A lot of people are talking about technological unemployment lately. This, if you’re not aware, is the idea that as technology gets better and better, a huge amount of jobs will become automated, leading to unprecedented levels of unemployment.
I’m not an expert on the subject by any means, but I’ve been paying attention to some of the discussions. Some of the more common reactions I’ve seen to the problem are things like:
- This is going to usher in a new golden age, where nobody needs to work and everyone can do whatever they want all day.
- Who cares, don’t be a luddite. New jobs will open up, we just don’t know what they are, because they don’t currently exist.
- Oh yeah, this is bad, we need a solution.
There are, of course, other reactions, but those are just the ones I see the most. I wanted to talk about the first two really quickly.
My first thought – hey, whoa whoa wait a minute, can I remind everybody that I don’t write proper essays on this blog, and it’s basically just a scrapbook where I throw things up? Yeah, I can remind everybody, and I am. Remember that.
Scenario 1 – Golden Age
So my first thought is in regards to massive job automation giving us all a wonderful heaven-on-earth type of situation. I actually don’t think this is an unreasonable prospect, although holy moly does it feel extremely unlikely and weird, just on a non-logical, visceral level.
I find that so, so often, I end up thinking about human evolution and how it relates to different issues (I understand that most of my readers are strictly into Raëlism, but please read the evolution part of this post anyway). In the case of technological unemployment, there’s one thing I keep focusing on is that the human brain has evolved over 6 million years in environments where it was really, really important to work, and do productive things.
So many motivations/behaviours that drive humans boil down to this, and I really wonder what would happen to most people if they suddenly had 8760 hours a year to kill. The human brain is not made to handle that, and I think the results would be incredibly unpredictable, and probably not great.
There’s a group of people we can look at to see how they might handle this: Rich celebrities whose careers have slowed down. They are in this same basic position: They have unlimited time on their hands, no urgency to make money, and they can just do what makes them happy.
I think most people could probably agree that, as a group, these people do not handle life very well. The good news is that there are a whole lot of them to test this out on (this is not great news for them). If you sit and think of a bunch of them, and Google their name along with the words “assault”, “mugshot” or “addiction”, you will probably get relevant hits on more than 50%.
It’s not a perfect test, but I don’t think it’s that controversial to say that a lot of people are not prepared for having too much money and too much free time. I think it’s because the human brain just hasn’t evolved that way, and it breaks badly when this happens.
So I’m not saying that what seems like an amazing utopia will lead to 7 billion individual versions of Foxcatcher, but actually I am. That’s exactly what will happen, wake up sheeple.
Scenario 2 – Everything Will Be Fine Somehow
I don’t like the logic behind just saying “Well, some people in the past thought we’d run out of jobs, but it turns out we invented a bunch of new ones, therefore it can’t possibly happen any other way”, but a lot of people do say this, and seem to think it makes sense.
The circumstances surrounding technological unemployment are much different than other huge labour disruptions. For one thing, this time, the insurgent labour force (computers, which are essentially salads made of microchips) will be able to do basically everything that humans are doing right now, not just a few tasks, as in previous eras. This means that even if we do come up with new job ideas, some smartypants computer will probably just think up a way to do them better than humans.
So one thing I’ve been thinking of is actually that maybe the ultimate job that computer programs won’t be able to replace will be: Customer support.
“Andrew”, you say, “This is actually the main thing that every gosh darn bot company is working on! This is maybe the perfect use for AI and bots and all that stuff! What are you, a fat idiot who doesn’t know what he’s talking about?”
Yes and no. I’ve seen a couple of demos of very cool computerized tech support chatbots myself, and they’re really impressive, but I’ve been thinking that there’s a limit to how much they should be used. Let me explain.
I have a million thoughts on customer support and tech support. My first job in the tech world was doing tech support, and then when I started my first web app, I did most of the tech support (with some help for a couple of the busiest years).
I don’t think you can overstate how important customer support is to a business. I find it incredibly satisfying to work on. I have to admit I haven’t always been perfect at it on my own sites, because of time/resources issues, but I’ve often thought that I’d really love to actually head up some sort of customer support research institute, or set up customer support systems somewhere, come up with guidelines, all that stuff.
I think support is becoming more and more important. It’s becoming easier and easier to create a good software product (obviously I’m not saying it’s easy, but there are more guidelines and known best practices, etc.) so providing really good customer support is a strong way for a business to differentiate itself from competition.
One thing I’ve found by doing so, so much support myself, is that customers care so much about dealing with a real person. People are not just there to ask a question, have their problem fixed, and move on. They love to chat about things that aren’t really pertinent to their technical problems, and they’re just generally very friendly, and grateful.
So this post is getting long, I’ll tell you what, I’m just going to wrap it up here with a quick summary, and then tell two anecdotes.
The summary is that I think one of the last jobs that might remain for humans could be customer support. Being able to talk to a real human to get your problems fixed could be a very powerful thing. Oh yeah and also the first half of the post was about how in the best case scenario, everyone is going to go crazy anyway.
These Two Crazy Guys from Thailand
As I mentioned, my first web job was doing tech support for a web hosting company. This was a very long time ago, and I’m not under any NDA, but I’m going to change the company name and my boss’s name anyway, just to maintain a little privacy.
Before I was hired, my boss did all the tech support. Once I started, I quickly saw that there were these two guys from Thailand who ran some quite nasty-sounding adult sites who were big customers. They paid a lot more for their hosting than the typical customers, and they weren’t super technical, so they had us setting up their billing pages and whatnot for them.
So these guys were friends, and if I recall correctly, they ran a couple of sites together, and then had their own individual sites. Their sites would have different niches, and they each had a number of them, and kept adding new ones.
I swear, every single morning, me and my boss (I’ll call him John) would log on, and at least one of these guys would write saying something like “John, you have to help, I was trying to change something on my billing page and it’s totally broken!!! You have to fix it!!! I’m losing money!!!” This happened so often, it was insane. They never stopped messing with their billing pages and cgi-bins and whatnot.
They also both (let’s call them Tom and Tim) wrote to us several times saying “Listen, John, you can’t tell Tim this, but his new site is doing really well, so I stole the whole thing and I set up this secret new domain. Can you set up the billing scripts, and DO NOT TELL TIM.” They both did this.
I wonder a decent amount about whatever happened to those guys and their businesses, but even if I remembered the exact set of words in their domain names (often following the pattern: adjective-nationality-gender.com), I’m not sure I’d go and check.
A Weird Tech Support Thing I Noticed
This isn’t as fun, but probably most people who actually read this site remember that I ran Diaryland, a blog-hosting community. About a year into the site, when blogs really started exploding and were the hot thing online, my site was getting tons of signups every day, and the premium, paid plans were getting a very constant, predictable amount of new customers every day too.
For a little while, the site was running on a couple of dedicated servers that couldn’t handle the big bursts of traffic that would occasionally come in, and there would be a lot of nights when things were bogged down and loading slow for an hour or two. Sometimes, this would lead to one of the servers completely crashing, and I had to perform a number of upgrades to more powerful servers, which always involved downtime.
The weirdest thing was, whenever the site had a downtime of a few hours, I’d be freaking out and working as fast and hard as I could to get it back, and I always assumed that people would be super mad at me and yelling at me in the tech support tickets. And of course, yes, people were mad, but shockingly, once the server was back up, the majority of the tech support tickets would just be people thanking me for getting the site back online.
I’d also get a lot of emails when the site was down who would be yelling at me, and I’d always reply and say sorry, and that I was working on it, and I’d say maybe 90% of people instantly responded apologetically, and suddenly were very friendly and supportive. The key here seemed to be that what people were really mad about wasn’t that the site was down, as much as their uncertainty as to how long it would be down, and whether anyone was really worrying about it or working on it. Just saying “Aaaah, I know, I’m sorry, believe me I’m trying to get it fixed ASAP!!!” assuaged all their concerns, it seemed.
Another odd thing: Right after outages, the number of people upgrading to a paid account would always spike for a few hours, and I’d often make double or triple the amount of normal revenue for that day. If the site was down for 8 or 9 hours, as it was a couple of times for major hardware repairs/migrations/whatever (this was a different time, running stuff was way harder, don’t judge me!), this pop in sales would actually be almost a little more pronounced. I guess there are obvious theories as to what was up with this, but it was always extremely odd to me that it kept happening. (And it goes without saying but this would be an exceptionally stupid thing to purposely try and replicate or exploit somehow.)