The Foretellix CTO Blog

Summary: This post takes a (speculative) look at the general implications of automation over the next 20 years: Employment declines. Democracy does not fold. Universal Basic Income becomes common. Education improves. New-age stuff grows. Quality-of-life inequality grows or shrinks, depending on how you look. And so on. A follow-up post will discuss the verification implications.

Note: Title changed (from the more-bland “The next 20 years of automation”).

My previous post, The rise of the mostly-autonomous systems, talked about a specific slice of the (speculated) future: The implications of systems which can almost, but not quite, replace humans in many jobs.

This new post is about the bigger picture: What will the next two decades look like, as a result of automation?

Now, this is all shamelessly speculative and simplistic. It runs the risk of sounding like a mix of the improbable and the obvious. And while I may (perhaps) have some shred of credibility when discussing the-future-of-verification, I clearly have none regarding the-future-of-society. So take it for what it is – just my thoughts, based on a lot of superficial reading. For instance, I am no economist, and a lot of this has to do with economics, but I read enough economy-related blogs to know that at least some economists agree with some of what I say.

I mainly wanted to do this lighting-tour-of-the-future so I can then go into what I really want to think about, which is the simulation / modeling / verification implications of all that. However, I realize that for many readers this might feel somewhat tacked-on (“Now that we are done discussing the zombie apocalypse, let’s consider its implications on the soft-drinks industry”). So I’ll put that part in a separate, follow-up post.

With that out of the way, let’s dive right in.

Assumptions

My main assumptions for the next 20 years (further details below) are:

• Automation will make more and more people unemployable
• No dystopia: Democracies will continue to function “reasonably”
• Some form of Universal Basic Income (UBI) will become common

Here are some quick clarifications:

• When is that happening? The processes I am talking about (more automation, less employment etc.) are already happening, and will just keep coming. There is no specific arrival date. For concreteness, I am discussing (somewhat arbitrarily) the next two decades.
• I am talking about liberal democracies here. With luck, this could be most of us on this planet in 20 years (e.g. if China becomes democratic and less corrupt, and India stays democratic and becomes less corrupt), but maybe not.
• How likely are these assumptions to hold? My intuition is to give each a higher-than-50% probability. For the whole set? Dunno. It is at least one reasonable future to consider.

Let me try to justify those assumptions:

Automation stealing more and more jobs: If unconvinced of that, perhaps start by watching this 15-min video: Humans need not apply. Then read this piece about the influence of driverless trucks, and finally perhaps read this long, thoughtful piece by investor Bill Gross.

Yes, some still think that there will be new jobs to replace the disappearing ones, and clearly there will be new job categories (such as those operators-of-mostly-autonomous-systems I mentioned here), but the overall trend is less and less employment – see chart I in that Bill Gross piece.

And this is accelerating: That video is two years old, and since then advances in machine learning, robotics etc. seem to come at an ever-faster rate. And at some point (In 15 years? In 30 years?) most people will be unemployable, even with good and adaptable education (which I think is coming – see below).

No dystopia: Yes, I know, doom and gloom scenarios make better movies / news stories (vs. “This just in: everything still basically OK”). But I think democracies are relatively stable, and have a history of finding reasonable solutions to changing conditions. And these unemployed people will still be voters.

You may feel the urge to call me a hopeless optimist. E.g. you may want to say that your country is already corrupt, or that the billionaire class controls everything and so there is no true democracy, or that in times of crisis the populists and the xenophobic are more likely to take over, and so on.  You have no idea (I hear you say) how bad our politicians can get.

There is a grain of truth in all of these, but I would call for a sense of proportion here. Yes, you guys (say in the US) may be more corrupt than Denmark, but you are still in pretty good shape. That corruption perceptions index should really be viewed as if it was drawn on a logarithmic scale: You may find it hard to appreciate just how bad corruption can really get. And it takes years of dedication to get to where e.g. Cameroon is (not to mention the People’s Paradise of North Korea, proudly holding on to 167’th / last place).

And similarly for those other worries. Over time (and with some painful missteps) democracies will probably auto-correct.

I also assume no other global catastrophe in this timeframe: No global war, no asteroid hurling towards earth while Bruce Willis is otherwise busy. Artificial General Intelligence risk starts creeping up in probability towards the end of this period (which means we should start thinking about it yesterday), but it does not happen yet.

Universal Basic Income: That idea, which says everybody gets some “basic” income (say $10,000/year) regardless of how much they earn, is becoming more and more popular. It is still controversial (“can we pay for it?”, “won’t everybody stop working?”), but less so over time.

It has a lot going for it: If we assume that a huge part (most?) of society will not be employable in the classical sense, then we should do away with the indignity of food stamps and standing in line in the employment office. And we don’t need the whole bureaucracy whose job is to check that you are eligible. And you are encouraged to try some side jobs – this will not make you any less eligible. And you have food and shelter security – that’s a huge thing.

How would we pay for UBI? Taxes will probably go up. Some reasonable solution will be found for taxing multi-nationals (after a lot of thinking and screaming – this is really hard to do right). And also (if you believe that Bill Gross piece) central bankers will simply print more money.

UBI comes with a basic change of philosophy. We used to tell ourselves / our children that work is a really important value. When, say, 50% of the population are permanently unemployable in the traditional sense, it is probably wrong to keep saying “OK, you are temporarily out of the workforce, but we’ll do our best to get you back there ASAP”. There is nothing temporary about it, so we’d better get ourselves some new views.

Some implications

I think many readers will agree that there is some reasonable probability that all three assumptions will hold for the next two decades: Less than 90% probability, but more than, say, 10%.

For the rest of this post, we’ll assume they do. What are some of the implications?

Note that this post is already too wide-ranging, so I’ll blissfully ignore some un-ignorable topics, like the effect on immigration, and the possible rise of nationalism, and how UBI relates to medical insurance, and the possibility of designer babies, and all that good stuff. Yes, it is all connected, but we can’t discuss everything (e.g. see this funny video).

OK, here we go:

Minimum wage will probably disappear: Unions will fight a rear-guard action (“It is especially at times like this that we need it”), but in a sense, once you have UBI, much of the justification for minimum wage goes away. And even if “officially” it stays, people will simply become independent and do “consulting jobs” which are exempt from unions

Capitalism will be alive and well: While UBI sounds a bit like socialism, and taxes will probably be higher, capitalism as we know it will continue.

All these people on UBI+ (i.e. people who live mainly on UBI, but may also have some irregular other jobs) will be a huge market of sophisticated customers. And then there are all the richer people.

Automation will accelerate: Once UBI exists, why would somebody do unpleasant work? Well, if it pays very well, or it is challenging. So there will be a bigger push to automate things which are currently done by people on very low wage.

Also, once many / most people are on UBI, I think luddites / unions / special interest groups will find it harder to block further automation: People on UBI will look unkindly on a group suggesting that a service / product should stay expensive just so that group will keep their jobs.

Will inequality grow? Depends on how you look at it: Income inequality will probably continue rising (even though taxes will be higher to finance UBI). But how about Quality-of-Life inequality? That’s a lot harder to measure.

If you are living on UBI+, say 20 years from now, then how does your QoL compare to that of the seriously rich? This really depends on how you compare. You could take the view that they have much higher QoL: They can have infinite foreign travel. They can go see that Sting concert. They can live in style in the most expensive cities. They can afford all those exotic medical procedures.

But if you compare yourself to the average bloke just 20 years ago (i.e. in 2016) then you could argue that the rich have just a somewhat higher QoL than you. For one thing, you now have the security of always getting reasonable food, shelter and other stuff, without having to do unpleasant work, and without being constantly under stress. That brings your QoL up quite a bit.

Also, remember that almost everything at some fixed quality keeps getting cheaper: Even now, the guy who drives a car 10 times as expensive as yours does not really get 10x quality out of it, right? Depending on how you count, maybe he gets 50% more quality out of it (if that), because pretty-good cars became relatively cheap.

So 20 years from now, you (on UBI+) will have pretty good food and shelter, a Virtual Reality set + infinite entertainment, excellent health facilities, excellent education options, democracy, safety and so on.

You can’t have that exotic medical procedure that those rich guys can, but your life expectancy and health are pretty good relative to 20 years ago (not to mention 200). And overall, life expectancy is increasing and health inequality is down.

Going to that Sting concert is $500 (travel and all), so you can’t afford it without lots of sacrifices. But the live-VR version is just $10, and you feel like you are really there (front row). And if you don’t need to see it live, it is free (or else you can always download it from some torrent). Virtual is cheap.

You can’t buy a car, but you have an electric bike and can summon an autonomous car at any time for a ride. And so on.

So how bad is inequality? Depends on your point of view. If you convince yourself to be happy and not constantly compare with the rich, you can be pretty happy. This, of course, is pretty hard to do: Our brains are wired to enjoy things only briefly, and then want something else. Evolution does not want us to be happy – happiness is just part of the control mechanism it invented, all geared towards having as many decedents as possible.

Which brings us to:

So what will everybody do?

 There will be cheap, good education: Education will be much better, and everybody who wants it will have access to it. For one thing, MOOCs are already proliferating. Also, eventually, somebody will come up with a good remote-learning / video-game hybrid (which helps you learn stuff but is as engrossing as a video game).

Also, as the implications of automation become clearer, politicians will suggest (and implement) better education for all. Admitting that employment is bound to decline is hard for politicians, so they won’t. Instead they’ll say “We need to improve education to prevent unemployment”. Regardless, education will improve, and this will delay (somewhat) the inevitable job decline, and create a legacy of many-paths-to-education.

There will be a strong push to learn programming if you can. While not all programmers will be immune from unemployment, some will. Also, programming (and related skills) will be useful for many other things, much like “using email / a word processor” is now considered a basic skill.

But more importantly, since the boundary between what-machines-can-do and what-humans-still-do will keep shifting, professional people will constantly need to instruct machines. So if you are psychologist or an artist or a game designer or an educator (or whatever the combination of all four will be called) you will often need to teach the machine something new. Yes, machine-learning-by-example will make some of that easy, but in many cases it would be much simpler if you could just program the damn thing (either visually or in text). At the very least you should be able to explain your wishes to a programmer in terms she will understand.

In a sense, people will have more fun: Once you remove the stigma from playing-VR-games-all-day, whole new cultural things may bloom in that intersection of video games / art / television / movies.

And of course, people will find other, non-money scales to excel in and impress others: Video game funny-money, various measures of popularity, etc..

People who do work will find it more satisfying. As I said above, unpleasant work will either go away or will be well-compensated.

It will also become easier to do all kinds of part-time jobs, Mechanical-Turk style. We’ll all be potential-part-time-entrepreneurs.

Everybody will be an artist: Everybody will be pushed to discover their creative side. And they’ll have a lot of time on their hands, and excellent, new (e.g. VR-based) facilities to express their creativity, alone or with like-minded people.

Selling art may be even harder than today (because everybody will be doing it, and because free art-by-machines will also get better). But art is one excellent way to look for meaning, so it will become popular.

Will people have more children as a form of self-expression? Dunno. Some of that probably depends on the finer details of UBI.

This may be an excellent time for religions / alternative / new-age stuff: Assuming meaning-through-work is blocked for most people, an enterprising religion/cult/whatever can be really successful. Obviously, I am referring here to US-style mostly-free-competition religion-as-a-service, not Pakistan-style religion-as-a-life-sentence.

Right now, the video-games demographic is not particularly religious, but as time goes by there will be a huge chunk of society looking for new meaning, with VR gear and lots of time on their hands. You just need to find the right angle for success.

Of course, the heavy-duty, orthodox branch of most religions is anti-internet (“nothing to see here – go back to the holy books”), but US mega-churches (and their spread to Africa and other continents) are testimony that free competition can push religions to modernize via trial-and-error.

Other forms of grouping which give members a sense of brotherhood, relevance, honor and direction will obviously flourish: Our proud city / nation / sports club / video-game community.

The possibility of a two-track economy: You can’t live in the really expensive cities on UBI+. On the other hand, most people will live in cities (they already do). So people will go to cheaper cities, where quality of life is good and there is a lot of outdoors. This, of course, is already happening.

One of the big worries will be the creation of a two-track economy: the growing split between the people who live the slower life in the less expensive cities, and the people who live e.g. in Silicon Valley, where everybody is forever busy and up-to-date (and making more money).

This may happen: Consider the fact that people find it very hard to do startups with twice as many people working half-time. You need to be all there. For a description of that intense, never-stopping, slightly-crazy atmosphere of Silicon Valley startups, read the long-but-excellent Uncanny Valley.

But perhaps this split will not happen. With virtual communities, fast internet, VR (and a mental shift) perhaps it will be possible to work well with remote teams and keep up-to-date. And cities will make a hell of an effort to adapt and become “one of the hot places”.

Helping others: One of the best ways to find meaning is to help others. Luckily, this is not going away: There will always be some people less lucky than you. With global connectivity and lots of smart, educated people, this will probably bloom. Interest in effective altruism, as in Doing Good Better, will probably increase.

When I was talking, a few paragraphs ago, about religion-as-a-startup, you may have suspected me of being a cynical guy. But I am not, really, and I think this helping-others stuff could be huge.

Summary

 I have outlined here the implications of one possible future. It is a somewhat strange future. As my friend Sandeep joked: “I can’t decide if I should encourage my kids to seek a career as a robot, artist, or cult leader. :-)”.  But I think it is a fairly-probable near-term future.

I assume this will happen (with variations) almost everywhere, with the US perhaps being the most visible. I am not an American, but I think Leonard Cohen is on to something when he sings:

It’s coming to America first,
the cradle of the best and of the worst.
It’s here they got the range
and the machinery for change
and it’s here they got the spiritual thirst.

Whether we’ll view it all as bad or good depends on our frame-of-mind (i.e. on psychology, culture, and the kind of messages Hollywood will sell us all). In my own view, it is what it is, so we may as well make the most of it.

Comments are very welcome.

Notes

Expect a post about the simulation / modeling / verification implications of all that, coming Real Soon Now.

Sneak preview: All that automation (and its interaction with people and other automation) will be pretty hard to verify. Even specifying it all will be hard, and It’s the spec bugs that kill you. Luckily, there will be lots of new resources to help in verification.

[Edit 29-July-2016: Here is the first post explicitly about those implications].

I’d like to thank Sandeep Desai, Yael Feldman, Amiram Yehudai, Shai Fuss, Avishai Silvershatz and Kerstin Eder for reviewing previous drafts of this post.