Wall Street Has Hit Peak Human and an Algorithm Wants Your Job
bloomberg.com
bloomberg.com
I'm still skeptical.
Ultimately, work is about solving problems which is a complex and multi-stage process: someone must identify the problem first, someone must formalize the problem into a domain in which algorithms can be applied, someone must code or train an automated solver, someone must deploy and maintain the code...
I expect that a lot of future work will be what we would call today "managerial" type of work. People will identify problems and manage available resources to achieve the desired outcomes. In the past, a significant amount of those resources were other people. In future, more and more of the resources under management will consist of computing power, datasets, robots and only very few other people.
And so I doubt we will run out of work to do. This would mean that somehow we ran out of problems to solve. On the contrary, we appear to be flooded with them: cancer, infectious diseases, global warming, clean energy, access to space...
Maybe in some very far away future we will have a god-like super-AI that will seek and solve problems completely autonomously with no need for our involvement. This is very far off though.
I'm less sure. Not so long ago self-driving cars were considered an impossible feat and DARPA self driving cars kept ramming into fence posts...
>a lot of future work will be what we would call today "managerial" type of work
And what of the 80% of the population that has no experience (ability?) to manage complex systems?
It probably won't be an issue soon, but it seems inevitable that a decent chunk of the population is unemployable.
I think people are smarter than most pundits believe, and will either adapt and learn new skills or opt to live simpler lives.
Oh, no, living off the land! How could anyone ever do that?!
Err, not really. A very significant amount of work is "grunt work" which we haven't automated because it's too capital intensive or for other reasons - that is, many humans simply take the place of relatively simple robots where it wouldn't be cost-effective to build said robots due to short lifespan of each specific type of job, or other issues.
As robots become more generalized, it becomes more cost-effective to fire the people. The programmers, the analysts, they're not the people anyone's worried about.
In the UK, huge portions of the country still haven't gotten over the effects of Margaret Thatcher moving us from a manufacturing economy to a service one. Doing that on a global scale is only going to hurt even more.
Which as a German sounds strange in my ears. We are proud to have factories.
Take the steel industry crisis in the UK at the moment. The steel mills were having to close because they can't produce steel at a competitive price. No-one is buying.
There are around 35,000 workers, and the figure for the government nationalising the industry was about £10bn. If instead, they let the foundry collapse and get sold off for pennies, they could use that £10bn to pay each and every worker over £285,000 each... That would be plenty enough to see the older people safely into retirement and a lot of the middle-aged into early retirement. The younger ones could use the money to retrain or start a business. It's a huge chunk of cash.
What happened instead was they scraped together a private buyer with huge government subsidy and now people have jobs. Or at least some of them. It doesn't change the profitability of the steel... But apparently it's the job that's important, not the outcome.
I'm sure the rest of the country wouldn't be too happy with that - Do we all get a safety net like this?
And you think young steel workers are any more likely to start viable businesses than any other young person?
Standard British crab mentality. You can waste £10bn nationalising a money losing industry for the benefit of just 35,000. But if those 35,000 were to get the equivalent cash in hand, nooooo....
The cost is the same however. And the payout plan is socially far better for the soon to be made redundant older workers. Sure the younger workers idea is a bit more sketchy - they might waste it, they might not - but it's better than just being laid off and sitting on the dole. Don't forget that this is all in just one location, without the steel mill the rest of Port Talbot will suffer also. Dole recipients don't typically have much money to spend.
The cost is the same, but the purchase is apples and oranges.
> it's better than just being laid off and sitting on the dole
Is it? some will find new jobs. Does the average dole recipient cost net more than £285K?
Why do steel workers deserve such an offer, as opposed to any dole recipient?
> without the steel mill the rest of Port Talbot will suffer also
Broken window fallacy. Lots of things have side effects, this is just one angle. Port Talbot needs to diversify, like any other place.
How's it going to diversify if all the money and labour's tied up in the steel industry instead of actually being spent on diversifying? Everybody's back to working in the steel industry just as before, and it's not like any of the workers have the money or training to bring new industry to the area. Oh wait, they could've had just that.
> bring new industry to the area
What keeps them in Port Talbot if they got a handout? Why does Port Talbot deserve this money more than any other place?
Exactly how? Who's paying for their retraining? We've wound up in a situation where an industry's about to die and no systematic plan for what to do when it does - again.
> Why does Port Talbot deserve this money more than any other place?
Why did the private buyer deserve a subsidy to take over the company? Why wasn't the whole thing left to die, if we're thinking of a way to spend as little money on actually helping people as possible?
Why would any one else pay for their retraining? If they can't do anything else, they drop a resume at mcDonalds - or take the same retraining offers any job centre might offer anyone else. that's what happens when you pin your career on a failing industry with no backup plan.
The government didn't kill steel, why do the steel workers get better treatment than anyone else who loses their job?
> Why did the private buyer deserve a subsidy to take over the company
Who knows - do you?
The steel workers are no different from anyone else on the dole, no more deserving. If you know of anyone willing to take over the company without subsidy, you might have a point, but there is no special treatment there.
And not wasting money excessively is different from "spend[ing] as little money on actually helping people as possible".
Because otherwise we have an entire industry's worth of people who provably want to work (or they'd be on the dole already) who very well might not be able to.
> If they can't do anything else, they drop a resume at mcDonalds
It's actually surprisingly difficult to get a job at McDonald's if you're not a deadbeat who has no future and has done nothing in their life.
> that's what happens when you pin your career on a failing industry with no backup plan
Believe it or not, not everybody is in a position to up and retrain on a whim. Some people actually have hard jobs and families and have neither time nor money to do so.
> The government didn't kill steel, why do the steel workers get better treatment than anyone else who loses their job?
Generally speaking, we don't lose industries very often - people can find work that's somewhat similar elsewhere. When your IT company shuts down, you'll no doubt find another IT company to work for. When we do lose industries, there's a problem in that suddenly, a whole lot of people's experience counts for nothing, and that's a special case that needs to be treated specially.
> Who knows - do you?
Because apparently it's more reasonable to outright give money to businesses than to people, and people defend the Government on this.
> And not wasting money excessively is different from "spend[ing] as little money on actually helping people as possible".
How is subsidizing an industry that can't exist on its own not "wasting money" by your measure? We could buy steel from elsewhere just fine, and obviously we were, or Tata wouldn't have exited the industry as they'd be making tons of money.
Not everyone has to, some people diversify.
I'll ask it again - why are steel workers more deserving of more than the average job centre?
I'm not falling for your "deadbeat" strawman. If the IT bubble burst, and I was out of a job with no backup plan, who'd be rushing to put me ahead of any other dole recipient?
> Because apparently it's more reasonable to outright give money to businesses than to people, and people defend the Government on this.
You're fond of hyperbole, and now this. Yes, it is more reasonable to give it to businesses. Yes. In fact your suggestion was that the steelworkers would start businesses themselves with the money, was it not? You didn't respond to that me questioning what stops steel workers walking out the door.
> How is subsidizing an industry that can't exist on its own not "wasting money" by your measure?
Maybe it is, who knows; what's your the point? That if this investment is a waste, we might as well waste it some other way?
I don't know how wasteful the government plan is, but it's certainly less so than you cash-in-hand scheme.
At the very least I wouldn't be saying it's a great idea to spend billions of dollars propping up your job when nobody wants your company's services at a price the company can afford to provide them at, and if we have to spend billions of dollars, we might as well spend it on retraining people.
> Yes, it is more reasonable to give it to businesses.
Wait, precisely why? What's the upside of propping up unprofitable businesses? There's basically two - ensuring that the industry stays in the country, and jobs. Ignoring the first due to it being nearly impossible to put a price on it without knowing what our trade risks re other countries producing steel are, why do we care so much about jobs that we're willing to spend money keeping them around? Well, the answer is that it's essentially welfare that we force people to work to get.
> Maybe it is, who knows; what's your the point? That if this investment is a waste, we might as well waste it some other way?
If this investment's primary goal is to keep people in jobs, then we might as well provide people with the ability to move into jobs that the economy actually requires, instead of ones which it's categorically stated that it doesn't require.
Trouble is, many of these types of problems are not profitable to solve as the beneficiaries lack the power (read: money) to demand a solution.
That's what I'm worried about, is when there's tons of problems that could be solved, but few people have both the means and the will to solve them.
I hope I'm wrong.
The problems left to solve can be tackled by an increasingly smaller percentage of the population. The guys that currently work as taxi drivers or paper pushers can't be reeducated to do cutting edge research.
That's why it's different to the old luddite fallacy. Machinery and robotic factories throughout the ages have cost millions to replace 100's of jobs. Every time you want to build a factory, it costs money.
With software that is not necessarily the case. You write it once and then distribute everywhere at little to no cost.
Automation requires really understanding the process and formalizing it. It requires knowing all possible failure points, or at least having a "A Bad Thing Has Happened" failure point of last resort and learning the details of those over time. This will annoy people no end.
The cost of bad software can be substantial.
Nothing fancy. "machine learning" has a lot of space taken up by fancy filtered "running averages" and other curve fit ... things. Then there are Kalman filters. I can't even categorically say that what I did was or was not a Kalman filter.
Once you do that, then there are degenerate and ... overflow cases that propagate like fruit flies. You keep talking about "use cases" in the presence of people for whom that is foreign.
The hard part is then accountability. One "thing" can take two smart people who know this stuff well hours to hammer out.
Never say never, but because of observer bias and path dependence, it doesn't feel like that gets any better. At some point, a GA has to admit it is powerless in the face of its own inadequacy and ask a grownup what to do. So what you're doing is distilling the set of ... opportunities for difficulty into an increasingly higher proof.
I tend to classify a lot of learning computer problems as pre- or post-general AI. As in, would it essentially take a human equivalent AI to solve this problem or can we do it sooner with shortcuts and good programming?
ie. I used to think that computer vision was completely in this post-AI field, but the recent huge strides that have been made in deep learning etc. have changed my mind considerably. Human-like ability to recognize and differentiate objects is a major step forward in replacing humans around the workplace.
The problem is that a huge part of the population won't have neither the skills or the inclination to work on hard problems.
At any given time slice, skills in a given field are a pyramid. If we assume that the unsolved problems require the highest level of skill to be approached and solved, then there will _always_ be a shortage of problem-solvers (as the problems left become increasingly more difficult and complex)
So yes, it is a lack of skills.
Just to be clear, I'm not convinced there isn't, either, I just think it's an interesting enough thesis that I'm willing to consider it (sorry for no cite, I have read interesting discussions of this point, may google around for them later).
I think the basic thesis boils down to this - there are, at any given moment, an "unlimited" number of scientific and technical tasks required to solve those deep problems you mentioned. However, there are (this is a thesis, not a conclusion) a limited number of those problems that can be worked on right now. Those problems need to be solved before the next set can be solved.
Now, if you have 10 research teams working on these problems, you have a better chance of solving them in a reasonable amount of time than if only 1 research team is working on them. 100 is better than 10. 1000 is better than 100... but at some point, you're maxed out on the number of research positions that can contribute effectively. Out of 1000, 100 will take a promising approach, 10 will get it right, and... 1 will solve the problem a bit faster than the other 9. There's a point at which you hit diminishing returns.
Is this true? (and btw, I'm constructing some of this thesis myself, as you can probably tell, I don't think I've read it in quite this form before). Is it true-ish?
I'm not really sure. I'm not sure it matters for the moment, because I don't think our budget for R&D and high value engineering work, or our educational systems to develop talent, are anywhere close to a point where we should be discussing whether we're reaching diminishing returns to scale.
But is there a theoretical point at which throwing more genius at problems won't make a difference in how quickly we progress? The question interests me, in an abstract way.
Consider the low interest among scientists in spending time and effort on reproduction of previous results.
Project proposal: gather a group of former paper-pushers with Excel experience and train them up in statistics and a scientific discipline (e.g. biochemistry, medicine, astronomy). After 2-3 years of training, take say 20% of the most promising ones and give them a grant to reproduce existing studies.
This fixes both the issue of idle workforce and of poor study reproducibility. It could also turn out to be a lot of fun for the affected. Who knows, you may even end up launching a successful scientific career or two.
Of course, this would cost money, but hey automation will increase profits which can be taxed and then redirected to these kind of projects.
God, I hope not. We already have done something akin to that to avoid job loss from prior technology advances; hence the increase in unnecessary bureaucratic and administrative jobs across government and private industry.
If the next evolution of the workforce is everyone is in management and the machines do all the work, then we're almost assuredly repeating the same mistake by giving people unfulfilling, pointless busy work (I've had managers who have viewed their responsibility as 'to bring leadership', and I've certainly seen plenty of middle managers whose existence was solely to take responsibility for failure and to pass bad news downwards; they had no decision making authority) just to generate a paycheck. Certainly, given the managers I've had, you're going to have a lot of bad ones, even if they're not pointless.
My take-away from it has been that the new "managerial" role will include much more responsibilities and more "hats" to wear. In other words, the manager will have to be the subject matter expert, the business analyst, the project manager and the administrator, all rolled into one.
So, in the case of banking, the "manager" will identify the need to say, automate the underwriting process for handing out institutional loans. They will design a list of tasks, create functional requirements, create constraints and the desired outcomes based on input values. They will also be responsible for providing everything necessary for the actual engineer team and their interaction with the QA team, as well as reporting to the executives.
In short - expect to see more overlap between senior IT roles and product/resource management, and more responsibilities rolled into single roles. People who can't adapt will be out of jobs.
>This mantra is being repeated over and over: algorithms and automation are taking away the jobs. I'm still skeptical.
My current work involves determining pain points in office "grunt work" and developing automation tools to reduce the need for temporary (and permanent) staff. Those people have been displaced from the workplace. No new jobs were created as a result of this automation. A great number of office workers will be displaced to due automation and the pace at which this will happen is going to leave a lot of people very unhappy and very unemployed.
Undoubtedly, I, or at least, people like me, will continue to have jobs. After all, someone must maintain the code, service the robots, et cetera. But that's a far cry from full employment of the workforce.
>And so I doubt we will run out of work to do. This would mean that somehow we ran out of problems to solve. On the contrary, we appear to be flooded with them: cancer, infectious diseases, global warming, clean energy, access to space...
How do you justify and fund higher education of all the out-of-work 40-, 50-, 60-year old truckers, office managers, office workers, and so on? Or will we just ignore them and wait for them to die?
There is definitely a short term horizon here and a long term one. The long term one is pretty wacky and probably no one alive right now has the mental capacity to understand or discuss it in any meaningful manner besides what is written in science fiction novels, and it probably is a post-human era.
You are right, there are still tons of problems to solve. Even those with a high standard of living continue to spend a large portion of time doing things they don't want to at home and work.
There is a dichotomy between a reduction of wages and all prices rising beyond affordability. Trillionaires turning the entire world in to giant ranches and using raw materials to build the biggest yachts permitted by the laws of physics seems like a stretch.
If we apply what happened to computing to other areas, and perhaps it isn't possible for everything but I imagine there are renewable/sustainable substitute goods for many things, prices should drop while power and quality increases by a lot.
Deflation is scary to the "old school" because all government budgets are based on borrowing money while the economy aka GDP keeps getting bigger and bigger. There is a lot of pressure all over the world in government policies to create inflation. Prices are all relative, but messing with functional markets at best accomplishes nothing and at worst causes an incredible amount of misery.
The other trend with job losses is that fertility rates in the developed world are dropping. The US continues trending down. Countries such as Japan are well in to their aging. There are going to be fewer people to do even the bare minimum required jobs to maintain existing infrastructure much less accomplish new things.
I used to want to hire a maid. Now I want a Roomba and a dishwasher.
I doubt that people who were a maid or a gardener will start solving cancer. Also, 'solving cancer' is currently getting more automated. IBM Watson is being used to analyze huge amounts of data that would've been pretty much impossible for people.
instead of thousands and thousands of companies with the same problems, the same struggle, and the same conclusions, there will be very few problem solvers and they will have large reach and influence
The same way the problem of famines is not the lack of food - Sen got a noble for pointing out that food was aplenty during the famine, but the poor didn't have the ability to purchase it.
Problems don't necessarily full employment make.
This has been going on ever since humans started thinking up better or different ways of doing things. Maybe we don't need as many money managers and brokers doing work that is done cheaper and better with technology. We will still need some, but not as many. There will be new areas where smart, driven people can make a great living.
Edit: One other thing to think about is how drastically automation could reduce the cost of living. In 1928 "A chicken in every pot and a car in every garage" was a campaign slogan. In 2016 where you can get a whole uncooked chicken for a few bucks in a grocery store it seems quaint that this resounded so heavily with the general public in 1928.
Sure! Now then, whose land will you be hunting/gathering on? You'll most likely have to pay for access to someone's land, and they probably won't barter access for chickens/goats.
It's basically the ice cream cone that licks itself.
You laugh, but... let's take the personal act of wishing someone happy birthday.
Before Facebook, you showed you remembered by calling.
Then facebook reminded you and you went to their profile wrote on their wall, where you saw tons of other short wishes like "all the best".
Then facebook made a little area where you don't even have to go to their wall, and you can just do all your "happy birthday" wishes in one go.
Then there are apps that let you automatically wish happy birthday to people.
A little further in the future:
Now people are complaining they are getting too many "personal", "heartfelt" happy birthday messages from people they barely know so they ask an algorithm to sift through them and summarize them.
Now we have bots reading happy birthday wishes from other bots.
A few years ago we built this at the TechCrunch Disrupt hackathon to prove this point:
To keep them from starvation, you would need to redistribute money to the ones who don't earn it themselves. Either via Unconditional Basic Income or Negative Tax. Because means tested welfare has the result of punishing people for earning money, so they aren't as willing to reinvent themselves and take productive jobs. Same with minimum wage laws. Eventually our society will be so rich that we'll be able to afford it, but how do we transition there gradually?
Computers will attract a larger share of the revenue every few years. They'll become "money sinks". So you tax the profits from laing off people, not enough to disincentivize R&D but enough so we don't have a severe demand shock for human labor and contraction of the money supply as we did in the Great Depression.
In general, though, it strikes me that we're only scratching the surface of what computer augmented humans can do, what we might call depth, while the major thing today's "algorithms" bring to the table is breath, such as what Google is (in)famous for.
It's up to you to get those things, or whatever things you crave. It's not like anything you originally owned has been taken away from you if you don't manage to do that.
That is a huge generalization, which seems intuitively wrong: why have norms towards conforming if they aren't effective?
I'm talking about the fact that the unemployed will face drastically lower qualities of life, or even life threatening situations. That this is acceptable comes from government policies, which in turn come from social norms. It's seen as okay to harm the unemployed, deny them healthcare, or let them starve. They're lazy, and they should be punished.
I won't speak for others, like you're doing, but conforming socially is the primary reason I work. I've been unemployed before, and I lost social status from it. It was embarrassing to admit to it. It was like losing a central part of my identity.
Social norms are important, but often you won't notice till you deviate from them.
How exactly do you "provide for yourself". You can't just go out and stake a claim to land anymore. The mechanisms for "providing for yourself" are purposefully limited to those sectioned and allowed by social norms. If you want to split hairs other "providing for yourself" please tell me of ways to do this outside the system.
If people didn't do their jobs there wouldn't be any food, homes or healthcare. People work not because of a social norm, but in order to solve each other's problems.
Note how ubiquitous work is: people engage in it across all cultures with otherwise highly diverse social norms.
The nature of work is really diverse across cultures. Hunter-gatherers hardly do any. Rural cultures have most people farming. I think about 1% of people in the USA are farmers.
Most jobs in the West are, well, bullshit jobs. Busy work. We do them despite the fact we could feed and house everyone with perhaps 9/10 people not working. It's merely a social expectation that we should all work, and a preference, not a necessity, to punish those who don't.
This seems like grasping entitlement. High productivity in farming or construction does not automatically give you the right to demand that they hand over their food or houses. Their preference to share their output in an exchange is not a punishment - it's reciprocity. If it wasn't for the prospect of exchange, why would anyone ever make more of anything than strictly what they needed? They could be hanging out with friends and family instead of catering to the needs of strangers.
> We do them despite the fact we could feed and house everyone with perhaps 9/10 people not working.
There is plenty of other problems to solve like cancer, diseases, global warming, clean energy, access to space and unfortunately in many places still housing and food. Claiming that people no longer need to work is like saying that we have run out of problems to solve. That this is it. That this is the world we want.
Here's a (mercifully short) article on the subject by Chris Hedges: http://www.truthdig.com/report/item/karl_marx_was_right_2015...
It would be interesting to see this proposal work out. Capital is movable whereas people are mostly not. How would you prevent the robots moving (or built and run) where there are no taxes on them? I don't see how this be preventable without all nations agreeing on it. It doesn't work even with tax havens within the EU so why should this work world wide?
Better yet, all the people without access to automation could in form an entire economy among themselves. They would live like we do now, or like we have done at some point in history. Even without any support from the more developed society, they could develop like we do now.
In reality saving is the engine of the economy. Consumer debt only temporarily increases the GDP. Especially in the form of payday loans it is not only bad for the individual but terrible for the rest of the economy.
I know right. Why don't poor people just get more money?
- Wow that looks cool. - Heh. Look! It sorta works. - Oh. It only sort of works. - Um, we can ship this but... - Next!