The Robots are coming, and if they replace everything that they're capable of replacing over the next 15-20 years, we're going to hit 65% unemployment.
That isn't an exaggeration.
I've been working on a breakdown, but as per usual, my post was far too complicated and dealt with far too many issues for me to be able to break it down simply into a coherent, concise post.
CGP Grey, the youtuber, was able to crystallize all of these issues into a perfect, 15-minute discussion of the new economy developing before our eyes.
The following is a summary of CGP Grey's video. All of these points are his ideas, and the summary works better than a transcript here. The automation of the 70-90s is archaic. General Purpose robots are now coming. They're here. We have just developed the first general-purpose robot, Baxter. Baxter is capable of doing just about any physical job a human can, but much slower. If I taught Baxter how to make my tomato sauce, Baxter could sit in my kitchen and make tomato sauce all day and night. For months, until he inevitably needed some kind of repair.
As long as I kept making sure he had the raw materials, he'd never need to stop making tomato sauce. I could then take said tomato sauce, freeze it, and sell it at farmers markets for a huge markup.
Purchasing Baxter costs less than a part-time employee's yearly pay at minimum wage.
The people who invented Baxter have the economic goal of achieving near-universal general purpose robot ownership within the next 5-10 years.
Grey draws a comparison between General Purpose bots and the General Purpose computers we're typing on.
The moment that Cheap-Ish general purpose computers appeared, it took about a decade for nearly every American home to acquire one. Again, paraphrasing Grey, baxter is not the apex, baxter is the Xerox Alto. He's the first of what's coming.
Grey uses the example of the self-check out at a grocery store. Where you once had 30 people, you now have one person supervising 30 machines. Even though people steal all the time, by mis-labeling goods (quinoa is more expensive than millet), or by labeling one expensive item as "onions" and then purchasing it by weight, or by weighing their avocados with their saffron, and putting them both in the bag at once, it is still less expensive for a company to eat those losses than it is to pay 30 cashiers.
And that's just idiot bots.
No one believes the phrase "better technology makes better jobs for horses." That's absurd. When the automobile and the tank appeared, the main two jobs of Horses, warfare and transport, were no longer necessary. We replaced the horse with air cavalry, and armored cavalry. Helicopters and Tanks. The horse population has since massively declined, because we don't actually need as many horses anymore.
When you replace the word "Horses" though and create the phrase "better technology makes better jobs for humans" we all suddenly believe it.
This is absolutely false.
"As mechanical muscles pushed horses out of the labor market, mechanical minds will do the same to humans. Not immediately, not everywhere, but in large enough numbers and soon enough that it's going to be a huge problem if we're not prepared."
"And we're not prepared."
You may think that you couldn't possibly be replaced by a robot, "but technology is getting better stronger and faster at a rate that technology can't match."
Grey points out that as the car showed the end of the horse, so now does the self driving car show us "the shape of things to come."
"Self driving cars aren't the future. They're here, and they work."
And they work better than human drivers.
Globally, that's 70 million jobs, gone. Those jobs are over. They are never coming back. From pilots, to train engineers, to truck drivers, humans are no longer needed in these professions.
And as Grey points out, the unions aren't going to prevent it. In fights between workers and technology designed to replace those workers, guess which side has never once been victorious.
The workers. They never, ever win these fights. Never once. And there are too many forces arrayed against them for them to ever win such fights.
Smart bots are coming. If you read a newspaper, and read an article with no listed author, you've probably read an article written by a robot, and then minimally edited by a human.
That's because we're teaching robots how to learn. No programmer can teach a bot to write. But they can teach a bot to learn to write. And the bot can trial-and-error their way to success much faster than a human can.
And this can be applied to all sectors of the economy. The age of the robotic CEO is coming. If a robotic mind can manage human wealth better than a feebleminded human can, then the board is going to fire the CEO and replace him with an omniscient AI that never sleeps, even if the upgradable AI in question is a supercomputer that costs $50 million a pop.
That's still cheaper, over five years, than paying a CEO 12 million a year.
The Kochs are likely to fire management teams, and replace them with AI bots, if it's in their economic interest to do so.
The value there isn't even going to be the hardware. The hardware can be upgraded, and it's likely that a CEO-bot would be able to manage it's own hardware upgrades.
The value is in the robot's learned, on-the-job experience, which it can share with other bots networked to it. So the board doesn't just own one CEO bot. They own a CEO-bot that can replace the CEO's of every company that the boardmembers own, saving themselves millions of dollars in bonuses and stock options every single year. Because the bot only costs the electricity it takes to run the bot.
And if you think that bots can't replace CEOs, look at wall street.
The floor of wall street is a TV set. The actual trading work is already done by Robots that taught themselves how to trade. And those robots are trading with other robots who taught themselves how to trade.
Nobody but TV personalities work the pit anymore. And that's just today. This technology is getting faster and better all the time.
This ends my minimal summary of CGP Grey's work. It's mostly him, with a few of my responses, but everything between the video and this line of text is either Grey's ideas or extrapolated from Grey's ideas. I give all credit in the preceding section to CGP grey.
The following is my work. I'm going to talk about what this means for police, military and government work, and the steps we need to take as humans to get ready for the future.
Robotic Policing
This is both potentially a good thing and the greatest threat to human liberty imaginable.
The good: Robots will have skill-sets created by programmers and politicians working together.
That means that robot policemen can be equipped with first aid kits, medical diagnostic equipment, and emergency lifesaving medicines. From epipens to adrenaline to naloxone, a police bot will be able to respond to a medical emergency and immediately save the life of someone having a heart attack, an overdose, or an allergic reaction.
A policebot doesn't care about its own life, because it isn't alive. Self-preservation is irrelevant. Robotic police won't gun down someone holding a wallet. They'll be able to see that it's a wallet. Also, they're bulletproof, so unless you have a rocket launcher, the policebot isn't concerned about your peashooter until you aim it at someone else. The policebot's goal is not to return to its wife and family: it has none. Instead, it's goal is to protect human life, and if it gets riddled with bullets, it doesn't die, it gets repaired. The policebot will not generally need firearms, because its goal is never to kill, but to detain a suspect, or print a citation.
Robotic police won't have "culture."
A robot cannot be part of a culture of violence because it has no culture. It's behavior can be modified easily, because it needs no union. Changing police behavior as a whole is as easy as changing a few lines of code.
Robots are incapable of malice or racism. They are not capable of feeling malice because they are incapable of feeling. They have no malicious desire to hassle the same individual day after day. If someone is selling cigarettes without a license, then the robot will approach them, issue them a citation, print out a ticket, and move on. It will not attempt to arrest the individual. It will not choke the person to death. It will simply print a ticket, and move on. Any racism inherent in a police bot will have to be specifically programmed into it.
Because robots don't fear punches, handcuffs become unnecessary.
A bot arresting a violent individual without a gun can simply grab their arm, and wait as the individual flails and beats the robot. The robot can then wait. For hours if necessary. Detaining a suspect without escalating violence, or needing to inflict any kind of pain.
There are no batons. No tasers. No pepper spray. No weapon of any kind is used unless the person is using a weapon on another person. You can hit an armored robot with a baseball bat all you like, the robot's not going to suffer anything but cosmetic damage.
In a situation where someone is being arrested, "You are under arrest. Come with me." Says the robot. And the person can either choose to cooperate and just follow the police bot, or the police bot can grab the person's wrist, and wait for them to become cooperative.
So there's no need for any violent arrest tactics. The robot arresting a person simply waits until the person is willing to cooperate with the arrest.
Robots can be hard-wired not to violate the legal rights of citizens.
A robot during a protest situation will not pepper spray innocent citizens lawfully exercising their constitutional rights to peaceably assemble. It will not punch protesters, or kick protesters in an attempt to cause a riot because it's angry about being on protester duty.
A policebot can be legally bound to refuse orders that violate a citizens rights. That means that if the police chief orders a park cleared of peaceful protesters, the robots are hard-wired to refuse that illegal order, because the people in the park are exercising their constitutional rights.
A robot during a riot situation can be screamed at, punched, kicked, hit with sticks and bottles, and not respond by brutalizing members of the crowd. You can punch a robot all day, and it will not feel angry. You can scream obscenities at it, and it will not care.
This is what can be, but it's probably not what will be. And that's because-
The bad: Robots will have skill-sets created by idiot programmers and evil politicians working together.
Robots may be incapable of malice or racism, but the people controlling them sure as hell are, and that's where we run into almost all of our problems with Robotic police.
Yes, we can argue that police-bots need to have hard-wired rules about not violating a citizen's rights, not killing, not even having guns, but some idiot, somewhere, is going to object to all of this. The police commissioner is going to want to have the power to clear out some pesky protesters, not have all of his robots tell him they're incapable of following his illegal orders. So there will be a huge resistance, by both the Patrolman's Associations and the police commanders, to policebots. From the Patrolman's Associations because they're losing their jobs, and from the Commanders because they're losing a private army and gaining a police force specifically programmed to prevent the kind of brutality that humans are capable of.
And we already have enough idiot tough on crime politicians out there, and one of these idiots is going to undermine a new paradigm of un-armed police bots by saying that if some violent criminal is shooting at police bots, then the police bots should shoot back. Then add in official police policies like stop and frisk, and you've got a system that is just as racist, and just as violent as the current one, but where there's no one to send to jail, because the police can just plead "Malfunction" rather than intentional malice.
So then our police bots are no longer nonviolent police bots, designed to protect human life, designed never to kill, hard wired never to violate the rights and liberties of citizens, but instead become low-level combat bots. Designed to engage in ruthlessly efficient military operations against civilians when their programming decides it's murder-all-meatbags time. Or murder-all-meatbags whose skin is darker than a paper bag time. That can be programmed in, too.
And that's going to be a problem if for "security reasons" the source code of police bots is classified. And this gets even worse if we're not just dealing with accidental murderbots, but actual intentional murderbots.
Robotic Military
Robo Military systems have exactly the same problems as robotic policemen.
To call Drones robots is to misunderstand what a robot is. A robot is not a remote-controlled kite with missiles on it. A robot is capable of independent thought and action, and our drones are not.
Yet.
The good behind robotic infantry is that it will not massacre civilians unless directly and intentionally ordered to do so. An infantrybot's goal is to obey orders. We can hard wire it to disobey unlawful orders, but that will be resisted by generals who won't want a military force that regularly tells them "I'm sorry, General, I can't do that."
The robotic infantry system's capabilities will far outstrip that of a human soldier. It will be able to detect where enemy fire is coming from based on sound, and instantly return fire with pinpoint accuracy. It will be able to kill an enemy soldier shooting at it before the enemy soldier has taken a second shot.
It will be able to see in complete darkness. It may be able to move silently. From snipers to heavy infantry, robots would be able to identify targets, separate combatants from civilians, and operate like a terrifying scalpel. Assuming everything goes well, and that the robot is used like it's supposed to be used.
If a robotic infantry battalion is ordered to exterminate all human life in a certain sector, then it will carry out its mission with ruthless efficiency. That action can then be marked up to a "Programming Error" and you can bet good money that the code for war bots will absolutely be classified. So the people who ordered the genocidal attack in question will be able to avoid any responsibility for the atrocity they committed, assuming that permanent, inescapable safeguards aren't instituted.
Some science fiction authors have suggested the creation of permanent, durable, Read-Only-Memory systems where battle recordings are created for legal, historical, and military intelligence purposes, which are impossible or nearly impossible to modify. But we'll get to the solutions later.
And neither of these problem sections deal with the question of what happens when something really and truly does malfunction.
Bureaucracy to Robocracy
Bureaucratic Robotics, that's the next step. Going to the DMV is going to be faster than ever before, especially since humans probably won't be doing much driving themselves, but the moment you need something human out of a robot, the moment you fail to be a frictionless sphere moving along society's trackways, that is the moment that the robotic bureaucracy can't help you. There's no form for what you need. No option on the menu. The virtual intelligence you're dealing with will apologize to you, repeatedly, but be unable to assist you further. There's no supervisor to call. There's no ladder to climb.
It's designed that way, just like a corporate call center. They are both the final arbiter and helpless.
Let's think about this for a minute. You've got a special case of some sort. There's a loophole in the law, something that hasn't been considered. The robot has no experience dealing with this, and as far as the city management is concerned, you'll never get to talk to anyone who isn't a robot. So what do you do?
Well, if this is still the US and not some kind of dystopian nightmare that we're writing here, you call your city councilor, your neighborhood ombudsman, or your state rep or senator, all of whom are still human. Human supervision is the solution to all of these problems.
The Solution
Nos custodiet ipsos custodes. We will guard the guardians. Or perhaps Nos custodiet ipsos automatons is a better use of that phrase.
In short:
We must never allow robot overlords.
There cannot be robots working for the government, police, and military which are completely and totally secret. In the republican tradition, there must be some independent, disinterested, civilian, human force established to supervise them. The code in electoral machines must be made visible to all so that we can watch it for "errors." Secure code, such as police and military code, must be reviewed by independent specialists who ultimately answer to the people, so that society can both have security and secrecy, but be protected from malicious code, whether created by a hacker or a corporation.
The people must be able to supervise and regulate robotics.
Civilian, human supervision is the most important thing here. An office as part of every city government made up of programmers and public policy workers who supervise the police, fire, and city bureaucracy. Appointed or elected, but with security clearance, and access to any and all classified information. Any policeman or commander who fails to deliver fully unclassified information to this body would be guilty of felonious obstruction of justice.
The same would go for the military. Congress, and the Executive branch should both have full supervisory authority over the military. There should be a civilian office set up by the executive branch whose entire purpose is to provide civilian oversight of military robotics.
Every action should have a black box showing all orders given by the commanders, in addition to random screening of operations, every time a civilian is killed, those black boxes are examined. If a black box is damaged or destroyed, it's a court-martial offense under military justice.
Civilian oversight of military actions is the only way to make sure that our current drone war isn't a permanent, commonplace thing that more powerful nations do to less powerful ones.
As long as we are capable of voting, as long as the people have the capability to control our government, we're going to be okay. Because a sovereign government, the nation state, is the only thing that gives us squishy humans the ability to triumph over the bots who answer to oligarchs.
And when it comes to the robo-bureaucracy the entire purpose of having a congressional office or a city council is to provide human supervision of a robotic bureaucracy. When the robot can't help you, when there's no one in the city management to talk to, you talk to your city councilor. Their office helps you. Their office is the override switch, because they're the people who are in charge of the robots.
We need to start changing he laws, today.
And that is why the following three goals are the most important things we have to do in the 21st century to prepare for what's coming, and what's already here.
1. The passing of anti-corruption laws that eliminate the American pay to play system of government, guaranteeing that it is the people who vote who control our government, and not monied interests (I.e. the people who are about to own a lot of robots.)
The growth and protection of our constitutional liberties and our right to self-government, and the need for it, should be obvious to all Americans. I imagine that even the Tea Party wouldn't want robots deciding their constitutional liberties.
2. The re-establishment of state sovereignty, globally, as the final legal authority in all matters.
From NAFTA to FIFA, we've let corporations trample state sovereignty lately. When a corporation can Seize an Argentine military vessel because Argentina refused to pay extortionate interest rates and fines, that's an international crisis. When FIFA gets to force the Brazilians to allow beer into stadiums on behalf of Budweiser, only to stand back and act surprised at the resulting violence, that's a problem. If states can't exercise sovereignty, then democracy loses its teeth, and the rise of oligarchs who control robots is assured.
3. Possibly most importantly, we need guaranteed minimum income. Every citizen must be guaranteed a living wage. If we're going to hit 65% unemployment, and we will, then the wealth created by that system needs to return to the people.
There will never again be enough jobs for everyone to work. The we are losing in transportation, construction, and food service will never come back, they will not be replaced by imaginary jobs of the future. Period. Full stop. End of story.
That's because the jobs of the future will be new ways to implement robots, not new places to put humans temporarily until we can figure out how to replace them with robots.
So when we have all these humans who are permanently unemployable, we're either going to need to make sure they have plenty of bread and circuses, or they're going to hack our death bots, and institute a communist revolution. Or they're just going to democratically vote in a communist state, because they'll have the numbers to do that. And while I'm fine with truly democratic socialism, the old style soviet state capitalist/communist system where people don't have freedom of speech is something that I'm not interested in at all. And I don't think anyone here who's anti-capitalist is interested in re-instituting a soviet style system. That's not what they're talking about. I know them well enough to know that.
But in a situation where there is an active revolution, there's no way to guarantee that those kinds of authoritarian assholes don't come to power.
So we need to implement some kind of system where the now permanently unemployable are given the economic power they need to thrive.
And there are some people unemployed today, who will never have a job again. Ever. They are permanently unemployable, through no fault of their own.
I, for one, am working on grabbing a technology job. The kind that I hope will survive the rise of the robots. It's basic self-interest there. But eventually, even I as someone who does the security side of IT will be replaced with a robot. I'll just lose my job later than everyone else. And not everyone is capable of doing IT Security work. To paraphrase Einstein, everyone's a genius at something, but fish can't climb trees, and monkeys can't breathe water.
This doesn't have to be scary, it can be beautiful.
Imagine a world where no one needs to work. Where the ones driven to do the available work, do. Where humans focus on the creation and exchange of luxury goods and services.
There will still be bars with human bartenders. Not because it's efficient, but because some people want to run a pub, want to cook for people, want to serve beer. And other people want to go to these bars where "everybody knows your name" and all that.
There will be robotic table servers working alongside waiters who will be Sommeliers or food ambassadors, there to discuss the artistic creations made by a human chef. The robots will seat people and serve the food, but the server will be there to provide a human touch, and human guidance.
And then there will be a McDonalds completely devoid of human workers, which exists as a bank of food-dispensing ATMs. The drudgery jobs, from dish-washing to line-cooking to fast food will be replaced. The jobs I just discussed won't be jobs. They'll be something people do because they wish to do it, not because they're being paid to do it.
People will cook for each other. We will grow luxury foods, heirloom tomatoes and such. And we'll do it not for economics, but because we want to. We'll give their tomatoes away, or trade them for some other luxury item created by someone else. Or just eat them ourselves.
We are approaching a world without need. A world of infinite resources. A world where those who desire a frontier are working on making mars livable. A world where the economic bounty of space is being returned to earth by robotic ships. A world where all human needs are provided for, and where everything is essentially run like Wikipedia is. No one is paid to write articles by Wikipedia. It just happens. No one will need to be paid to do the work that is coming, because everyone will already have the economic support they need. The work itself will become its own reward. The gratitude of the community will be the profit motive.
As long as we protect, grow, and repair our democracy, we will have the tools we need to achieve this world.
Because when 65% or more of society is made permanently unemployable by robots, we'll have to vote ourselves the economic resources we need to survive. We'll have no other choice.
And if we fight the hard fights now, for guaranteed minimum income, no corruption in government, and a return of national sovereignty, we'll have a society capable of dealing with the coming human wave of permanent unemployable humanity without resorting to warfare or violence.
Oh, and if it comes to violence, the rebels will win. Easily. With little bloodshed. Revolt against a robot-supported oligarchy isn't just going to be possible, it's going to be easier than revolt against our current government. The riot police and the army can't be hacked and reprogrammed.
And as long as the people have more hackers than the oligarchs, (Pro-Tip, there will always be more of us than there are of them) the people will be able to overwhelm the oligarchy. Just look at what happened to HB Gary when it decided it was more powerful than the people.
If we the people need to fight a war against the wealthy for things like guaranteed minimum income, then we're not just going to take guaranteed minimum income. We're going to act punitively in such a circumstance. And I think we've seen enough gritty newsreels from Russia in 1917, or Romania in December of 1989, to know the kind of retribution that happens to the powerful when the people are forced to take up arms.
So it's really in the best interests of the wealthy to go along with the plan of giving everyone who's becoming permanently unemployed a guaranteed minimum income that's enough to thrive, not just survive.
And I think that over time, the rational self-interest of people like the Gates and the Soroses will override the royalist ideology of people like the Kochs.
As long as we do the hard work of building society and preparing for post-scarcity economics, and as long as we protect our democracy, we'll be able to achieve not a utopia, a world without problems or pain, but certainly a world without poverty, and squalor. There will likely be new problems coming down the pipes that we haven't even imagined yet. (Like the question of when synthetic life becomes sentient and sapient and gains certain rights.)
But we're on the cusp of a revolution that is about to change everything we do. A revolution akin to the agricultural revolution. We've seen it coming for a while, imagined it for a while.
And with the invention of things like Baxter, and Scribe, that day is here. And things are going to change very, very rapidly in the next few years.
And we've got to get ready, because things are not about to change, they are already changing.
So we've got to start thinking about the problems we haven't imagined. And we've got to start solving the problems we know are coming.