A couple years ago I wrote a piece on dementia using artificial intelligence as a jumping point. I alluded to AI taking jobs, what’s commonly referred to as job automation. One way I framed this was artificial intelligence, or machines, can do a lot of things. Things we often consider very hard. You might have a calculator which can do more math than you ever could, but you don’t have a computer which can walk up a flight of stairs. Something we consider very easy.
Then this video came out:
It might not be as smooth or widespread as a human (yet), but AI is climbing stairs pretty damn well now.
Annnnnd then this shit came out:
-> Though not as impressive when thought of in this light.
Job automation talk has been going on for hundreds of years. As we tend to think we’re unique, every time each generation has said “this time is different,” which is what’s going on now.
“The pace of improvement is unprecedented, the job losses will be unprecedented. Silicon Valley and software is truly starting to eat the world. Other than those who own / make the robots, we’re screwed!”
It all makes sense. Computers can do more and more, replacing more and more of what humans can do. Computers typically do these things better, so they supplant us.
Fortunately, what humans consider logic isn’t what statistics considers logic. Saying something doesn’t make it true. We need to test these things. Our logic is more often than not merely assumption. Instead, we can look at research papers like:
There’s beauty in some of the simplicity of this paper- Bank tellers and ATMs. ATMs kill bank teller jobs, right?
“The ATM is sometimes taken as a paradigmatic case of technology substituting for workers; the ATM took over cash handling tasks. Yet the number of fulltime equivalent bank tellers has grown since ATMs were widely deployed during the late 1990s and early 2000s (see Figure 2). Indeed, since 2000, the number of fulltime equivalent bank tellers has increased 2.0% per annum, substantially faster than the entire labor force.”
Not so logical after all.
“Why didn’t employment fall? Because the ATM allowed banks to operate branch offices at lower cost; this prompted them to open many more branches, offsetting the erstwhile loss in teller jobs (Bessen 2015). At the same time, teller skills changed. Non-routine marketing and interpersonal skills became more valuable, while routine cash handling became less important. That is, although bank tellers performed relatively fewer routine tasks, their employment increased.”
Or maybe it is logical!
The tasks of bank tellers, some of them, have gone extinct. The occupation, what we care most about, has only done the opposite. Again, this is not complex. We hear over and over software-rules-people-drool, but once you do something basic, like look at how many bank tellers there are, you then look like a moron when you say “ATMs kill bank teller jobs.” You need to actually look at some statistics!
Paralegals are another commonly referred to “brink of extinction” job. The study references growth for that profession too.
Overall, the paper shows jobs with the most growth in computer use are jobs which have grown the fastest. Furthermore, non-routine jobs, the jobs we consider to be safer from automation, have had their computer use grow faster than routine jobs since 1997. The non-routine jobs are where the automation has moved towards.
There is also a nice reference to ecommerce, which is another common automation-job-killer arena. Yet ecommerce makes up only 7% of retail sales! Even if Amazon owned all online shopping, that would mean 93% of the market it has no piece of. Amazon has been around two decades. If only 7% of the market has been penetrated, is it really likely that all shopping is on the brink of being non-brick and mortar?? (And that all brick and mortar shopping is on the brink of using significantly less employees?) Borders died, but Barnes and Noble is still kicking. And Amazon has opened two book stores! (Seattle and where I live, San Diego. I recommend visiting if able. They’ve done a very nice job.)
This is not luddite fallacy thinking. This is what’s actually happening.
The fitness industry has been privy to an onslaught of technology since July 9th, 2008. The date the app store opened.
The fitness industry includes more than just personal trainers, but looking at trainers seems best for what’s going on. Gym memberships and the like can grow regardless of the infiltration of apps, because people may go to the gym merely to use their apps.
If we do a google search up until July 8th, 2008, for “number of personal trainers bureau of labor”, we get not a whole lot. But a Men’s Health article from 2005 cites the Bureau of Labor as saying there were 182,000 trainers at the time. Seems close to enough for 2008 statistics. (For 2000, this article says there were only 62,000. Showing the growth within that timeframe.)
So we’ll call it maybe ~200,000 trainers in 2008, right before the app store opens. Before Xbox Kinect took off, Nintendo Wii, Apple Watch, FitBit, lord knows how many more home workout DVDs, the list is very long.
In 2015, the Bureau of Labor says there are 279,100 trainers.
1) It’s worth considering if the market for trainers is currently outpacing app adoption, albeit at a decreasing rate, where eventually the lines will cross, and trainer jobs will decrease. This is unlikely considering the insane rate of adoption of things like smart phones the last decade. We’d be saying people are adopting personal training services faster than they are adopting fitness tech. We know that’s not happening because in the last 10 years there are less than 100,000 new trainers, but billions of smart phones, millions new fitness devices, thousands of fitness apps. Personal training jobs are growing in spite of smart phone / band / watch / app adoption.
What we could possibly say here is if the smart device revolution didn’t happen, perhaps personal trainer jobs would have grown even more.
2) While I don’t know, it wouldn’t surprise me if using new technologies lead to increasing personal trainer adoption. It’s rather common a new client starts off talking to me with “I was trying [insert technology (home treadmill, Bowflex, P90X DVD, running / nutrition app)] but then I fell off. I need more accountability than that.”
Thinking back to our ATMs and bank tellers, it doesn’t have to be an either or thing. I have plenty of clients who use me and an app. There is room for augmentation rather than solely replacement.
3) Whenever you try to make predictions about the future of work, the future of human behavior, the future period, you open yourself to looking very silly. Unless you can account for every variable you can’t be 100% sure, and it’s impossible to account for every variable, particularly considering how many variables won’t exist until years later, and the fact human behavior changes with time.
3) Michael Lewis is one of the most popular authors in America right now. He’s written Moneyball, The Blind Side, Liar’s Poker, Flash Boys, The Big Short. He’s very well received, and a very sharp mind. (His books are highly recommended. The Big Short is a great movie too.)
He also wrote a book about Silicon Valley back during the ’99 boom, and he lives in Berkeley (right by the Valley). In a recent interview he stated the place is “basically a bunch of autistic people wandering around.” I’m sure way too many people overreacted way too much to a joke. A joke, like much comedy, which is funny because it has truth to it.
When you spend your days behind an object, no matter how much it can interact with you, which is not human, an object you’re obsessed with, likely because you can supplant human interaction with it, in the least it’s plausible to state there are potential trickles of autism there. (It’s a spectrum after all. Not an on / off thing.)
-> Jobs, Gates, Musk, Bezos, Zuckerberg, read the biographies on these tech entrepreneurs and one commonality is the horrors people have had communicating with them. Though some examples may be called sociopathic rather than autistic!
Any future predictions are ones to be careful listening to, but even more caution should given when listening to proclamations on the future of human interaction from those who minimally interact with everyday people / those who might not have a great grasp on how human interaction happens to begin with.
-> A tangential example:
No wonder advertisers have thought the future was online! Highly recommend The Ad Contrarian for more examples of this. The difference in the thoughts on what digital ads would do and what they actually are doing is staggering.
In case a disclaimer is needed: I love the Bay Area. I visit often; I’ve met great people there. I own numerous tech gadgets. None of which is in opposition to the above.