ve done that sort of work, and left it almost immediately, because the setup is hostile. You do work, and it’s presumed your work is bad, so you have to run through checks to make sure it’s all right, and usually “all right” isn’t about anything very smart; usually it’s depressingly dull. And then you’re paid like the person who’s paying you thinks it’s smart to screw people out of money, which is also dumb, and depressing. Inherent in the whole thing is a deep mistrust and dislike of people, of employees.
The nicest work I’ve done, apart from teaching — which isn’t always nice, especially if I’ve done it badly — has been bookstore work, because there, assuming you’re in one of those nearly-extinct things, a decent bookstore, you’re there doing considerably more than handing over books for money; you’re involved in making part of the world, making culture. Because you’re talking with people. They come in and hang around, and talk, or come in as strangers to bookstores and you engage them in conversation, and the next thing you know books start coming off the shelves, and the conversation is continuous with the thought and images in those books, the stories and ideas. It’s just ordinary, but it’s an important kind of ordinary. And everyone involved has value and is a whole human being — you, the customer, the writers, the other people working in the store, the other people shopping, the world that can produce a book and a bookstore.
I seldom buy from amazon because of how they treat their workers, writers, publishers. People. When I do, generally I order from used bookstores that do their own shipping, rather than run people around in amazon’s warehouses.