Brad Wardell's site for talking about the customization of Windows.

imageI don’t like sounding like chicken little. Nevertheless, the sky is falling.

For those of you who don’t know me, I run one of the oldest independent software companies in the world. My specialty for the past 20 years has been in AI. It is my job to research, evaluate trends and invest accordingly to stay ahead of the curve. And I am here to tell you that the sky is falling.

It is automation. It is inevitable. It is irresistible. And if you think that automation always creates new opportunities look back to horses. Technology made them more useful too…for a time. Now it’s our turn and in this Facebook post I’m going to walk you through, in plain terms, why I think the sky is falling.

I’m not going to try to persuade you. I’m just going to put out the data. I suspect anyone reading this is intelligent enough to reach their own conclusions whether they agree with my assessment or not.

Before I begin, I want you to refer to this page: http://www.bls.gov/emp/ep_table_201.htm This is the Bureau of Labor statistics. It’s the government’s site that tracks what people are employed in.

To summarize we Americans work primarily in these areas:

1. Factory work (8%)

2. Construction (4%)

3. Retail (10%)

4. Transportation (3%)

5. Business Services (13%)

6. Healthcare (12%)

Next, I would like to quote you this statistic: “The maximum unemployment rate during the Great Depression was 25%.”

The sum of the percentages I give above is 50%. You be the judge on what percentage of the areas I am going to discuss below will likely be out of a job in the near future.

Like I said, my day job is to evaluate technology and try to predict where it’s going to go next. And with that, I am telling you the automation revolution isn’t happening soon. It’s happening right now.

Amazon Prime

Do you use Amazon Prime? It’s pretty great right? How are they able to do it so cheaply? It’s because it’s largely automated now. Over the past few years, Amazon has been quietly laying off thousands of employees and replacing them with machines.

http://www.techrepublic.com/article/amazon-robots-and-the-near-future-rise-of-the-automated-warehouse/

Amazon currently employs over 200,000 people, most of whom will be out of a job in less than five years. Right now, you go to their website, you order something. That signals a AGV to go over and pick it up in the warehouse which then takes it over to another AGV (automated guided vehicle) that in turn takes it over to the auto packager which in turn sends it to be sorted and packaged.

I’m not talking about some future technology either. Did you get something from Amazon Prime recently? Look at the box you received. You will see an MSI code and a Code-128 code (very similar) along with a Datamatrix code (a box with graphical blotches). Right now, some of this is still handled by a person. But this will soon be completely automated.

Right now, the Code 128 code is used by UPS or Fed Ex staff (people) to load trucks and get them to you. But this is not going to last much longer. The transportation industry is already in process of being automated. You don’t hear much on this publicly because no one wants to talk about it. But this isn’t a 10 or 20 year away thing. This transformation is happening right now.

Products and goods will soon be transported to you through autonomous vehicles. I’m not talking about drones. That’ll happen too but that’s a distraction. I mean that UPS trucks and Fed Ex trucks will soon be autonomous. Loaded at the warehouse by machines and transported to you by AFVs (autonomous freight vehicles).

And even if you think “how will they get up to my doorstep” remember most shipments occur from business to business who have their own loading docks and warehouses. Moving stuff from point A to point B is a huge part of our employment.

Going…going…gone

Retail, where a lot of people now work, is going to be hit soonest, hardest and most obviously. We are familiar with self-checkouts but that’s really not that big of a deal. It’s the stocking that is going to go away and you won’t even notice. Walmart, Target, you name it, will quietly and not-so-gradually replace their stock people with machines. Read a bar code, go to the proper location in the store and place it. 24 hours a day, 7 days a week.

Chain restaurants? The only reason why McDonalds and Burger King haven’t automated already is the relatively low minimum wage. But that’s going away. Kiosks will replace the order taking and the food preparation will be handled by machines. And machines in 2016, already do voice recognition better than most humans for the drive through (that’s something I never would have thought possible even 5 years ago). And people will be happy for this because it’ll be more convenient and the results more consistent. They’ll never “fuck you at the drive through” because the order will be perfect every time.

What about office jobs? They’re safe right? No. Again, I want to emphasize that I am not talking about some “20 years from now” thing. I am telling you that this automation revolution is happening as you read this. It’s not something to prepare for in the future. It’s already upon us. And with that in mind, Walmart just announced that it is cutting 7,000 office jobs.

http://fortune.com/2016/09/01/walmart-job-cuts-layoffs/

This was a week ago.

These are administrative jobs. Accounting jobs. All those jobs that involve paperwork, inventory management, producing invoices, handling payroll. Do these jobs sound familiar? They’re not going away in the future. They’re going away right now. And it’s accelerating.

Today, you walk into a Walmart and pick up a can of soup off the shelf. That soup was placed there by a person. You probably still go to a person to check it out because you have a bunch of stuff and it’s still a pain to do self-checkout. Nevertheless, everything you bought is automatically deducted from Walmart’s inventory. The acquisition of that item and its purchase doesn’t require people anymore so it was automated.

But relatively soon, every item, from food to your clothes, will have a tiny passive RFID tag in it. You’ll simply walk through a checkout and everything will be deducted automatically (for those with a NFC device like an iPhone or an Android phone). There will still be a person handling stuff for old people. But most people will naturally prefer to take their cart full of stuff through the RFID scanner and have it handled automatically.

Ironically, the service is likely to get better because they’ll probably soon have auto-baggers so you won’t even have to bag your own stuff anymore. But that’s probably around 7 (2023) years away from becoming mainstream. You’ll have RFID tagging sooner than that.

If you’re feeling stressed and want to go to the doctor, well, they’re going to be automated away too soon. And this will be a good thing for everyone. Today’s doctors will become more focused on dealing with patients’ needs while the machines handle the diagnosing and prescription writing.

You, reading this right now, when was the last time you went to the doctor not knowing what you already had? You probably just needed the prescription and had to wait. The machines, networked with each other across the world and able to sample billions of people’s anonymous data will make Dr. House look like an amateur and prescribe you with what you need vastly faster than having to wait for the doctor.

However, this won’t be good news for a lot of people in the health industry. Your doctor today with the thriving practice will be fine. He or she will save up and buy these diagnostic machines that will handle the vast majority of cases he or she currently handles. But those next generation doctors? They’re in for a rough time. Those in the medical profession can comment below and explain the problem a lot better than I can.

What about lawyers? They’re screwed. As someone who routinely gets sued (intellectual property is a mine field), I have a lot of experience with lawyers. The most expensive part, by far, is discovery. This is the part where each party sifts through the other’s sides stuff to determine what bullshit to put into their motions to convince a jury that their narrative of the case is the correct one. 99% of that time is wasted. Machines could handle that 99%.

It is unlikely that there will be such a thing as a paralegal in 20 years. They’ll go the way of the gas station attendant.

http://technoccult.net/archives/2013/10/08/report-47-of-u-s-jobs-at-risk-of-being-automated-out-of-existence/

Now, I’m not suggesting all these jobs are going to be gone in 5 years or even 20 years. Not all of them. But a lot of them. And unlike in the past, there’s no job for these people to go to. There’s no “training” for a new job because this time, the machines aren’t creating a new type of job in their wake, they are simply replacing the existing jobs without creating a new one.

We are not ready

We are not ready for this. We are oblivious now and we will remain oblivious until it’s far too late for our society to adapt carefully.

People will continue to be oblivious even as they watch their malls close down just like people shrugged when their bookstores went away.

http://time.com/money/4327632/shopping-malls-closing/

They’ll continue to be oblivious when their neighbor’s kid loses their job at the coffee shop because there’s a machine that makes the perfect Mocha Latte every single time.

They’ll still ignore it even as their sister’s husband loses his job at DHL (the world’s largest logistic company).

http://roboticsandautomationnews.com/2016/04/20/dhl-opens-supersize-logistics-centre-featuring-130-robotic-shuttles/4068/

They’ll only notice when their job at the local dentist’s office handling appointments and other office duties suddenly, without notice, disappears because Dr. Benning, such a great guy, has bought a Wavenet Office bot that can call patients to make appointments, reschedule, and handle all the tasks they previously did.

https://deepmind.com/blog/wavenet-generative-model-raw-audio/

And what about the unemployed? Think that they’ll just raise up a mob and take on the 5% of the population who owns the means of production now? Think again. The one area automation is already doing very well is security.

The new uber class (calling them rich isn’t even the right term at that stage) will be way ahead of the mob. They’ll have machines to protect their holdings, homes and families from the rabble. So those who think “they’ll have to do something for everyone else or they’ll be a revolution” think again. Don’t assume a universal basic income is a definite.

/">http://www.digitaltrends.com/cool-tech/security-robot-knightscope-/

What we do as a civilization, will define my generation. I pray we figure it out. Nothing scares me more than a future of split between the Gods and the Useless.

Further reading:

Economic Singularity


Comments (Page 3)
on Sep 15, 2016

zuPloed

 

Quoting Kazriko,

Note the larger lump towards the top of that graph, from the 50 to 67 range. [...] Just eye-balling it, that looks like about 30% of the workforce, which is in the 25-50% range you were highlighting

Good.

Now eye-ball the 3-20 region and subtract it from the retiring 50-67 region. Will the net change still be 30%? Will it even be over 10%? I would estimat this change to be about 5%. Don't forget, if we are looking at the work force, natural growth will have a delay of about 20 years: Even if everyone stopped having children right now, it wouldn't change anything for the workforce in 20 years.

Yes, I noticed this after posting. I was mainly looking at the fact that towards the bottom, it keeps getting thinner, without noting that it was still thicker than the thin part of the 49-67 segment. It's much harder to calculate having solely a graph though.  

zuPloed

Quoting Kazriko,

People who planned ahead though would be better off.

I am not familiar with US retirement plans, so all I can do is drop this, I guess. I am sure everyone can afford retiring 5-10 years early. Especially these guys who worked in amazon packaging.

the US retirement consists of two things that are still viable. There used to be a 3rd, Defined Benefits, but it's basically reserved for government employees and those who are part of massive unions at this point. Social Security, and Defined Contribution. Social security was originally designed to cover people who were long past the average age of mortality, primarily to take care of widows and those who were so extremely old that they simply couldn't work at all. It's a defined benefits package that gives just enough to where they can live modestly in an area with inexpensive costs. The only requirement is a certain, relatively small amount of work through one's life from which social security taxes were taken. This can be gotten early, but at a greatly reduced rate. If someone retires early it's usually best for them to just live off their other savings until they are 67.

The second is Defined Contribution. Generally, you negotiate this with your employer. You put in a certain amount, and your company puts in a percentage of this amount. It is stored in a mix of various financial instruments tax-free (for the standard 401k) that starts much more aggressive, then is transitioned to safer assets as you get closer to retirement. How much you can afford and how early you can retire is based entirely on how much money you accumulate in this account. 

Though, I wasn't referring to early retirement when I was posting, just standard retirement at 67. 

That link is right though. A lot of people, myself included, are terrible with money. This is why the companies cheerlead people towards contributing to the retirement plan several times per year.

 

Following the post yesterday, I started thinking about the original post and the 25-50% claim in general. 

It used two main things as examples. Amazon and Walmart. Two companies that are extremely aggressive about cost cutting. They don't follow the norm of US businesses when it comes to laying people off.

My own experience in the area of Oil and gas field automation indicates a much different trajectory, one that is probably more typical for the bulk of the workforce. When you have absolutely no automation on a well, you need someone sitting there 24/7 to monitor it. Given shifts and days off, you probably need 4-5 people total whose full time job it is to maintain this well. When you add the bare minimum of safety features to a well, such as shutting in when the tank fills in, or closing if the temperatures get too far out of range, or shutting when the flow drops below a minimum, then you can theoretically fire 2-3 of those people, but in reality this does not happen. Instead, they drill 2 more wells, put the same safety features on them, then have the workers rotate between the wells, doing the tasks of maintaining production on them. The more and more automation that gets added, the more wells get drilled and the more of the field that a pumper maintains. Each time moving up to focusing on just the top level tasks, replacing worn parts, tweaking the timings, etc. What you mainly get is that the automation is allowing them to expand more quickly than normal, and is slightly reducing the number of new people hired.

Because these people are doing things that are worth more money, their salary goes up as well. The wages and the capital costs rise in tandem as the field grows. 

Where you do end up firing 2-3 people is when suddenly the prices plummet and you're no longer making as much money as before. The workers are cut, the remaining workers have to work longer hours maintaining the wells that the cut workers used to maintain. But they're not really spending anything on new automation or any other capital expenditures either because they don't have the money. They reduce new well production to just enough to keep their rate steady, if any at all. If the remaining people are overworked, they might bring in a cheaper, younger person to handle some of the work. They may shutter some very poorly performing wells and reclaim them instead. If the prices start rising again, they'll evaluate how overworked people are, how many opportunities there are for expansion, and hire more people based on that and if they can afford it.

So, I think that the rate of decrease in jobs isn't going to be exactly the same as the rate that automation is added, but rather a cyclical effect. Automation added during the good times, but no job losses until the cyclical bad times. This would probably slow down the 25-50% losses described by a few more decades.

on Sep 15, 2016

Ok, so automation increases productivity per employee.

Which means you can use your experienced employees to expand.

So far so good.

But what I can't wrap my head around is: Why is expansion allways favorable? To take the example of super markets, at some point you have good coverage, so everyone has one nearby. How do you profit from expanding then?

Amazon could have more distribution centers and push delivery time down to a day or less everywhere, but will people buy more then? Pay more?

In that scenario I only see two ways: fire people to optimize profit or keep them due to basic human decency.

on Sep 15, 2016

Basically more people mean bigger demand. Bigger demand means we need a better way to make goods. Also due to advertising people need what they don't need. That basically answer's your question. Why everything is getting bigger. Now you communicate in a smaller world where you don't know where people are from. Now this can be supplied by small businesses helped out by stocks, so a rich man doesn't need to invest into your company to make it work. Except a big business has more resources.

on Sep 16, 2016

zuPloed

Ok, so automation increases productivity per employee.

Which means you can use your experienced employees to expand.

So far so good.

But what I can't wrap my head around is: Why is expansion allways favorable? To take the example of super markets, at some point you have good coverage, so everyone has one nearby. How do you profit from expanding then?

Amazon could have more distribution centers and push delivery time down to a day or less everywhere, but will people buy more then? Pay more?

In that scenario I only see two ways: fire people to optimize profit or keep them due to basic human decency.

Well, they're not necessarily going to realize that they have more people than they need. They put the automation in to make it easier on their employees and to smooth out problems they encounter, but don't necessarily know how much idle time those employees have. They may not go looking for that until profits drop and they need to cut costs. 

Aside from that though, the old adage in business is that if the business isn't growing, it's dying...

on Sep 17, 2016

There's this book I read recently, "The Artificial Intelligence Conspiracy", by one Arthur Shakarian. The subtitle is "How the World's Elites Plan to Replace Everybody Else with Intelligent Machines". (found here: https://www.amazon.com/Artificial-Intelligence-Conspiracy-EverybodyElse-Intelligent-ebook/dp/B01AHZU5ES/ )

I used to think it was bull, what with some of the richest people in the world actively getting together and planning the last push for automation, so they could get rid of the masses (i.e., us) for good this time. But with everything I've seen over the past few months, this article included, it seems like the author might not be that far off the mark.

The most frightening thing in the book, to my mind, was the idea that the elites won't even have to suppress us with force or kill us all once they no longer need us (because their cars drive themselves, their kitchens cook on their own, and their laundry gets folded by little laundry robots). On the contrary, we'll go peacefully, even blissfully, thanks to (wait for it) interactive Virtual Reality porn.

There's a whole chapter about this, and it's positively horrifying. (would quote, but the book's on my other machine.) IIRC the argument goes something like: Look at porn today, how bad it is and 2D and non-interactive, and people get addicted to it anyway. Now look at the very early stages of VR we're in, still choppy and expensive and all, but we're already seeing bad cases of VR addiction, disorientation and dissatisfaction with the real world. Now put the two together and wait a few years (or months, at the pace of today's progress), and you get mature interactive VR porn that will be the drug to end all drugs.

Even if we don't end up like these autodarwinators who forget to eat and drink because they're so hooked on computer games, actual real-world sex will be so clearly inferior to VR sex with AIs that learn exactly how to please and tease and edge us, that we'll die out within a few generations max. And this goes for women as well as men. Who wants to procreate if you can have perfect sex tuned to your exact preferences, no matter how unrealistic or specific, forever?

I can actually see that happening. And the most frightful thing? I'm not even entirely sure I would mind.

 

on Sep 17, 2016

The same people that get "addicted" to porn today, are the same people that would waste away on vr porn...

 

It's not much different from alcoholism, only a small percentage of the population actually flushes their lives down the drain over the stuff, the rest of us have more self respect.

on Sep 17, 2016

To say that this was a great article would take understatement to a new level.

I've written on this subject as well, here and elsewhere, as in my article on a solution to poverty, as well as my On Morals piece.  I'm effectively blacklisted by many libertarians for the solution piece. 

Thanks


 

on Sep 19, 2016

Kazriko
Aside from that though, the old adage in business is that if the business isn't growing, it's dying...

In the long run, sometimes conventional wisdom turns out to be very wrong. For example, this old saw could also roughly describe the 'success' of a cancer.

One of the things I love about the GalCiv series is that it has given me a long run of playtime that also feeds into my past training and current hobby-thinking as a political theorist. Among other things, I've had plenty of opportunities to question assumptions, both others' and my own.

Despite having tried some of the more common aggression-centered approaches to GC, I'm still strongly inclined to start a game as what I've come to call an 'optimizer' (that's as opposed to a maximizer, the approach taken by the numbers masters who quickly figure out how to thrash the AIs at max difficulty levels).

Parts of this engaging, disturbing thread have set me wondering again if our era might be able to devise a middle way between extreme capitalism and command economies that is different from traditional social democracies in a way that can balance desires for broad liberties with the need to resolve crises of overproduction and maldistribution.

As a lapsed anarchist, I remain fairly convinced that a solution here cannot be achieved through top-down power. But we surely do need more thoughtful discussion of the situation from all 'sides.' Including speech from ardent capitalists like Brad Wardell, who, despite being able to inflame many of my 'fellow travelers' to spew mindless vitriol by the barrel, does seem to have a strong conscience working alongside his entrepreneurial streak.

p.s. That bastion of sold-out-to-big-money Democrats, NYT, has an article that seems worth including amongst the references in this thread http://www.nytimes.com/2016/09/19/technology/artificial-intelligence-software-is-booming-but-why-now.html

on Sep 19, 2016

I don't believe he was motivated by greed, but by a disatisfaction on how games were going. Realizing he could make games he like if he could fund it started a software company to fund his game making.

on Sep 20, 2016

Philocthetes


Quoting Kazriko,
Aside from that though, the old adage in business is that if the business isn't growing, it's dying...



In the long run, sometimes conventional wisdom turns out to be very wrong. For example, this old saw could also roughly describe the 'success' of a cancer.

 

I can say to this though, that the one bit of wisdom to the saw is that existing business tends to decay over time. Each game in a series that is just "holding steady" will get a few fewer purchasers than the last. Each oil well will slowly decrease its production over time. Every magazine if it doesn't advertise for new subscribers will slowly shed them over time. You need to be making at least some effort to increase your business at all times even if all you want to do is just to hold steady, otherwise it will decay into nothing. 

As for finding real world wisdom in games, It's possible in some limited instances to do so, but you should be aware that even the most realistic simulation will not entirely represent a realistic view of the world. They will often more represent the biases of the one producing the simulation.

 

on Sep 20, 2016

Kazriko
As for finding real world wisdom in games, It's possible in some limited instances to do so, but you should be aware that even the most realistic simulation will not entirely represent a realistic view of the world. They will often more represent the biases of the one producing the simulation.

I didn't mean to imply that I was divining 'truths' in the GC series, just that playing this sort of game can be a sort of 'exercise' that can feed into deliberate thinking about political economy.

A couple of times over the years I've poked at the devs about taking away a major 'bias' of the series by doing something analogous to the US going off the gold standard. I really want to see the BC economy 'demoted' to just one way of accruing, transferring, and using excess resources. I still find it very, very odd that a civilization like the Yor have any use for a currency. Of course I understand that my 'little' fantasy there is the opposite of little when it comes to the code planning and building.

One day, maybe around GC 6 or 9, I'd really love to play a game where an alliance led by market-fueled Iridium and Terra might go up against one led by post-currency Altarians and Iconians, with the merry-mad Snathi Revenge poaching both currency and the econ-strategic resources (maybe energy, basic macro-molecules, and industrial metals).

on Sep 20, 2016

I agree the Yor wouldn't use money. I think you could substitute production for lack of economics or government. Intimidation for diplomacy. Using this replacement we could have a more version of two's tree. Of course the Yor couldn't lease or rush production, but wouldn't pay for maintenance, or be able to trade with money.In three the Thalan got an extra power plant to make up for a population penalty. The Yor just got screwed on that with a population penalty, and no feasable power plant to be used like a power plant. No manufacturing or research capital.

on Sep 21, 2016

<<What about lawyers? They’re screwed. As someone who routinely gets sued (intellectual property is a mine field), I have a lot of experience with lawyers. The most expensive part, by far, is discovery. This is the part where each party sifts through the other’s sides stuff to determine what bullshit to put into their motions to convince a jury that their narrative of the case is the correct one. 99% of that time is wasted. Machines could handle that 99%.

It is unlikely that there will be such a thing as a paralegal in 20 years. They’ll go the way of the gas station attendant.>>

 

Unlikely.  For one thing, lawyers make the law and it would be easy to prohibit the practice of law by a non-human.  LOL.  Second, lawyers need to speak and most robots can't speak. No a trained parrot like Siri does not qualify.   Third, computers are widely used by attorneys as a tool in the discovery process, there are no machines or software that can do what lawyers do.

More broadly, robots are good at very limited tasks and there are no signs that autonomous cars are likely to hit the streets in any great numbers any time soon.  We can thank lawyers for that because every time there is an accident, there will be a lawsuit alleging defective software caused the accident. 

I'm still waiting for flying cars from the 1950's shows or even flying hoverboards (yeah I know but the ones that are supposed to be coming out are not the right ones).

 

 

on Sep 21, 2016

Kazriko

It's not going to be a blood in the streets, lets kill all the useless people or let them starve sort of thing.

But is what is now happening in the big cities with the black lives matter not a prelude on this? A lot of people do not have work and have little hope for a better life which makes this problem only bigger.

(I could be wrong, but it looks that way from my side of the atlantic)

on Sep 21, 2016

Gabberkooij


Quoting Kazriko,

It's not going to be a blood in the streets, lets kill all the useless people or let them starve sort of thing.



But is what is now happening in the big cities with the black lives matter not a prelude on this? A lot of people do not have work and have little hope for a better life which makes this problem only bigger.

(I could be wrong, but it looks that way from my side of the atlantic)

It's actually still quite rare. The rare issues are made much bigger because of the media. That's not to say there aren't problems, but they're not as wide spread as the media would like you to believe. Most of the issues descend primarily from the ridiculous drug war and the way that, just like Prohibition before it, it's encouraged criminal gangs and mafias. That leads to all kinds of police racial profiling of those who resemble the gang members and training problems surrounding their reactions to said people, and some false-positives that get wall to wall news coverage. When false positives mean dead innocent citizens though, they should probably do something to err on the other side, but for them, they believe that false negatives mean dead police officers. 

Unfortunately, records being kept on this aren't as expansive or easily collected as one might want, so it's difficult to pin down exact numbers of how many of these are false positives and excessive use of force, but we're hearing about maybe 10-20 per year out of the 1000 or so people shot by police, but that false positive number could easily be higher. Police officers shot has been hovering around the 50's per year, way down from the highs of the 70's or the high's of the alcohol prohibition era. (also half that of a couple decades ago.) A bit higher this year because of a few nutballs. There's been a number of documented cases where the incentives are all backwards for this sort of thing. Police officers being fired for "failing to eliminate a threat" and such.

If you were a conspiracy nut, you might conclude that drug prohibition was specifically designed so they could do this, but I don't believe that line of reasoning.

I'm still an advocate for ending the prohibition, and working towards robust systems for mitigating drug use instead of criminalizing it.

Meta
Views
» 107132
Comments
» 63
Sponsored Links