Maybe, but it seems a weird argument to make during record low unemployment. Yes, I'm aware of labor force participation rates, and that the pay for many of these jobs is too low, AI is predicted to destroy jobs, etc, but that's not what we're talking about, right?
Amazon employs way more people than B&N, the company they put out of business. Maybe those jobs should be better (OK, they definitely should), but that's different from saying there are no jobs, isn't it?
> Maybe, but it seems a weird argument to make during record low unemployment. Yes, I'm aware of labor force participation rates, and that the pay for many of these jobs is too low, AI is predicted to destroy jobs, etc, but that's not what we're talking about, right?
Admittedly, I'm driving the conversation towards a general problem from this specific example. But I think that even if I accept the premise that right now, things are OK, we still need to prepare for the future that's coming. Automation is coming for truck and taxi drivers, for restaurant workers, for the last remaining factory jobs. It's even coming for relatively skilled workers in some fields. It's not directly what this post is talking about, but it's the context in which this post takes place, and we shouldn't avoid it.
> Amazon employs way more people than B&N, the company they put out of business. Maybe those jobs should be better (OK, they definitely should), but that's different from saying there are no jobs, isn't it?
Amazon employs more people than B&N, but B&N is by no means the first or only company they have destroyed, and there are more to come. And as soon as Amazon can make the same amount of money while employing fewer people, they will do so. Every company does. The cycle repeats, only this time more people lose and fewer people gain.
we still need to prepare for the future that's coming
I disagree, honestly. Not because I don't think all the things you're saying will happen will happen, but because of all the things that are also going to happen that we're not talking about, because we don't know what they are. Yes, tons of jobs are going away (although I think it'll be 10-20 years for most of these, not 3-5 years like many seem to think), but what new sectors, opportunities, jobs, etc. will be created? I'm an optimist, so I genuinely think we'll adjust and adapt. I think we'll end up with most people doing jobs that don't exist today.
I could be wrong though.
However, no one knows the future, and the prediction that tech will destroy more jobs than it creates has been made since the start of the industrial revolution. So you basically have to fall back on the claim that "this time it's different". Which has also been made repeatedly during that time period :)
But honestly, at the end of the day, it doesn't really matter whether we should prepare or not, because we won't. Not in America at least.
There is no chance in hell that our political system will get its act together in the next two decades to get ahead of this problem and put policies in place for a bad situation that's coming but hasn't arrived yet. Look at all the problems that we know are problems because we're already living through them that we won't even fix now. We just put patches in place (maybe) and keep kicking the can down the road. Seriously, if half the jobs in America get destroyed in the next 20 years, it'll take us 40 years (or violent revolution) to address it.
> Yes, tons of jobs are going away (although I think it'll be 10-20 years for most of these, not 3-5 years like many seem to think), but what new sectors, opportunities, jobs, etc. will be created?
How many of the people whose jobs are destroyed will be able to move into the jobs that are created? In something like 30 states [0] 'truck driver' is the largest single occupation; where would those skills be re-applied?
I agree that nothing will be done regardless due to a lack of political will, but I don't share your optimism. I suspect that the change will be gradual to begin with, but that suddenly there will be a whole category of people unable to earn a living with their skills, and unable to afford to re-skill into something that can be used.
I think we are pretty far from truly replacing truck drivers. There will be a long period where the truck driver is in the car but not really driving, maybe even a very long time depending on if it is enshrined in law.
Even if the truck did drive itself, would it also load/unload and deal with flat tires and other vehicle damage conditions? I don't think these problems are insurmountable but I do think they will further extend the life of the truck driver profession.
I mean trains and subways don't even drive themselves right now somehow, so I think truck drivers will be fine.
I think the first step will be replacing long-haul truck drivers with autonomous trucks, they'll start at a depo near a freeway and end at a depo at the other end, then a human will drive to make the local delivery. Freeways are (relatively) easier to navigate than local streets.
That takes a 30 hour long haul job and turns it into a 30 minute job -- a single driver can do 10 of those a day.
Autonomous trucks will handle flats and other roadside breakdowns the same way many human drivers do -- pull over and wait for the repair service to come.
There's less incentive to replace train/subway drivers due to the high driver to passenger/cargo ratio (plus union rules get in the way) -- 1 truck driver to 40 tons of cargo versus 1 train engineer to 10,000 tons of cargo
>Autonomous trucks will handle flats and other roadside breakdowns the same way many human drivers do -- pull over and wait for the repair service to come.
Not to put too fine a point on it, but you clearly have no idea how the trucking industry works.
There are dozens of fully-automated train lines, some with attendants but many with absolutely zero personnel onboard, and the number is expected to quadruple within 10 years.
You may be right, I don't know. I'm not saying we shouldn't worry about this, just that I'm personally skeptical, and not really in favor of restructuring society for a problem that might be coming. If we had done that all those times in the past...
To take the truck driver thing, here's a plausible scenario (to me anyway).
First of all, "top job in X states" seems like a really weird way to express this. It honestly feels like a way to make it seem as dire as possible.
There are something like 3 to 3.5 million truck drivers in the US. That's a lot. However, BLS reports that only about half of these are heavy and tractor trailer truck drivers [1]. So maybe the rest are driving smaller trucks and vehicles around town? Those seem safer from job loss to me, I suspect that any job that requires driving a truck here and there and loading / unloading things will not be replaced by AI as fast.
Side note: BLS is predicting 6% growth in this job from 2016 to 2026, which is about average. But let's assume they're wrong.
So let's say that 1.8 million are long haul truckers that can be relatively easily replaced by AI. How long will that really take? I know everyone thinks that self-driving cars are around the corner, but I'm really not convinced. We've been hearing that for a few years now, and it seems like lots of teams are hitting some roadblocks (heh) in terms of their ability to solve some of the remaining issues. Just my ambient observations though, I'm no expert.
Even if we do get there with some cars in the next 3-5 years, are we really going to be ready to put heavy loads on highways unattended? We're going to handle the technical issues, security issues, legal issues, insurance, in a few years? I'm really skeptical. Maybe we can do point-to-point on highways in 5 years, but then what? Do we have huge lots where these trucks park and then drivers ferry them to local warehouses to be loaded / unloaded? What about fueling these things along the way with no drivers? How do we keep these loads safe from theft? None of these are insurmountable, but are we really going to solve these things in the next few years such that we're ready to start laying people off?
I think 10-20 years is much more realistic to put the majority of truckers out of a job. Not all, but the majority. But truckers are already old (49 average) and many will be retiring over that period anyway. Why can't we just not replace them as those jobs dwindle?
The other stat I think is relevant is that the average car today is 11.5 years old. Or looked at another way, in 11 years, HALF of the cars on the road are the cars that are on the road today. I guess if we cut down the number of cars total due to efficiencies of self-driving carshare maybe that goes down but still... If Level 5 autonomous drive were available TODAY, it's still a long metaphoric road towards complete switchover.
It's hard to tell what the average tractor age is, but it looks between 6 and 10 years old so you have the same rough math in the 6-10 years after we have fully autonomous self-driving trucks where we'll still have half the trucks on the road as human-driven.
Plus if autonomous trucks increase capacity and/or efficiency and cut shipping costs, that might also stimulate shipping demand and therefore keep human drivers on the road longer as well.
Your 10-20 years guess feels right to me, but only after we actually have the autonomous truck thing fully sorted out, IMO.
You say that in 10-20 years, tons of jobs are going away. You say it seems very unlikely the American political system is going to address it without violence in any successful manner.
But you also say that you think we will adjust and adapt and thereby keep the work a job, make a living system that we have. I don't see how that optimistic result follows from the pessimistic future you predict. And I think we're talking about the same problem - a period of volatility due to a lack of available work and a refusal to acknowledge the failures of our current system - but looking at different resolutions. I don't accept violence to restore status quo as the only possible solution.
This is what I think: Your predictions are relatively spot on, if nothing changes, but we have the opportunity to change them if we act now. If we install a government dedicated to addressing these problems, if we begin to think seriously about solutions rather than spending time repeating the same arguments about whether the problem exists at all, we can fix it. Yes, we've failed to do this on other things (e. g. climate change, North Korea, etc.) until it was more than a little too late, but the fact our society still exists and functions suggests that we must have had some success, too.
Remember that it's only in the last couple centuries that we transitioned from most countries being monarchies or what we'd now call dictatorships to most countries being at least some level of democracy. It's only in the last couple centuries we've tried systems of economy and government other than direct rule by a ruling cast. Some of those systems didn't work (full communism, fascism); some have worked out quite well thus far (modern capitalism, democratic socialism a la much of Europe), and of course many are in the middle. We can work out systems which are better still, and we can transition into them without violence or (too much) societal strife, but only if we can at least acknowledge that it's necessary or even just wiser to do so than to try and maintain the system we've got just because it's what we have.
TL;DR: You're right that we don't know what's coming, but, in my opinion, it's a mistake to assume we can't decide for ourselves if we move quickly.
You say that in 10-20 years, tons of jobs are going away. You say it seems very unlikely the American political system is going to address it without violence in any successful manner.
But you also say that you think we will adjust and adapt and thereby keep the work a job, make a living system that we have. I don't see how that optimistic result follows from the pessimistic future you predict.
Very fair, I wasn't clear. I don't think this terrible future is coming, but even if it is, and the majority come to agree that it is, we're not going to do anything about it. Even if the number of jobs does plummet and we have record-high unemployment, we still will take forever to do anything about it, assuming anything can be done. It might be that all the ideas we've come up with (which really just seems to be variations on UBI) won't work and are just rearranging the deck chairs again. Except now the AI is Amazon and we're all B&N :)
if we begin to think seriously about solutions rather than spending time repeating the same arguments about whether the problem exists at all, we can fix it
Yeah, I'm not really in favor of that. The arguments are necessary. We shouldn't be putting huge changes in place without even having the argument as to whether it's necessary.
I think you're underestimating how destabilizing and destructive these changes are going to be. If AI is going to be as bad for employment as many think, the solution is massive and represents a completely fundamental shift in economy, culture, politics, etc. For example, switching to meaningful UBI would entail huge tax increases, slashing of other benefit programs (and dealing with their supporters), and decades of fighting about the amounts, administration, waste / fraud / abuse, etc. There has never been a coalition in American politics strong enough to get something like this done. The abolition of slavery would be the closest thing, and we had to fight a full-blown bloody civil war to get that done, and we're still dealing with the cultural fallout of that change 150 years later.
I also think you're way ahead of the vast majority of Americans as to whether it's as necessary as you claim. To wit: Republicans have exactly zero interest in fixing any of this (or a bunch of other problems) and 18 months ago voters decided they should hold all three branches of government. Yes, popular vote, gerrymandering, voter suppression, etc, etc.
The point is that it's all well and good to say "enough talk, let's put these solutions in place!", but I don't think you realize how hard the opponents to those solutions are going to fight, and how powerful and entrenched they are.
The part you're missing is that jobs will also be created while others are destroyed. We lost tons of horse caring jobs as we switched to the automobile but it didn't wipe out society.
The folks who are worried about AI destroying jobs are aware of this cycle, but they think that it's different this time.
Specifically, that technology has previously allowed workers to be more productive by shifting away from manual labor and towards cognitive work that requires the human mind. But AI is a unique threat because it can match (and then quickly absolutely dominate) humans for any kind of work that requires cognitive power.
It's not a ridiculous argument, but I'm not convinced for several reasons:
1. I'm not convinced that AI can actually achieve something like the general intelligence that's required to perform in rapidly changing and dynamic environments.
2. Even if they can, there are lots of artistic and creative jobs that AI may not be able to do. Maybe we'll all be writers and poets and songwriters and sculptures in the future.
3. Also, doesn't incredibly smart and capable AI that can do any job that humans can do mean that the price of everything should plummet? Maybe post-scarcity will look different than we expect; maybe all the things we need for life will be effectively free, and companies will give them away the way they do with advertising schwag now.
4. If AI gets so cognitively powerful that it puts most of humans out of a job, we'll have either much bigger things to worry about (war with AI), or nothing to worry about (they'll either just squish us quickly or save us from all this).
At a certain point, rather than trying to radically reengineer American society for a post job world, I think are far more palatable solutions (palatable to our crazy system as it stands).
Shorter work week!
Higher taxes, spent on social services. There is tremendous need for all kinds of social services in the US. Such occupations could fill a lot of jobs. Social workers, teachers, judges, police, EMTs, soldiers, programmers, managers, infrastructure building, doctors, financial counseling, career advisors. If the Republicans could let go of their incoherent,burn the house down strategy of governing, things could really be put on a better footing.
Limitations on amount of AI that can be utilized in a company. It’s simply not viable to exclude large percentages of citizens from the working economy, and think that the technicians and managers will capture increasing proportions of the economy’s output. There’s no rule that we simply have to accept unfettered and increasingly Darwinian capatilistic competition, as it destroys society. I doubt it’s even possible in the US, as the sort of brutality required to maintain this type of extreme system is antithetical to the spirit of America. US citizens have some notion of themselves as having worth and rights.
The history of the labor movement shows that constraints on trade in the name of improving working conditions is possible.
My view is that we see the ugly rumblings of what a backlash might look like in the election of Trump. I understand some of the frustrations behind Trump as protest. I don’t think that is working that well. Things aren’t actually that bad in the US, compared to what they could be. There will be time to change course if things really start to disintegrate.
>But AI is a unique threat because it can match (and then quickly absolutely dominate) humans for any kind of work that requires cognitive power.
Only at very specific tasks. As you point out later there is no evidence that AI general intelligence where it needs to be for this actually to be a threat to humans who can adapt to unexpected circumstances.
General AI that can totally replace humans in a job isn't really the problem. The real issue is specialized AI that lets one person do work that previously required five or ten, whether because it increases their productivity directly, or the human is working as an AI-minder, letting them do the routine stuff, supervising them for quality, and taking over when they hit unexpected circumstances.
You are totally right. In the very least no politician would be willing and tell people that there no longer is a need for someone with their skill level to participate in the economy and that we are just going to give them money so that we don't have to worry about them.
>>And as soon as Amazon can make the same amount of money while employing fewer people, they will do so. Every company does. The cycle repeats, only this time more people lose and fewer people gain.
Not every company does, every public company does. It seems hard to believe in this day but there are some privately owned family-run businesses that care deeply about their employees and aren’t looking to cut labor costs as a means of profitability.
> Amazon employs way more people than B&N, the company they put out of business. Maybe those jobs should be better (OK, they definitely should), but that's different from saying there are no jobs, isn't it?
Amazon doesn't employ them, they contract certain services from them. Your first post talks about "getting laid off" but it won't even be that, it'll be not extending their contract.
Most people can still, for the time being, contribute to the economy - but already not at a level where the economy is comfortable hiring them for the long term. The "jobs" Amazon offers are explicitly designed to be automated away.
Amazon employs way more people than B&N, the company they put out of business. Maybe those jobs should be better (OK, they definitely should), but that's different from saying there are no jobs, isn't it?