Andrew Yang's ideas on automation and mass unemployment are not based in reality | Hacker News Reader

Andrew Yang's ideas on automation and mass unemployment are not based in reality

slate.com

108 pointsMilnerRoute109 comments

In a 2018 paper, UC–Berkeley sociologist Steve Viscelli suggested that in the most likely scenario, long-haul truckers, who tend to make middle-class wages, will be replaced by poorly paid drivers tasked with steering autonomous vehicles through tricky city streets, which onboard navigation systems handle less well than highways. That’s not all jobs disappearing. It’s jobs changing.

I'm not sure if the author is claiming that the job displacement will be 1:1. Does anyone honestly think it would be 1:1? I don't think so. And if more jobs are disappearing than are being created, there is a net job loss; also the author agrees that the new jobs would be lower pay. Whether jobs will be created in entirely and unpredictably new industries is besides the point. Yang's point is that everyday people are inadequately prepared to make that transition to new jobs. He frequently cites data that the US federal government sucks at job retraining.

IF various tech companies are successful in their endeavors (and yes, that's a big if), it will be true that a lot of people will need to transition to something new. And given recent demos by various tech companies for voice AI tech (yeah, Google's really was impressive when it came out, and it will probably get better until it's commercialized one day), self-driving tech (I'm a skeptic for the short-term, but it will probably happen in the long-term), and etc, I think the subject is ignored at the working population's peril. Never mind self-checkout, which is already a reality and steadily gaining ground.

Yeah, honestly I don't really understand the article's point.

Is it that say McDonalds switching to electronic kiosks to order your burger is going to create as MANY tech jobs as it took away? Because that doesn't pass the smell test.

There were a slew of anti-yang articles that came out from the leftmost media sources immediately after the debate, owing to the fact that Yang made Elizabeth Warren look unprepared in her answer about job loss.

Andrew Yang, snake oil salesman: https://www.nationalreview.com/2019/10/andrew-yang-tech-entr...

Is Yang doing more harm than good to the case for universal income?: https://www.politico.com/magazine/story/2019/10/16/andrew-ya...

Andrew Yang is Full of it: https://slate.com/business/2019/10/andrew-yang-automation-un...

There was one more I can't find. It's good to be generating hit pieces; it means someone views you as a threat.

I watched the debate, and I'm not a big Warren fan, but it most definitely wasn't my impression that Yang won that argument or made her look unprepared. If anything I would give the point to Warren for citing data arguing that automation to date has had significantly less impact than sending jobs to China and Mexico.

I think Yang has a solid argument that automation is a considerable job risk in the future, but we don't know how long that will take. I really doubt truckers are going to be obsolete so quickly. The tech will be good enough no doubt, but there's going to be a kerfuffle that'll slow it down, especially if you get a pro union candidate like Sanders.

I think Yang is a good talker, charismatic, and smart, but I don't buy his urgency... yet. That being said, I think he'd make a good candidate (personally prefer Bernie, but would happily accept Yang).

No doubt a lot of media is trying to tank him (we see this with Bernie constantly), which just makes it really hard to sort the reality from the BS and polarizes people further.

National Review a leftmost media source?

Or Politico, or Slate. When I hear "leftmost" I'm thinking like, Jacobin or Chapo

"Ah yes, noted Leftist bastion" checks notes "the National Review"

What is the basis for saying Politico has a bias in some direction?

> Yeah, honestly I don't really understand the article's point

To discredit Yang's main running ideal.

Perhaps the article's reason for being is simply that, if you publish a headline with a professional-looking article behind it, then large swaths of people will have their opinions swayed without actually reading through.

It will free up a lot of workers to do other things

Right now if you want to employ them you have to compete with McD burger-flipping, now they will be available for your next enterprise

First, jobs are work in that they’re emotionally boring repetitive tasks the bean counters have whittled down into cost saving nonsense. So fuck them: https://aeon.co/essays/what-if-jobs-are-not-the-solution-but...

Second, there is a ton of work out there our society likes to ignore the value of because it’s not shallow novelty: https://aeon.co/essays/the-key-to-jobs-in-the-future-is-not-...

And really if automation does take more jobs than we can replace, being really nice to each seems like a good idea. Cause most people won’t be needed. Are we to descend into Mad Max life or could we, you know, kill the notion alpha attitudes are useful.

I am reminded of a paper by chimp researchers who watched a community of chimps gang up on some alpha male chimps and kill them off. Then noted how their community became much more ethical after such selfish, entitled behavior was eradicated.

Applying ethical standards to chimps seems weird. By what measure did they become more ethical?

“In these crises there breaks out an epidemic that, in all earlier epochs, would have seemed an absurdity—the epidemic of overproduction. Society suddenly finds itself put back into a state of momentary barbarism; it appears as if a famine, a universal war of devastation had cut off the supply of every means of subsistence; industry and commerce seem to be destroyed; and why? Because there is too much civilization, too much means of subsistence, too much industry, too much commerce.” - Marx

Yeh. Though I think it’s not explicit enough? A bit open to literal interpretation only.

I actually see it as “too little” new information that’s of utility to the general populace.

We spend then “too much” time focused on “too few ideas that matter”.

Like what celebs eat is discussed more than environment, and on and on. Too little new progress is made to satisfy these less discussed but ultimately far more important problems.

Thinking Keynes, our demand has been nudged to focus on consumerism, celebrity, gainz!. But we maybe need to balance that with the demands of the literal majority being shat on.

I’ve seen it discussed here too, that focusing on poverty via a statistic emotionally numbs analysis, ignoring that it’s still hundreds of thousands or whatever number it is, of real people barely scraping by.

We’re fetishizing opinions that we’re obliged to import as education from our parents and institutions, which is unavoidable. What is avoidable and manageable is the content of those opinions.

Free speech is generating whatever syntax you want. You’re not owed a society that kowtows to the embedded semantics.

You can read the study here: http://laborcenter.berkeley.edu/driverless/

"Autonomous trucks could replace as many as 294,000 long-distance drivers, including some of the best jobs in the industry. Many other freight-moving jobs will be created in their place, perhaps even more than will be lost, but these new jobs will be local driving and last-mile delivery jobs that—absent proactive public policy—will likely be misclassified independent contractors and have lower wages and poor working conditions."

If tech companies reach their holy grail of what they want to do with self-driving, last-mile delivery jobs won't exist. And as I already said, yes, that's a big if. But it's something the tech companies are aiming to achieve. And besides that, I am highly doubtful that proactive public policy would enable last-mile delivery jobs to have higher wages and better working conditions. The public policy isn't there today, I'm not confident it would be there in the future. Most public policy trusts macro economic theory's invisible hand to naturally raise wages if there are truly more jobs than there are people. And if it's true, that's what will happen, then there's no need for proactive public policy. So... /shrug

All of this ignores Yang's other point that all of the roadside diners and other hospitality industry on the highways that depend on the trucking industry would be affected.

I'm not saying that the tech companies will achieve what they are aiming to do. Again, I'm a short-term self-driving skeptic. But I think it's dangerous to not have the conversation.

Those last mile delivery jobs also already exist. That market isn’t going to grow, necessarily, because long distance becomes autonomous.

Wouldn’t lowering costs for freight mean an increase in trips ?

More things can be shipped farther away therefore more last mile jobs?

If only we hadn't invented the printing press, scribes would still be making good wages. Somehow everybody found a job afterwards, though -- the economy adjusts with an increased supply of labor. UBI is a promising antidote for short term automation shocks, which Yang is also a proponent of.

>If only we hadn't invented the printing press

The difference is that scribing was a limiting factor to new and expanded methods of communication. Removing this limit allowed industry to expand due to massive decreases in costs of communication and expansion into entirely new forms. A similar argument can be made for many of the usual purported counter examples against automation alarmism. Steam engines, combustion engines, trains, cars, highways, the internet, they all were massively reducing friction and transaction costs, thus opening up new forms of commerce which brought new kinds of jobs.

Automation is a different kind of beast. The promise of the automation revolution is not orders of magnitude increases in efficiency, its removing humans from the equation. A long haul truck doesn't get across the country and order of magnitude faster, it just gets there more cheaply without the constraints human involvement. This does create a market for self-driving trucks and the industries needed to support them. But they too can be automated. The potential for making human labor irrelevant to large chunks of the economy is real and shouldn't be hand-waved away.

It really is not different, and people of the time certainly made the same exact arguments against most of the inventions on your list. Whether it's caused by a cotton loom or an immigrant, there is something about having to learn how to do a new job that really sets people the wrong way.

The only difference is that we are getting closer to the same asymptote we have been approaching for centuries now, where human labor is less about supplying material property and more about providing human service. Sure, it is possible that within the next few decades all but the most highly skilled and educated workers will be displaced into a continuously growing service industry. A decent chunk of people are not cut out for making a living on human interaction, so this is where a base UBI makes a lot of sense. It does not, however, imply that the labor force as a whole will not have any work to do.

Why would the new jobs be lower pay? If a truck driver can operate one truck while an autonomous truck operator can operate say 10, then the latter's marginal revenue is 10x more. Technology makes people more productive, which leads to greater wealth and higher pay.

Technology does make people more productive, but the claim that it’ll increase their pay is dubious. We’ve seen many times that companies will choose to keep the money rather than paying their employees more if they provide more value than their position requires.

Making productivity easier decreases the value of that productivity. Their wealth & pay may increase, but not as a 1:1 ratio to productivity.

Were that not the case, there would be no incentive for customers to select the new service.

A more results-based analysis is that technology makes companies more productive, which leads to greater profits. What you described sounds inexperienced and naive, not matching historical results without collective bargaining.

The people were just biological machines with mutually exclusive temporarily aligned self-interests.

I'm not a Yang supporter but this article is total garbage. Slate has been very disappointing lately.

The key takeaway seems to be while Yang says automation has displaced jobs there's a lack of academic consensus on the weight placed between trade policy and technology improvements; but everyone agrees both have had a significant effect.

Ok... so we shouldn't be coming up with a plan for how to address economic changes due to technology based on "it's historic effect has possibly been outweighed by other factors" even though we all agree it will continue to have an effect in the future?

Also just want to point out that anyone using the reported federal unemployment rate as a measure of overall employment health has already lost all credibility to write on the topic.

I’ll go a bit further. The article was just rambling. The article title is super click baity and the substance of the article doesn’t support it. They cited a bunch of studies to try to show that the job losses can be attributed to many things.

Why not compile the various studies and make an infographic showing the different ideas out there? It might be much easier to understand.

Agreed, there are some good points in the piece, but the headline and concluding paragraph are not supported by the body.

Curious as to what you identified as a good point? I didn’t really see any to be honest

I think Yang has been overstating automation’s impact, in part because that’s necessary for political messaging. Sounds like the Ball State study has some issues.

"[...] Long-haul truckers, who tend to make middle-class wages, will be replaced by poorly paid drivers tasked with steering autonomous vehicles through tricky city streets, which onboard navigation systems handle less well than highways. That’s not all jobs disappearing. It’s jobs changing."

Uhm, that's halfway decent careers turning into far fewer and shittier jobs. That's jobs disappearing. And it ignores that eventually someone will automate even that part away, it'll just take a bit longer.

The article just seems to just be a series of quibbles that don't show Yang to be wrong in any meaningful sense.

That's BS about being capable of navigating Interstate highways autonomously.

Whoever does the research needs go actually go along an Over The Road truck driver and actually see what challenges technology face.

Different highways with crumpling road markers. Severe weather. Unpredictable human drivers. Fueling. Construction sites and their differences between displaying markers.

I do not see autonomous technology taking over long haul truckers anytime soon, especially if the state and federal government doesn't maintain the highways and keeps cutting budget.

The few scenarios where the present technology can handle is a local route that has greatly maintained roads and stable weather. Even in this case, it's foolish to think human drivers will not be as careless as before. In fact, I predict they will think its so safe that they will cut the truck off ever closer, thus creating more disasters.

Source: I'm a trucker and know the challenges computer vision and its sensors could possibly face.

There are many routes which are possible to automate today. Those will become human-free first, leaving the humans to compete for the remaining 95%.

Then another 5% of routes will become computer navigable, and the remaining truckers will compete for the remaining jobs.

Then between automation and improved maintenance due to the higher regulatory fees afforded by automation, more roads will become auto-navigable by dint of maintenance and standardised marking.

There will be a pincer movement, roads coming up to standard, standards converging on what automated trucks need, and automation becoming better at navigating any road.

By 2030 there will be far fewer humans driving anything.

What? 10 years?

Probably caravans where you have one driver per 3 trucks will be first or something similar to that.

Exactly right. Not to mention it's the most common job in most states. That's a lot of people losing what quality of life they had.

Yes, I disagree with some of Yang's views (and his proposed solution) and the general idea that we're living in the fastest age of technological process (I'm generally quite pessimistic if one tallies up the resources we keep using, the education we have and how little change there is), but many of the criticisms of Yang seem to be worse than the caution he brings to the table.

Even if Yang is only slightly right, and I do agree on some sectors like Amazon replacing brick and mortar stores and the erosion of mid-skilled work and good jobs, that can already be back-breaking for many developed economies, and in particular the UK and the US with their strong focus on deindustrialisation have already felt the effects.

One only needs to leave London for a few hours and travel the country and go to the poor places that were former hubs of industrial production. Many of the places are in astonishingly bad shape.

Truck driving has a incredibly high turn over rate. Not that many drivers will lose their jobs but a lot who would have entered the field won't have that option. The net result as far as number of truckers goes might be the same but it is easier to divert into something else instead of retraining after investing time into a career. Of course this necessitates there being something else to be diverted into, which certainly isn't guaranteed.

Meanwhile what will demonstrate Yang wrong—conveniently for him—is just waiting a decade.

Since I like Yang and have Googled him before, Google thinks I should know about every article written about him and keeps prompting me about them as they come up.

This one, like so many lately, all have been hammering me with this idea that Yang is predicting a dystopian horrible future.

On the contrary, I think Yang is fundamentally optimistic about the future and technology. He predicts that we will lose a lot of our current jobs and that we need to put in place a government safety net in the form of UBI, but I don't think he considers it a dystopia.

I'm frustrated by all the editorializing in the headlines that I'm seeing. If all you saw were all these headlines of late, you'd think Yang were some kind of "sky is falling" candidate. But actually I like him because I feel like he has a quite hopeful vision of the country and our future.

Agreed. If we're honest about it, labor sucks for most people. You have to get up early, spend the day doing something you don't really care about, and get home too burnt out to enjoy even the rest of your day. It's absurd to think that liberating people from that labor regime through technology and a new social contract must be dystopian.

I wonder what the source of those articles and the "dystopian" sentiment is. Do people reach these conclusions on their own and go out to feverishly write about it, or are they told to do so by funding/larger agendas who feel the emotional reaction from such headlines will benefit them (in the presidential race)?

I read the whole thing. It’s not so much that Yang is wrong but that the automation-is-taking-all-the-jobs mantra is incomplete and that outsourcing to places like China is a substantial part of it. From what I can tell from Yang-like arguments outsourcing could be seen as a form of automation. The general idea is that companies are incentivized to reduce labor costs so whether it’s outsourcing or automation the end result is the same. Presumably if we restrict outsourcing then just more investment into automation will take place. It’s inevitable.

The most charitable way of looking at this is that yang wasn’t wrong, but he wasn’t right either. He made explicit claims like “this isn’t a rules problem” in response to trade policy; actually, there’s a lot of evidence that there _is_ a trade policy problem.

This is also the corporatists argument for a high minimum wage - the higher the cost of labor the more aggregate pressure to automate it away. And the more abrupt the upset the more likely the necessary cultural shift occurs. The alternative, generational slow burn automation, would just perpetuate the current state of affairs - new generations being poorer than their parents gradually till the two classes - capital and labor - have totally diverged into post scarcity luxury and destitute helpless permanent poverty.

I stopped reading after the author mentioned that Yang thinks that Ubi will solve issues. Have you actually read into some of the things he said, Ying explicitly says he does not think it will solve everything. Rather it gives people a bit of leeway in the short-term to focus on other things. With the hope that it will lead to more entrepreneurial things, but knows more likely it will go towards random things in each individual's life

I think the best summary of Yang's view of UBI comes from his own FAQ [1]:

> We are experiencing the greatest economic and technological shift in human history, and our institutions can’t keep up. Without the Freedom Dividend, we will see opportunities shrink as more and more work gets performed by software, AI, and robots. Markets don’t work well when people don’t have any money to spend. The Freedom Dividend is a vital step to helping society transform through the greatest automation wave in human history.

[1] https://www.yang2020.com/what-is-freedom-dividend-faq/

> over the very long term, automation probably has played a role in limiting manufacturing’s share of employment in the U.S. Factories really have become more advanced and efficient since, say, the 1970s.

Can anyone at Slate read a chart? US employment in manufacturing as a percentage of non-farm labor dropped from 32% post-WW2 to less than 10% today [1]. This change didn't start "sometime in the 1970s", but has been a consistent downward trend post-WW2. The low hanging fruit of automation (from mechanical feedback mechanisms to programmable logic controllers) are all pre-1970s inventions.

And so what if different jobs are created for those that are lost? The article itself admits those jobs will either be lower paying or require more advanced education. Job re-training is mostly a myth [2]. No matter what, an entire generation of unemployed will suffer. I grew up in Detroit. I know what happens when the jobs base shifts.

[1] https://fredblog.stlouisfed.org/2014/04/the-decline-of-manuf...

[2] https://www.nytimes.com/2017/02/23/magazine/retraining-jobs-...

> And so what if different jobs are created for those that are lost? The article itself admits those jobs will either be lower paying or require more advanced education

And even this will only last so long. With ever more advanced robotics and AI, there will be fewer and fewer possible lower paying jobs as well as advanced jobs that average Joe could even be educated for. Humans have finite muscle power and brain capacity. Is it not obvious that as technology becomes more potent, we'll reach a point where it just doesn't make sense anymore to make the masses spend their days working bullshit jobs?

Yes, self-driving cars and trucks will very quickly unemploy 3.5 million truck drivers (US only), 3 million Uber drivers, 1.4 million Lyft drivers.

And that's just one area.

I agree on retraining. An average driver will not become a surgeon or a programmer (or whatever other hard-to-automate jobs are left).

> Job re-training is mostly a myth

Job re-training is more than a myth, it is another Reagan-era privatization scam, designed to funnel public money to private for-profit training centers and trade schools. Gordon Lafer's _The Job Training Charade_ is a great book on the subject.

> Following the debate, a “fact check” by the AP claimed that Yang was right... “Economists mostly blame [manufacturing] job losses on automation and robots, not trade deals,” it stated. But this was incorrect. No such consensus exists

So they claim AP and Yang are incorrect but also state there is no consensus. This piece seems more like it is trying to justify the warren gaffe in the last debate by muddying the waters. The truth is be it automation or outsourcing our economy is changing and we need to be prepared for it.

I would like to know which mainstream economists blame trade deals. Ideally stuff that was published more than 5 years ago.

Yang is totally wrong about automation. The main cause of the shift of manufacturing from US to China is trade policy and China’s monetary policy.

If automation was the problem, then US would be price competitive on goods with China. Somehow US automation is only good enough to eliminate jobs in the US, but not good enough to compete with cheap labor in China.

Manufacturing is probably the worst example because trade policy has definitely had an impact there. However, look at other industries: Tier-1 tech support being replaced by chat bots, order-takers being replaced by kiosks and apps, realtors being replaced by self-service apps (ie redfin).

Whether by outsourcing or automation, the cause is the same: Loss of low-skilled jobs in return for cheaper products & services. The challenge is to find the best method to leverage those cheaper products and services to improve quality of life for everyone, especially those who's skills become obsolete.

Sorry that is just plain wrong. As of 2019 support services are mostly outsourced to India/Philippines or other places. US call center jobs lost to chatbots vs India is negligible at this point.

Redfin has a tiny fraction of real estate sales, and hasn’t broken into it anywhere locally, or is showing any real signs that it will. Turns out it’s a little harder to self serve a home purchase vs a cab or vacation rental.

Whatever automation you can do in the US you can also do in China but with lower regulation. Whatever labor costs left over are also cheaper. What I mean is that automation is not an inherent advantage of the US.

Absolutely. In fact industrial automation components (robots, servos, linear bearings, etc) have been in huge demand from China. Suppliers often sell the same components there for half the price as they charge in the US for the same quality and quantities. Domestic Chinese companies are becoming major suppliers of a lot of factory automation equipment like presses, laser welders, coil winders, etc.

There's no reason to think the US has an advantage in automation. In fact, it's much easier as an automation engineer to design automation processes for the factory that's around the corner than to do so for a factory in another time zone that 90% of your team has never seen in person.

Nowadays 100%. 20 years ago China didn’t have a high technology sector that could have supported large scale factory automation, and factories where built to scale labor.

100% agree with regulation. That’s why US trade policy is the main cause of the decline. It should reflect regulatory arbitrage. It’s utterly pointless to save the environment in the US and poison it in China.

The US's position as a prestige currency of choice certainly contributes, too. Foreign savers really want dollar safety, so the US has a net capital outflow in the form of Treasuries, which means it must take a net inflow of goods and services. The end result is that American consumers pay less than they should for goods, all because other governments can't or won't provide the safe assets that their citizens desire. These low consumer goods prices are definitely not good for domestic manufacturers.

> If automation was the problem, then US would be price competitive on goods with China.

This (and the article's similar approach) is only half of the equation.

'If automation was the problem and cheaper than offshored labor, then US would be price competitive on goods with China'

There are two independent variables: the cost of offshore labor (China recently, SE Asia and Africa now) & the cost of automation.

Businesses are going to pursue whatever is most cost effective.

But people are incorrectly ignoring automation because it isn't a problem now.

Put more restrictions on offshoring labor (thereby raising its cost) and automation is going to start looking very important.

Don’t forget persistent currency manipulation meaning that goods manufactured exactly the same way with the same cost of goods sold would be 30% cheaper in the USA because of the devalued Yuan.

it is a characteristic of technology that what costs too much in one generation will in each following generation cost less - at some future point the quality of automation found in the U.S now will no longer be prohibitively expensive in China.

I generally favor UBI because it's basically wealth transfer from the wealthy to the less wealthy and I support those policies.

UBI seems like a poor response to the shifting nature of work. Yang himself implies that automation will "come for" some groups before others. I.e. subsequent subsets of the population will lose 100% of their income. Increasing the income of everyone seems like a poor response. If we want to support workers who worked or trained for a job that has been automated away why don't we...do that?

If we want to overhaul our society into a place that no longer forces people to work to live I'm also in favor of that! It's just different.

Because UBI is the only welfare system that can't be gamed. How do you identify "workers who worked or trained for a job that has been automated away"? Non-universal welfare wastes a lot of resources on conflict between potential welfare recipients and the state, and it inevitably mis-categorizes people, which causes suffering either directly when the "deserving" are denied it, or via resentment when the "undeserving" are granted it.

>Non-universal welfare wastes a lot of resources

It seems like you're saying that UBI is more efficient we don't need spend resources on targeting it. I 100% agree for the general case of wealth transfer. I think it's very weak for addressing specific job losses.

So, let's say we have 100 people and 10 of them lost their jobs to automation but we don't know which. One option is the "Yang plan" of $12 a year ($1 a month), so we have a budget of $1200[0]. That spends $1200 giving $120 of assistance to the people we want to help. This plan is 10% efficient if the point is to help those who have lost their jobs. In fact, with a budget of $1200, we can spend up to $108 per person identifying if they lost their job due to automation and have the same efficiency (10%). That is a really high ceiling on the cost of conflict.

I'm very skeptical that in any real world situation the cost of identifying worthy recipients is greater than the "loss" of giving everyone else money. Again, only if the goal is helping workers who lost their jobs. Hell, even if you're wrong 50% of the time and give 15 people $12, you would need your method of identification to cost $60 (5x the benefit) for it to be 10% efficient!

Again, I favor UBI and I favor transforming society to function differently, but that's not helping workers who lost jobs.

[0] These are ofc not the real numbers Yang is suggesting

I always assumed part or perhaps most of the funding would come from flattening the tax system.

Yeah. I feel the same as you. On a macro scale distributing funds to middle and lower class is a net gain for the economy since I don't believe the trickle down effect works and those groups will spend the money and keep it flowing through the economy. It just makes sense to me that the economy is stimulated by demand instead of supply.

However, I agree UBI seems to be going to people who may not need it. However, I like the simplicity of it.

If you start into differing requirements down the line those will accumulate and you hit a slippery slope and it devolves into our current tax code nightmare with thousand pages of what it takes to be qualified for UBI.

So it's heavy handed but simple, which I like.

> It just makes sense to me that the economy is stimulated by demand instead of supply.

Maybe my experience is atypical, but I've seen, over and over, hiring driven by either 1) we have more demand than we expected, or 2) we have good reason to expect more demand and should hire now to support that.

I've never seen "we have money, but no extra demand and no reason to expect any, so we better hire". "We got a tax cut but have nothing for extra workers to do—better hire anyway" said no business owner ever.

Which is to say I have no friggin' idea what supply-side is going on about. Yeah there's probably some hypothetical situation in which growth can be supply-constrained, but in the presence of demand that problem will fix itself pretty quickly, guaranteed. The opposite clearly isn't true at all.

China or not, most of the jobs won't come back, they were eliminated by automation. The bottom line of the so called UBI is the redistribution of wealth. It's not only for individuals, but regions too. I somewhat image that the money you give to individuals is able to encourage them to move to regions where there are less jobs, but lower living cost. meanwhile, the newly moved in individuals are able to revive the region, creating new local business and jobs...etc.

Notably, jobs in China are and will be automated away just the same as in the US and everywhere else. Technology and science are universal, you can't just put a border in and think you're insulated. Apart from why one would even try to obstruct progress?

I don't think Yang is claiming the _only_ thing causing job loss is automation. It's just one major factor, which will continue to rise into the future.

I also don't think he believes it will put everyone in those industries out of a job. I can see how the sound bites from the debate sound that way, but unfortunately he probably has to use the worse-case scenario type examples.

Use your eyes. Have you walked into a Supermarket? Or better, when's the last time you saw MORE retail store hiring MORE help and not actually cutting (because Amazon's choking them off). Better, ATM machines are now charging $5.00 per transaction fees because cash is no longer needed. Automation is a bad word for all this -- it should be called simply "job replacement" -- and it comes in all forms of technology than can help us in many ways, but also has ramifications like job loss. People need to use their eyes. He's not wrong. He's not full of it.

That's a strangely angry article and seems to miss the bigger point.

America changed its trade situation with China, which either made the US better off or worse off. (I'm guessing the economic consensus for is better off).

Basically the same applies to automation.

No individual factory worker or doctor or police officer made that happen, so the benefits and/or costs should be spread around equally. UBI is one tool for doing that. People who lose their job and need to retrain automatically get cushioned, which lets your society be more ruthless about creative destruction.

> either made the US better off or worse off.

It's very hard (meaning impossible) to make such an assessment for the US as a whole. Some groups net benefited from globalization and trade liberalization, others net lost, for many it was a wash.

Remarkably, this is one of the predictions of mainstream economists (MSEs) that is empirically true. What they should have stressed more, therefore, is the distributional aspect of trade liberalization: that the gains from trade must be distributed to compensate the losers. The gains are so large that this could have been done and still leave something over for the natural beneficiaries of free trade.

What Yang is saying, and I TOTALLY agree on this point, is that automation or progress in AI is like that only ten times worse. We all know that tech is full of natural monopolies, while the economic gains are HUGE because they are aggregated over so many people/transactions/etc. and marginal costs are low.

Therefore we should think about reorienting our economies (talking globally here, not just US) to ensure that we don't see the same that happened with trade now with automation and AI. Namely huge segments of the population left behind "because market" and all the gains accruing to a handful of data oligarchs.

I develop key performance indicator software for my company's manufacturing plants. It's basically a real time analytics dashboard that also trends data over time. One of the results of this software has been a better understanding of the efficiencies (or really the lack thereof) and bottlenecks in our processes. Identifying and understanding these KPIs has allowed our pilot plant to cut significant labor costs without losing any productivity (no one was fired, the positions in question weren't refilled due to attrition, which is always high in a manufacturing job).

My point being, it isn't just robots and automation that reduce the labor workforce. Better data and analytics are also having an effect, a trend I don't see slowing down. As the market continues to shrink for the blue collar workforce, something will definitely need to happen so as to not disenfranchise large groups of people. Whether a freedom dividend is the answer or not is up for debate. I think it will help more than it will harm. There's always going to be someone who finds a way to game the system (the perceived welfare queens). For many, though, an extra $1000 a month would literally be life-changing.

At the risk of continuing the non-conversation here, I feel like it's important to consider countervailing long-term trends. I was listening to a fascinating episode of the Macro Musings podcast with Alex Tabarrok of George Mason and he brought up a very thought provoking point: the largest single group of workers was the Baby Boom generation, the next largest is (broadly speaking) their children, the Millenials. The number of Americans in the workforce is currently at its peak for about the next hundred years simply by number of people in their prime working years. This means that unless automation vastly outstrips the decline in the number of prime age workers, this problem won't directly be nearly as bad as people are saying. This is further reinforced by the demographic trends among the changing generations of workers -- Millenials aren't going into trucking or welding or other not-currently-automated jobs at nearly the rate their parents did.

The obvious followup to that whole point is "what are the comparative effect sizes of all of these various causal phenomena?" I don't happen to have the answer to that, but I feel like it's worth pointing to at least some mitigating factors.

The article's most compelling point for me is the graph it links to showing all U.S. manufacturing jobs from 1939 to 2009.

https://fred.stlouisfed.org/series/MANEMP

There's a big, sudden drop in 2000.

It also cites a paper that says that the number of factories in the U.S. has been dropping (between 2000 and 2014)...

I don't understand the weight assigned to factory closures. If I could make all the product in one robot-factory vs 3 people-factories, I would expect 2 factory closures. I think it must be coupled with the assertion that productivity isn't rising. But I think it is very interesting to look at the graph you linked compared to the graph for non-supervisory employees:

https://fred.stlouisfed.org/series/CES3000000006

>There's a big, sudden drop in 2000.

actually it is in 2001. Also in 2008. The years of big busts.

Yeah, but the jobs never come back after those drops. (It's almost entirely a one-way trend over the last 19 years...)

There's a big three-year drop that starts in 2001. Then it goes from that steep drop to a slowly-declining drop -- before that next massive "2008" drop (which lasts through December of 2009).

So it seems plausible that the loss of jobs was caused by specific events. (Whereas if it were the creep of automation, it seems like it'd be happening more gradually, factory by factory....)

I guess you can argue that the economic hardship of a bust forced companies into sudden and widespread cost-cutting measures like replacing humans with robots. But another study cited in the article argues against that theory. (The article acknowledges that automation has taken some jobs, but that it's a much smaller fraction, and that latching onto that as the single explanation makes us miss the other factors -- as well as possible solutions.)

The real threat to jobs is business process automation. The ideal is to change business processes to be like self serve gas stations where in the usual case the customer does all the labor required and the individual transaction never requires an employee's attention. Better sensors and actuators, embedded technology, and artificial intelligence to handle all the corner cases of the business process will allow 100% flow through for many other business processes. This will result in a significant decrease in administrative, back-office, and low level managerial jobs.

Trucking is just a nice example that is easy to explain to people. But the overall trend is much broader and more complex involving 100s of thousands of IT workers incrementally removing costs.

> Meanwhile, the overall job market is fine...

In constant (inflation-adjusted) dollars, according to the US government's Bureau of Labor statistics, current earnings for production, nonsupervisory employees on private non farm payrolls is $23.65, in constant inflation-adjusted 1983 dollars for September 2019.

In February 1973 that was $23.33 an hour. Almost a half century of economic growth, in a good economy, and workers are making only 32 cents more an hour than they did 46 years ago.

If pundits were describing wages in the Soviet Union, this would be called 46 years of stagnation and so forth. Whereas we have someone from the Bill Gates funded corporation Slate telling us "the overall job market is fine" and other niceties.

Wondering aloud if there are any studies that figure out how much these workers should be earning proportional to the productivity increases in the last 40 years?

What is the “cut” workers can expect from their output charted through the last century - this number must exist

Even in China labor is getting more and more expensive. So sure there was some loss of manufacturing jobs to China in the 2000s but now even China is losing it to automation and outsourcing as labor rates go up.

This complaint about robots and AI taking all our jobs is hundreds of years old. Still hasn't happened. US is at record levels of employment. Humans are terrible at predicting the future. Let's just wait and see and react when it actually happens. My guess... there will be plenty of jobs for humans for thousands of years to come.

Ever since the industrial revolution jobs have gone by the wayside. We no longer have wagon wheel makers, or people cutting pins from long spools of wire. If you just look at the jobs going away, you're missing the other side of the picture, the jobs being created.

Jobs are essentially just things that humans do to improve our lives. If you imagine that you were infinitely rich, what kinds of things might you want someone to do that isn't getting done? I've got tons of housework and yardwork, construction work, programming ideas, hell my brain is overflowing with stuff that will never get done. Too many jobs. All those things are jobs. Most of them cannot be automated anytime soon.

If you just look at the jobs going away, you're missing the other side of the picture, the jobs being created.

I don't think Yang is missing the other side of the picture. He's saying that society sucks at getting displaced people to the other side in a humane manner.

I don't disagree with that.

I take real issue with the unemployment rate claim. The federal U-3 rate is a politicized number that has over time been reduced to an incredibly narrow definition. U-6 is much better because it includes discouraged workers who have given up on seeking employment and part-time workers who want full-time (an increasingly large population). In 1994 the BLS changed the U-6 to exclude long-term discouraged workers which had the effect of lowering the rate. As John William's Shadowstats shows, when you adjust long-term discouraged workers back into the U-6 rate unemployment is actually well over 20%.

http://www.shadowstats.com/alternate_data/unemployment-chart...

    const capital = () => [capital(), capital()]

Humans are an implementation detail in that scheme.

Economic power (the ability to extract, move and transform resources to perpetuate the economy) is gradually moving from people to capital. At some point in the future, it will not make sense for the industry to keep on feeding people, and people won’t have any power left to do something about it.

The jobs you mention are peripheral things meant to keep meatware happy while it is still economically relevant.

There are people who are starving today, whose will to live and imagination is as strong as yours. It turns out they are not economically relevant, and thus we let them die. I don’t expect an automated industry to care more about about us than we do about the unfortunate folks I just mentioned.

And global warming is a scam as well right? Weather seems fine to me.

Just more of a smear against Yang. The establishment is afraid.

Also the tint of anti Asian American racism - similar to SV when the hiring team or management group tries to talk themselves into finding reasons to not like a qualified candidate that does not fit the mold of the obediant Asian worker drone. Usually done insidiously to cast doubt on a person's character etc.

The other part of the equation is that retraining are a complete failure. The idea you can teach a truck driver or a coal miner to code is a fallacy.

Retaining programs have a 0-15% success rate. It’s basically useless. What happens is that middle aged people get their jobs automated away and they are left with nothing. Look at what happened after manufacturing jobs left the Midwest. This is the bigger problem is that large swathes of the population will have no viable jobs after automation.

I’ve done some consulting at quite a few bigcorps in the Houston area. The amount of outsourcing of IT generalist work that I’ve seen is downright scary. Regardless of political affiliation, this should be front and center for both parties. It is the go to move for executives that want to spin stagnant growth into profitability. I generally disagree with tariffs. However outsourced work should be subject to a 100% tariff so that it’s disincentivized.

Don't know how quickly it is coming but I think ubi is the right direction for civilization. People are fighting tooth and nail for meaningless bullshit jobs that low key function as a form of government assistance but waste your entire adult life in the process.

I could understand if they questioned universal basic income or whether Yang's 1k/m will do any good, but plainly denying what is happening before your very eyes indicates a hack or a hit job. Also, focusing on manufacturing jobs alone is simply misleading.

If the need for low to moderate skilled labor is going away, why are we still importing so many workers in those skill ranges?

So the article is probably correct that trade has been a bigger issue than automation. Well let’s assume that’s true, to be charitable to the mentally challenged Slate authors.

So what happens after you put the screws to trade? That certainly isn’t going to decrease automation as an alternative. What happened from 2000-2010 doesn’t really matter, they should save their breath.

So I guess this means that Stephen Hawking was wrong as well?

Andrew Yang is the Trump of tech. And his brand of politics is more a dangerous threat to democracy than Trump’s is. His messaging and policies are just like Trump’s in that they’re dumbed down, optimized for social media likes and subtly enforce a world view where extreme income inequality is inevitable so let’s make the best of it. AI/tech isn’t going to kill the middle class by itself. This destruction will be assisted by technocracy propagandists like Yang.

Obligatory CGP Grey “Humans Need Not Apply” spruiking: https://youtu.be/7Pq-S557XQU

Deals with the issue of why current types of automation are different to industrial automation an why we won’t be creating new “machine fixer” jobs when the new machines take our old jobs.

“You're creating a slippery slide transfer straight into socialism”

This seems to beg the question: "Would Amazon be less successful if it paid a living wage to those who labor in its warehouses and those who work to manufacture the products it sells?"

1. Does Amazon get away with paying warehouse workers less than others?

2. Are you presuming no one else will pay them a living wage either so they MUST work at Amazon?

3. Are Amazon warehouse workers prevented from getting living wage jobs elsewhere?

Those are stupid excuses posing as questions, but I'll play along and answer them...

1. Yes 2. No & Yes 3. Yes

And therein lies the real problem with Yang’s outlook. It’s not just unrealistic. It’s lazy.

And this is why Yang is the most prescient of any of the candidates. He looked at the problem, broke it down into its basic components, extrapolated what was going to happen, and found the most straightforward answer (UBI) from the large search space of possible solutions. You know, like a computer programmer would.

So I fundamentally disagree with the article for the following reasons:

* If I had a budget of a few million dollars and a small team of geeks like me, I could write a (non-general) AI myself in 2-10 years that would pass the Turing test in any specific area, a bit Like IBM's Watson. And I'm nobody. The reason this won't happen is that I don't have the social proof (I haven't won the internet lottery or proceeded through the graduate level accolades or submitted papers so anyone would listen to me) so I'll likely spend the rest of my life making rent instead of making any meaningful contribution to society. And there are a million software developers around the world just like me. And a million architects who will never design a building. And a million biologists who will never cure a disease. This is the great tragedy of the commons of our time, that creates the underemployment that contributes to the wealth inequality that is the one thing holding AI back. You know, the very thing that Yang is trying to solve with UBI.

* I agree that shortsighted trade policies started by Reagan in the 80s with his "I'm from the government and am here to help" deregulation anti-pattern, exacerbated by Clinton with NAFTA, and beaten like a dead horse with George W Bush's offshoring/tax cut for the wealthy blue pill orgy, is the second biggest reason that we may have passed the point of no return on preserving (at least American) jobs. But half the US population won't concede this point or think outside the box for empathetic solutions, so I don't predict that this will be fixed before the singularity. It's a moot point now.

* I predict that the tech industry will go the wrong direction with AI, in roughly this order: industry invests heavily in cybersecurity and profit-driven artificial intelligence research to better exploit users by leveraging their shopping addictions, social/mobile bubble pops as independent developers are undermined further by receiving an ever-smaller slice of the pie, a global recession rivaling the dotcom crash and housing bubble pop hits which halts progress in social causes similarly to what happened during the lost decade of the 2000s, the US doubles down on trying every wrong thing before being forced to do the right thing (further tax cuts for the wealthy, privatization of remaining government services, no UBI because there is no money for that, further regression into service industry feudalism), sometime between 2025 and 2030 the world comes out of recession but now all money passes through an online gatekeeper (whatever follows Amazon) as the world transitions from sovereign nations to having its monetary policy determined by corporations, Stockholm syndrome far exceeding the Trump phenomenom grips the world in luddite-inspired rejection of magic technologies like AI, climate change falls from the forefront of concerns as simple human incursion into every last wild place on Earth creates an instability in civilization due to having so many insurrection cells crossing national boundaries that civic discourse gives way to political infighting and canabalistic government policy, a third world war or dystopian new world order or singularity hit all at about the same time between 2040 and 2050 so that only the rich and powerful have oracle-level AI technology and live in gated communities while the world burns, the masses find that they still need food in VR and then rise up and something happens which unfortunately can't really be predicted past this point.

I really hope that I'm wrong. But the last 20 years have been so bizarre compared to what I thought was going to happen while watching the Matrix in 1999 that I'd argue that we're already living in the early stages of dystopia. So I think we have a year or two left before the next planned recession takes the choice away from us. If even Hacker News is skeptical of the things that Yang is suggesting, then we are well and truly hosed.

> If I had a budget of a few million dollars

But 1k a month would make this happen?

If 1k UBI is causing this tragedy of the commons then you owe it to yourself to fix this, earn more, move to LCOL area and start curing diseases

$1000 a month 12 months a year 10 people team 10 years.

That's 1.2 million.

Sounds like this is a paid article for an agenda. Yang sees where automation is going and as far as I know he's the only candidate talks about automation and coming up with solutions.

Reply on news.ycombinator.com