United States California approves 1st state-funded guaranteed income plan

A large part of my doubts regarding the "every job is going to be automated away" I think boils down to two points:

1) Humans as machines are extremely cheap to upkeep. At current prices humans can be fed fairly good food for about $4,000 per year, a little over $10 per day. Many people make do with $1,000 a year or less. Medical insurance can be had for most functional people for about $3,000 to an acceptable quality. Good quality housing and utilities can be done at about $3,000 in material costs (ignoring extremely high land prices).

Therefore, the human machine can be supported, in conditions fairly comfortable even for western humans, for about $10,000 dollars. Third world conditions, or just low first world, can make do on about $5,000, once again not at particularly squalid conditions. Indians for example seem to support themselves (and probably more than themselves) on an average monthly income that implies about $5,000 per year wages.

This is the price needed to beat for machines vs people. We would expect wages to be regressing to this level.

2) Wages globally have been rising. Even in the West I recall people saying wages have been going up, and that the static wage stats are not accurate. The number of people employed has been increasing. In just the US, the labor force has increased by about 30%, from 125 million to 165 million. The complaint generally does not seem to be people not employable at any wage, but that the opportunities are such that working for $5 an hour, which is a $10,000 dollars a year job full time, enough to cover the above expenses. Much more actually if the job has benefits, since the $10,000 variable cost assumes out of pocket health care.

Much of the automation I'm hearing about now seems driven, not by the automation being cheaper than cheap labor, but because they can't find enough people to do it at $30 an hour.

The labor problems as they manifest, at least as I can see, do not look like something coming from automation being cheaper than people. That is not the problem most seem to be dealing with. Instead, its that people are so expensive that it necessitates investment in expensive automation.

JagerIV said:

This is the price needed to beat for machines vs people. We would expect wages to be regressing to this level.

Click to expand...

Eventually, we're going to get to the point where people will have to accept Chinese levels of wages (think about $0.50 an hour) in order to compete with automation. It may not happen within our lifetimes, but it will happen.

JagerIV said:

Wages globally have been rising.

Click to expand...

Not when you factor in inflation.

JagerIV said:

A large part of my doubts regarding the "every job is going to be automated away" I think boils down to two points:

1) Humans as machines are extremely cheap to upkeep. At current prices humans can be fed fairly good food for about $4,000 per year, a little over $10 per day. Many people make do with $1,000 a year or less. Medical insurance can be had for most functional people for about $3,000 to an acceptable quality. Good quality housing and utilities can be done at about $3,000 in material costs (ignoring extremely high land prices).

Therefore, the human machine can be supported, in conditions fairly comfortable even for western humans, for about $10,000 dollars. Third world conditions, or just low first world, can make do on about $5,000, once again not at particularly squalid conditions. Indians for example seem to support themselves (and probably more than themselves) on an average monthly income that implies about $5,000 per year wages.

This is the price needed to beat for machines vs people. We would expect wages to be regressing to this level.

2) Wages globally have been rising. Even in the West I recall people saying wages have been going up, and that the static wage stats are not accurate. The number of people employed has been increasing. In just the US, the labor force has increased by about 30%, from 125 million to 165 million. The complaint generally does not seem to be people not employable at any wage, but that the opportunities are such that working for $5 an hour, which is a $10,000 dollars a year job full time, enough to cover the above expenses. Much more actually if the job has benefits, since the $10,000 variable cost assumes out of pocket health care.

Much of the automation I'm hearing about now seems driven, not by the automation being cheaper than cheap labor, but because they can't find enough people to do it at $30 an hour.

The labor problems as they manifest, at least as I can see, do not look like something coming from automation being cheaper than people. That is not the problem most seem to be dealing with. Instead, its that people are so expensive that it necessitates investment in expensive automation.

Click to expand...

That isn't the case anymore. You're still using a series of assumptions that -simply put- have been annihilated by the evolving technological context. Machine learning has literally changed the game to the point that there is no game, so to speak. The only limit is the time to get the metadata necessary, and with sensor density becoming -quite literally- yes, that isn't going to take all that long or with the employee's knowledge.

Terthna said:

Eventually, we're going to get to the point where people will have to accept Chinese levels of wages (think about $0.50 an hour) in order to compete with automation. It may not happen within our lifetimes, but it will happen.

Click to expand...

That isn't the case, actually. It'll happen within our lifetimes being the most probable unless a nuclear war starts (and I'm dead twenty-eight days from Sunday due to the fact that I'm sitting next to a major nuke target, just down the hill of a second, and across town from a third).

Not when you factor in inflation.

Click to expand...

If the minimum wage kept up with inflation? Well back in the 20-aughts, that number would have been somewhere around $25. Now? ~$30. If it kept up with productivity? Back in the 20-aughts, it would have been ~$30 and now likely ~$50.

Bear Ribs said:

The significant part here is that the horse was a component in transport, specifically motive power, and when that component was replaced horses became useless and were reduced to pets. Humans were not affected because the part of the transport they represented, steering and control, was not replaced. Now the human component, steering and control, is being replaced while we watch.

"A horse is not a human" is a pointless objection. An AI is not a human either.

Click to expand...

A horse and an AI have more in common with each other, than a human.

A horse can do a very few things better than a human can. It can lift and carry larger loads, and run faster. A human is more durable, adaptable, intelligent, capable of finer manipulation, more flexible, etc, etc.

An AI can do a very few things better than a human can. It doesn't need to sleep, it can process large numbers and do the exact same task for as long as it's provided power without ever growing bored. It cannot take care of itself, it is capable of at best extremely limited adaptability, it lacks the ability to intuitively value concepts, and link them in ways critical to simply and efficiently accomplishing most tasks, etc, etc.

Within their specific niche, a horse and an AI can both outperform a human. Outside of those very specific niches, they cannot.

The primary difference, is that a horse is a horse, and no amount of selective building will fundamentally change that. You can custom-craft AI for a task, so they have a future as economic tools, while horses are now more or less permanently consigned to a status as pets and entertainment.

Bear Ribs said:

You've already been given examples in this thread, specifically how things are going down right now.

Click to expand...

No, I've been given examples of industries that are probably going to see a decrease in the use of human labor. I've pointed out how there've been industries that have had similar drops in use of human labor, or outright gone extinct in the past, yet people have always ended up in new jobs.

What you need to give for it to count as supporting proof for your position, is an example of an industry that is not in danger of going under, but has gone under, and the people who worked in it then ended up permanently unemployed. Not a few members around the fringe becoming permanently unemployed, but a majority, hell I'd settle for a third ended up never getting a job again.

Bear Ribs said:

@DarthOne presented a vid that pointed out how the increase in jobs and hours of work is not keeping up with growth in population anymore and hasn't been for some time. Real wages have not kept pace with inflation since the 70s. Basic necessities have become more expensive, inflation-adjusted, and have been getting more expensive for decades. Thirty years ago Homer Simpson, stupid, lazy, and with no college degree, was seen as reasonably able to support a family of five and own his own suburban home. Today that would be considered ridiculous.

Since you missed the point of the dinosaur analogy let me present you with a different one. Can you give one historical example, just one, of where a human being walked on a different planet than earth? Just one?

If not do you believe that it's physically impossible for humans to ever set foot on another planet, say Mars, because it has not happened before? Or do you accept that, actually, as technology advances things that haven't happened before in history can happen?

Click to expand...

The idea that people are poorer than they were in the 70's, is bogus. I don't know what kind of statistical manipulation that video is using, but it is quantitatively untrue, and I think I've had this argument with you before.

It is now standard in America in all houses middle class and above, to have both heating and air conditioning. The majority of poor houses in the US have both. In addition to that only the absolute poorest of the poor do not have:

Cell/smartphones.
Microwaves.
A refrigerator.
A television, these days probably a large screen.

In addition, a majority of the 'poor' also have:
One or two cars owned in their household.
A laptop or desktop computer.
Either their own washer and dryer, or access to free-use ones as part of their apartment complex.

Among these, a minority would have possessed the fridge or laundry machines, microwaves were just barely starting to see home use in the 70's, and air-conditioning at home was for the rich, or at least the comfortably well off.

The idea that people are poorer now than they were in the 70's is complete bullshit that is easily disproven by a basic look at what people have now, that they did not have.

That is not to say that there are not problems. Interference in property markets via things like rent controls and regulatory burdens blocking new home construction are serious problems, and have grossly and disproportionately inflated the amount of their income that people spend on housing. There are lots of other problems as well, but this is one of the worst, especially for those of lower incomes (like myself).

The value on-balance of hourly wages can also legitimately be a problem. There's a variety of ways to try to fix that too, but BLS/UBI would again make the problem worse, not better, for the same reason it'd fail the unemployment problem: Paying people not to work solves no problems, and creates enormous new ones.

The sort of problems that usually are key to making a civilization collapse.

Aaron Fox said:

and with sensor density becoming -quite literally- yes

Click to expand...

You have never substantiated your "smart dust" bullshit before. Show me the evidence that sub-millimeter fully contained anything is remotely able to be done with current technologies, as in literally shit like silicone semiconductor chips. Can you substantiate the basic methodolgies currently in use are actually at fucking all capable of this?

I have looked at what the machine learning papers bring up. They cannot be relied on to handle shit themselves, the basic methodology we use is too unstable to function without extremely specialized programmers setting up the exact system of weights and bugfixing the results. Period.

It does not matter how many sensors you plug in, how many weights you use, how carefully you design it. The core methods of current machine learning cannot generate an Artificial General Intelligence as necessary for this shit to happen. Because, quite simply, the modern technology of machine learning is iteration.

The systems do not adapt over time once implemented, they're far too sensitive to be functional that way, they'd rapidly turn into useless garbage. That is how the current technology operates, you'd have to make a fundamentally different kind of machine learning to cross this bridge.

And before you bring up search engines, those things are ancient code by machine learning standards. They crunch very well-known variables in very well-known conditions. The things are just highly sophisticated best-fit algorithms, not actual machine learning (which, mind you, is actually still derivative of such), for all they have appearances of such. Very different programming processes.

By appearances, your mind has been consoomed by the brainbug of Skynet bullshittery and infinite technological progress. Throw that in the trash, and throw that trash into a fucking black hole: Modern machine learning systems cannot replace truckers, fundamentally. Because the basic way the image recognition operates has several well-documented impossible to remove vulnerabilities, and fixing the edge-cases requires an open failure to know the edge case is even there to fix.

Fine in a lab, but fixing those issues in real world application means every instance of the system that encounters the edge-case actively fails. Every car that goes by an adversarial patch is promptly shoved into wildly unintended behavior. Every truck that ends up encountering novel lighting conditions may well careen off the road. Machine learning isn't solving that. The main solutions being used for that are extremely painstaking and specific manual code to idiot-proof the lauded "machine learning".

Oh, and even once you do have the machine learning able to do all the physical labor from robotic chassis, you still need all the same technical staff to design the physical machinary that actually does the building for those programs to run. Because we don't have general-use manufacturing, and the current basic methodology of 3d printing is not suitable for replacing a meaningful portion of the current manufacturing industry.

One cannot 3d print an airplain driveshaft. Metallurgy does not work that way. One cannot 3d print a fully-functioning computer. Semiconductor processors do not work that way. There are reams and reams of massive fundamental techbase changes that have to happen before this is a realistic possibility. We overwhelmingly still produce bulk steel the way we did forty years ago, just more precisely. Same in most other industries. Because once the first wave of robotics got set up, we more or less settled into the best paradigm our current technologies could offer, just refined it with more precise machinery and more powerful computers.

Machine learning is currently like smartphones. It opens up new things, but only by making old theoretical work into practicalities. And much like smartphones, we were working with it for a very long time until physical hardware got to the point of supporting it. We're never getting a 1nm chip on current technologies. The physics do not work. Truly, dramatically, new technologies are required to make the leap to AGI.

Quantum computers of the basic kind we have now aren't going to support your image of the world. Supercooled servers aren't going to be a functional proposition. If they're centralized, they're too vast a security issue and the latency fucks over a load of applications. If they're local as needed for the "print everything" model, you're losing obscenely vast reams of capability because it scales exponentially.

Local McDonald's runs at like a third the manpower of 20 years ago with way higher throughput and quality. You dont always need high end computers and software to kill off a lot of jobs.

Like how much longer until grocery stores are a use app and pickup or get delivered by robotruck? Fraction of the workers, a lot less theft (see San Francisco) higher quality/better date on things because nobody else touched those tomatoes and bruised them looking for unbruised ones etc.

I expect that as is right now if it wasnt for business inertia at least half the population would be on a position where the value of the work they can do is less than the money needed to keep them to an acceptable standard of life.

Maybe it would be a better tactic to try to lower the costs of living?

Captain X said:

Maybe it would be a better tactic to try to lower the costs of living?

Click to expand...

Post stalin style Soviet apartment blocks and plant based only food ration packs infused with pacification drugs for 2/3rds+ of the population is what I consider a good outcome long term.

Edit: Do this for basic human storage then allow luxuries to be earned by attending education non disruptively, but your attendance records and behavior are effectively a resume if a job does open up.

Stormin said:

Post stalin style Soviet apartment blocks and plant based only food ration packs infused with pacification drugs for 2/3rds+ of the population is what I consider a good outcome long term.

Edit: Do this for basic human storage then allow luxuries to be earned by attending education non disruptively, but your attendance records and behavior are effectively a resume if a job does open up.

Click to expand...

Because that doesn't sound like a total nightmare or a system that could be gamed with ease by governments to force people to toe the line....

DarthOne said:

Because that doesn't sound like a total nightmare or a system that could be gamed with ease by governments to force people to toe the line....

Click to expand...

The alternative is of course that civilization will collapse as the surplus to the needs of the workforce decides to bring everything down. Guillotine memes will become real, people will turn to communists and worse etc.

The people wont starve quietly but there is not enough resources to have luxury lifestyles like current lower middle class or upper lower class just given out. And since there will be no hope for a better future for the people in 99% of cases then some form of control is needed.

To those who doubt how insane 3d printing is getting and called it a tool for prototyping and toys











And this is still new tech who knows what will happen in a decade

Separate from the 3d printing just watching the Tom Scott video on robot grocery stores should make you aware of how close we are to a huge shift in how we will have to think about how humans work

Captain X said:

Maybe it would be a better tactic to try to lower the costs of living?

Click to expand...

I think the best tactic would be to explore ways to actually increase how well humans can accomplish things in order to keep a human relevant. We've been doing that for a long time by increasing education, but we're starting to run into some hard limits on just how much education you can give a person before you start cutting into their useful career years. I suspect we can actually make education far better than it is now and probably need to explore augmentation options, though again that also opens up a pretty massive can of worms itself.

LordsFire said:

A horse and an AI have more in common with each other, than a human.

A horse can do a very few things better than a human can. It can lift and carry larger loads, and run faster. A human is more durable, adaptable, intelligent, capable of finer manipulation, more flexible, etc, etc.

An AI can do a very few things better than a human can. It doesn't need to sleep, it can process large numbers and do the exact same task for as long as it's provided power without ever growing bored. It cannot take care of itself, it is capable of at best extremely limited adaptability, it lacks the ability to intuitively value concepts, and link them in ways critical to simply and efficiently accomplishing most tasks, etc, etc.

Within their specific niche, a horse and an AI can both outperform a human. Outside of those very specific niches, they cannot.

The primary difference, is that a horse is a horse, and no amount of selective building will fundamentally change that. You can custom-craft AI for a task, so they have a future as economic tools, while horses are now more or less permanently consigned to a status as pets and entertainment.

Click to expand...

And in comparison to a human, an AI is more durable, capable of finer manipulation, and more flexible. They aren't more adaptable or intelligent yet but both of those traits are increasing by the year. The issue is that AI's specific niches happen to overlap with most of humanity's specific niches.

No, I've been given examples of industries that are probably going to see a decrease in the use of human labor. I've pointed out how there've been industries that have had similar drops in use of human labor, or outright gone extinct in the past, yet people have always ended up in new jobs.

What you need to give for it to count as supporting proof for your position, is an example of an industry that is not in danger of going under, but has gone under, and the people who worked in it then ended up permanently unemployed. Not a few members around the fringe becoming permanently unemployed, but a majority, hell I'd settle for a third ended up never getting a job again.

Click to expand...

Which cannot be accomplished, because the definition of "Unemployed" used in the US only counts people currently looking for a job, so Unemployed who give up finding work, aka those permanently unemployed, are automatically removed from consideration, which allows the data to be manipulated to never count the permanently unemployed. Nice Catch-22 you have there.

And I'll repeat, do you think it's impossible for humans to walk on Mars or do you concede that sometimes things change due to technology?

The idea that people are poorer than they were in the 70's, is bogus. I don't know what kind of statistical manipulation that video is using, but it is quantitatively untrue, and I think I've had this argument with you before.

It is now standard in America in all houses middle class and above, to have both heating and air conditioning. The majority of poor houses in the US have both. In addition to that only the absolute poorest of the poor do not have:

Click to expand...

Show me the numbers then. Present proof, I'm tired of being the one using sources while you just make vague assertions (that have already been proven false in-thread) and just fail to quote any point you can't argue with over and over. You say it's quantitatively untrue so you must have actual quantities, show them to us.

Cell/smartphones.
Microwaves.
A refrigerator.
A television, these days probably a large screen.

In addition, a majority of the 'poor' also have:
One or two cars owned in their household.
A laptop or desktop computer.
Either their own washer and dryer, or access to free-use ones as part of their apartment complex.

Among these, a minority would have possessed the fridge or laundry machines, microwaves were just barely starting to see home use in the 70's, and air-conditioning at home was for the rich, or at least the comfortably well off.

The idea that people are poorer now than they were in the 70's is complete bullshit that is easily disproven by a basic look at what people have now, that they did not have.

That is not to say that there are not problems. Interference in property markets via things like rent controls and regulatory burdens blocking new home construction are serious problems, and have grossly and disproportionately inflated the amount of their income that people spend on housing. There are lots of other problems as well, but this is one of the worst, especially for those of lower incomes (like myself).

The value on-balance of hourly wages can also legitimately be a problem. There's a variety of ways to try to fix that too, but BLS/UBI would again make the problem worse, not better, for the same reason it'd fail the unemployment problem: Paying people not to work solves no problems, and creates enormous new ones.

The sort of problems that usually are key to making a civilization collapse.

Click to expand...

It's funny how utterly predictable this response is. It's such a canned response, I have yet to ever have this discussion without somebody bringing up "But muh luxury goods!" in response to pointing out that luxury goods are cheaper but basic necessities are more expensive. Well no duh luxury components are cheaper, I'm the one who pointed it out.

Actual charted data shows that necessities are more expensive. Luxury goods have gotten cheaper.


There's a reason we see people talking about how eating bugs is the future, necessities like food are getting exceedingly pricey and people who are looking to the future expect that meat will be out of reach for average humans eventually.

And regardless of Luxury Goods, you run headfirst into the problem of who's paying for it, which you gloss over every time and just don't quote those points since you can't refute them. Only one person needed a high-school diploma and job to support a family of five, their car, and owning their own home, and they began having children right out of high school. Today both parents need college education and full-time jobs to afford a small apartment's rent and people on sites like SB get the heebie-jeebies at even the thought of children because kids are so expensive. Even if the house has more air conditioners, if you have to have all the adults working to pay for it instead of just one you cannot indicate that things are getting cheaper from that.

Terthna said:

Eventually, we're going to get to the point where people will have to accept Chinese levels of wages (think about $0.50 an hour) in order to compete with automation. It may not happen within our lifetimes, but it will happen.


Not when you factor in inflation.

Click to expand...

Well, that's the thing, Chinese wages have doubled in the last 10 years. And that's on top of a double in the previous 10 years. Average Chinese wages are in the $1,000 per month range.

I think its the non-thinking parts that are actually the hardest part of the human to replace. Machine learning doesn't load and unload boxes. Or do plumbing work. That requires an actual robot, and that runs into the issue of, well, the human being a fairly cheap machine to work.

I think Peter Thiel's is correct that adding a billion foreign workers doesn't look very much like an automation crisis.

Some of this doesn't necessarily mean we won't have an automation crisis, the issue is with any argument we are already seeing an automation crisis. We are not seeing any of that. The crisises were seeing seem to be more or less unrelated to automation:

1) terrible, overly expensive dysfunctional health care system. Sky high prices, extremely high wages, and medical companies going bankrupt does not sound like an automation crisis.

2) Terrible, overly expensive dysfunctional education system. This does not seem like an automation crisis.

3) Extremely high property prices (in certain areas). This does not seem like an automation crisis.

Edit:

@Aaron Fox , inflation adjusted minimum wage peaked in 1968 at $11.39, not $20 or some other ridiculous number.



And notice that it immediately started falling down. This peak also happened right before 1971, which is one of the dates people point to of when things started going crazy.


That's one of the reasons for doubt I have, is that there's a lot of people blaming automation, and specifically machine learning automation, for trends that have been going on for 50 years, and well, then your dealing with the earlier issue that a time period (1970-2020) where we added basically 1-2 billion workers to the global economy in the industrialization of China, Japan, Taiwan, and India, which didn't really start till this period, at least en mass does not look a whole lot like a general automation crisis: if what's happening is factory jobs are moving to China, UBI doesn't seem like it solves that. Most problems seem to be tangentally related to automation.

So WTF did happen in 1971? I'm not seeing a cause.

Typhonis said:

So WTF did happen in 1971? I'm not seeing a cause.

Click to expand...

Well, something clearly happened in 1971, but like you say its not clear exactly what. Listening to an interview with the people who put it together, they're classical libertarians/austrians, so their instinct of what to blame is the closing of the gold window and moving to more "funny money". They admit though that closing the gold window is as much a response as a cause to several other problems, as part of the broader Nixon shock.

I think abandoning the Bretton Woods system may be the reason things become really noticeable in 1971, but it wasn't just abandoned willie nilly either, and the system set up post WWII was deeply flawed and probably doomed to collapse at some point anyways. Notably, Germany had already dropped out of the Bretton Woods system pre the US, so its somewhat questionable if the US could have kept on the system post 1971 anyways really.

But, its definitely an open for debate conversation. 50 year trendlines are probably always going to be multi-variable issues.

JagerIV said:

Well, that's the thing, Chinese wages have doubled in the last 10 years. And that's on top of a double in the previous 10 years. Average Chinese wages are in the $1,000 per month range.

I think its the non-thinking parts that are actually the hardest part of the human to replace. Machine learning doesn't load and unload boxes. Or do plumbing work. That requires an actual robot, and that runs into the issue of, well, the human being a fairly cheap machine to work.

Click to expand...

Machine learning has already advanced past loading and unloading boxes.

The first plumbing and construction robots are also coming online right now. They're far from able to fully replace humans but it's undeniable that they're advancing.

Jettyrobot's particularly interesting, as it's a drone that can enter the pipe and cruise down it in a way to human plumber could possibly do. Plumbing is still far fuzzier than what current AIs can handle alone (Little standardization and a lot of weird edge cases) but computers have advanced ridiculously in a short period of time and continue to do so. Automated 3D printed houses will lead to robots having a much better idea of how the house is constructed and make it easier for them to take over maintenance of the home because they also built it.

There's also the issue that anti-automation debaters tend to pull out "Demand is infinite so we'll always want more stuff" but things like plumbing absolutely don't have infinite demand, there's a limited number of sinks and toilets needed and people won't demand infinite sinks and toilets.

I think Peter Thiel's is correct that adding a billion foreign workers doesn't look very much like an automation crisis.

Some of this doesn't necessarily mean we won't have an automation crisis, the issue is with any argument we are already seeing an automation crisis. We are not seeing any of that. The crisises were seeing seem to be more or less unrelated to automation:

1) terrible, overly expensive dysfunctional health care system. Sky high prices, extremely high wages, and medical companies going bankrupt does not sound like an automation crisis.

2) Terrible, overly expensive dysfunctional education system. This does not seem like an automation crisis.

3) Extremely high property prices (in certain areas). This does not seem like an automation crisis.

Click to expand...

I would agree with you on most of those, but not education.

Extremely expensive education systems are a direct result of automation (albeit not the only cause, just one of them). Simply put, jobs that can be done by the uneducated and incapable are now done by robots. You can't get by with a grade-school education like you could a hundred years ago because robots/tools that let one person do the work of pre-automation ten are doing all the grade-school education jobs. You can't get by with a high-school education like you could forty years ago because robots are doing those jobs now. You have to have a college education of several years to do anything a robot can't anymore, and that's pushed expensive education (plenty of other issues too, to be sure). This leads to people not being able to afford low-pay jobs because they have to pay for their expensive educations... but in real-dollars adjusted for inflation they're getting paid less, not more, for the same work, a trend that's been going on as humans have to work for less in order to compete with increasingly capable and decreasingly expensive machines.

Typhonis said:

So WTF did happen in 1971? I'm not seeing a cause.

Click to expand...

All the countries that had massive oil supplies, that is most of the Middle East, banded together to form OPEC a few years earlier, and in 1971 formed the Tripoli Agreement where they enriched themselves by strangling the oil supply and raising prices. Oil went from 2.55 to 3.45 a barrel in the space of a few hours on April 2, 1971. The world never really recovered from this.

The 70s were also when the first home computers and major business computers started to appear, and automation began to really take off, leading to the current situation we're facing.

The 1970s marked an increasing trend within the machining industries towards automation. Advances in technology allowed presses, more so than ever before, to perform multiple operations at a single location simply at the touch of a button. A single operator was required to press that button and make sure the machine ran smoothly. Although these advances had positive benefits, primarily in terms of safety for workers, increased automation also meant that fewer people were required to produce an equivalent output. As a result, companies reduced their numbers of hourly workers drastically.

@Bear Ribs

My main argument with the boxing is that machine learning isn't doing that, a forklift is doing that. Things that are pure information, like tax filing or playing music, are very automatable, and potentially very cheap. Automating for example a trumpet player is dirt cheap: I'm pretty sure there's free software at this point which will mostly play a trumpet part if one was composing some song that needed a trumpeter.

Physical labor is much more difficult and expensive to automate, because, well, you need the robot, which means manufacturing costs, upkeep ("healthcare") costs, and energy costs. Its an even more difficult and expensive system if you need to combine physical with thinking. My cousin who does landscaping work probably has way more job security than I do as a Tax Preparer. A machine that can cut grass, trim a hedge, dig a ditch, clear a drainpipe, plant flowers, discuss better lawn layouts, maintain all the equipment, and negotiate pricing for all that is a much more expensive ask than automating data entry of W-2s and such.

Is it possible to automate all of that? Of course. If he stays in the landscaping business and doesn't just decide to retire at 40 and live off his accumulated assets and rents (the landscaping business has been very good to him) he'll probably end up buying some. Is it cheaper and more efficient than a person doing it? Well, probably not at current manufacturing capacity.

The education part would be where I think we have serious disagreements: education, mostly, especially as we have it, is completely worthless to do any of those things. White collar work which is what all this education is mostly geared towards, if the education is worth anything at all, are exactly the kind of work machine learning replaces. Its clerks and paperwork people who make $10 or so an hour, while laborers make $20 an hour.

If we had the grade school education of a 100 years ago, most people would be more employable with that and 5 years of work experience than even completing high school, let alone college. Much of current college is a giant wealth destroying black hole actively impoverishing the lower classes.

The problem with education is that education has objectively gotten worse while becoming super expensive, and providing no meaningful job training or even signaling value at this point. Partially its my understanding the current obsession with it college degrees partially grew out of banning of other, cheaper sorting methods. As another thing that happened in 1971, Griggs v. Duke Power Co. basically banned general aptitude tests for employment. This was part of more broad Disparate impact nonsense that really ran wild in the 70s, though obviously it continues.

JagerIV said:

@Bear Ribs

My main argument with the boxing is that machine learning isn't doing that, a forklift is doing that.

Click to expand...

But that's meaningless. By your own standard there, a human also isn't doing the boxing, a forklift is. So what's your point?

That makes everything else you're arguing is irrelevant, the cost of the forklift is meaningless because the company has to buy a forklift either way. The difference, then, is if they have to pay a human an hourly wage to operate or have a copy of an AI operate it for free.

Physical labor is much more difficult and expensive to automate, because, well, you need the robot, which means manufacturing costs, upkeep ("healthcare") costs, and energy costs. Its an even more difficult and expensive system if you need to combine physical with thinking. My cousin who does landscaping work probably has way more job security than I do as a Tax Preparer. A machine that can cut grass, trim a hedge, dig a ditch, clear a drainpipe, plant flowers, discuss better lawn layouts, maintain all the equipment, and negotiate pricing for all that is a much more expensive ask than automating data entry of W-2s and such.

Click to expand...

For every physical task that isn't purely muscle-powered you need a machine anyway, so it can be automated by putting an AI in charge of that machine instead of a human. Self propelled lawnmower? Put an AI in it and don't pay the landscaper anymore. Actually, we already have a wide range of robotic lawn mowers for sale. Trim hedges? Add a hedgetrimmer attachment to the lawnmower. Dig ditches? We already use tractors for that, put an AI in it and you don't need pay a human anymore. Clear a Drainpipe? I linked to multiple plumbing robots above. Plan a lawn? They're training machine learning systems to do that right now. Negotiate pricing? They're already using machine learning systems for that.

Bear Ribs said:

I think the best tactic would be to explore ways to actually increase how well humans can accomplish things in order to keep a human relevant. We've been doing that for a long time by increasing education, but we're starting to run into some hard limits on just how much education you can give a person before you start cutting into their useful career years. I suspect we can actually make education far better than it is now and probably need to explore augmentation options, though again that also opens up a pretty massive can of worms itself.

Click to expand...

No, we've been exploring that primarily through more advanced technology making labor more efficient.

Education fills a head with knowledge, some of which might be false. It has its use, but better tools and actual work experience are what improve 'how well humans can accomplish things.'

Bear Ribs said:

And in comparison to a human, an AI is more durable, capable of finer manipulation, and more flexible. They aren't more adaptable or intelligent yet but both of those traits are increasing by the year. The issue is that AI's specific niches happen to overlap with most of humanity's specific niches.

Click to expand...

It's less durable, less flexible, and only capable of finer manipulation after enormous investments of man-hours and finances to make it so.

Using current transistor-based technology, AI will never be more adaptable, because the fundamental nature of how AI 'thought' operates compared to human thought sticks them with fundamental limitations they cannot overcome. They can operate faster in specialized math-based tasks, and in that specific way be more intelligent, but easy grasp and understanding of concepts is something that is not possible based on current technology.

Now, if a completely new type of computing is invented, we may get computers capable of doing such. But the alternative to the transistor to base that off of doesn't even exist yet, much less having been developed into a mature technology capable of surpassing human thought.

Bear Ribs said:

Which cannot be accomplished, because the definition of "Unemployed" used in the US only counts people currently looking for a job, so Unemployed who give up finding work, aka those permanently unemployed, are automatically removed from consideration, which allows the data to be manipulated to never count the permanently unemployed. Nice Catch-22 you have there.

Click to expand...

This is not a catch-22 for me, this is a catch-22 for you.

I can and have pointed out the entire history of the first world over the last 200 years as evidence that new technology doesn't result in permanent unemployment. You can't point to any piece of new technology that has, and now you're actually claiming that it's impossible to do so?

When you start claiming that it is impossible for the key evidence supporting your argument to exist, you're in a bad place.

Bear Ribs said:

And I'll repeat, do you think it's impossible for humans to walk on Mars or do you concede that sometimes things change due to technology?

Click to expand...

Yes, things change due to technology. Human nature does not, and it is human nature which governs economics, because humans are those who are at base responsible for both production, and consumption. The details change, and some of those details are very important, but a machine is never going to earn a paycheck, go home, pay rent, and then buy food and/or luxuries.

Bear Ribs said:

Show me the numbers then. Present proof, I'm tired of being the one using sources while you just make vague assertions (that have already been proven false in-thread) and just fail to quote any point you can't argue with over and over. You say it's quantitatively untrue so you must have actual quantities, show them to us.

It's funny how utterly predictable this response is. It's such a canned response, I have yet to ever have this discussion without somebody bringing up "But muh luxury goods!" in response to pointing out that luxury goods are cheaper but basic necessities are more expensive. Well no duh luxury components are cheaper, I'm the one who pointed it out.

Actual charted data shows that necessities are more expensive. Luxury goods have gotten cheaper.

Click to expand...

Well gosh, would you look at that. There's a direct relationship between how expensive a thing has proportionately become, and how much the government has intervened with and distorted the market for that good or service.

It's almost like more government intervention makes things worse, not better.

Bear Ribs said:

There's a reason we see people talking about how eating bugs is the future, necessities like food are getting exceedingly pricey and people who are looking to the future expect that meat will be out of reach for average humans eventually.

And regardless of Luxury Goods, you run headfirst into the problem of who's paying for it, which you gloss over every time and just don't quote those points since you can't refute them. Only one person needed a high-school diploma and job to support a family of five, their car, and owning their own home, and they began having children right out of high school. Today both parents need college education and full-time jobs to afford a small apartment's rent and people on sites like SB get the heebie-jeebies at even the thought of children because kids are so expensive. Even if the house has more air conditioners, if you have to have all the adults working to pay for it instead of just one you cannot indicate that things are getting cheaper from that.

Click to expand...

This is such bullshit I don't even know how to respond.

You want some cites? I'll give you some cites.

1. I am staying in a 3-bedroom (the third is small, so let's call it a 2.5 bedroom) apartment, the monthly rent is 830 dollars a month. If I paid for that myself, rather than splitting it with one roommate, I could still afford it on a single full-time job at minimum wage. That's not possible for everyone though, so let's look at some actual numbers.

2. Let's look at rental rates in some non-crazy cities around the nation:
Top two listings in Superior, WI, ~800 a month for 2 bedroom, 1 bath.
~850-1200 a month in Cheyenne, WY, for 2 bedroom, 1 bath.
800-1200 a month in Grand Rapids, MI, though you have to scroll down a ways to get past the expensive stuff at the top of the listings.
Here I actually had to start sorting for cheap stuff; Atlanta, GA, 800-1200 for a 2 bedroom, 1 bathroom.
I know from personal experience that if you're actively looking for a cheap part of the country to live in, you can do so for considerably less.

3. Let's look at food costs. Let's start with an easily-findable list of groceries that'll get you through a week at a decent price. This outta do it. Granted, it's based on 2016 grocery prices, so let's generously slap 20% inflation on that, and say it's 24$ a week, instead of 20. Multiply by 4 for a family of 4 (closest to average American family size), and that's 100 dollars a week. Four and a half weeks in a month, and even though we've rounded up at every stage, we're still coming out to ~450 dollars a month.

We're now up to 1250-1700 or so a month.

4. A car. You can get one as low as 1500$ but I wouldn't recommend going under 3 grand unless you already know the seller and/or the car. Split that up into something spread out by 12 months, and you have 250$. Add insurance, which in Michigan runs around 300$ for six months, and Michigan is one of the most expensive states in the union for car insurance (mine dropped by 60% when I moved out of the state), and you're up to 300 a month. Tack on another 100 a month for gas and minor repairs.

1650-2100 a month in living expenses.

5. Medical treatment. This is an incredibly volatile thing. If your have good genes, don't take stupid risks, and don't have terrible luck, this could be as low as a couple hundred per person per year. One major medical expense though, can run from tens to hundreds of thousands of dollars, and aside from not deliberately doing stupid things, there's very little one can do to control that. For most people though, in their younger and healthier years, the low end of costs is the norm. If we call it 300$ per person per year (which is much more than I generally spend myself, but I'm deliberately erring high), then that's 1200 for the family of four, or another 100 per month.

6. Cell Phone/internet service. This'll run you another 100-200$ a month for a family plan, generally speaking.

This brings us up to 1950-2400 a month.

This brings together bare-bones essential costs, though I did forget to count utilities with rent, for places where that is not included. Based on past experiences, that'd probably run 200-300 a month for a family of four, though bringing our total up to 2250-2700 a month.

Now, what's the income at Federal minimum wage? 7.25 per hour. If we had two parents working full-time at minimum-wage jobs, that's 80 hours a week, or 580 per week, coming to 2320 per month, which is barely enough to scrape by at the bottom end of the spectrum. 1160 dollars personal income per month is enough to live on by yourself, if you're smart, but two incomes at that level is hard for a family to get by on. It's only 29,000 a year, assuming both parents take two weeks worth of time off over the course of the year.

This is a real financial problem for this poor family, but let's look at some expanded statistics...

How much of the labor force is earning minimum wage (or less)? 1.9% a of 2019.

Boy, that really changes the picture now, doesn't it?

Just move up to the 9 or 10 dollar minimum wage common in many states, and you easily exceed the 2700 monthly cost estimate. That's still a really tight budget, but it is livable.

Now, let's look at the situation that you claimed, two college graduates working full-time to keep living in a small apartment. Now, what is the average starting salary of a college graduate?

$53,889 a year, according to one of the top results I found on a quick search. (Granted, I bet they're excluding a lot of graduates with completely useless degrees from their results, but its those graduates own fault for pursuing useless degrees, so they're not terribly germane to this discussion. If they are included in these results, then college graduates are doing way better than 54k a year once they stop dragging the average down.)

Wow. That really changes the picture, doesn't it? That's almost twice what the two minimum-wage earners brought in together. Obviously an average isn't going to account for everyone, but if that's your average, anything approaching a normal distribution is going to mean that a vanishingly low number of people have as low an income by themselves as two minimum wage earners would have together.

And if you put two of these average earners together, you have enough income to live on twice this calculated minimum living expense, leaving plenty of funds for paying off student loans and saving for the future.


So, to sum up, your assertion that the cost of living is so high that college graduates can barely make it on two incomes, is complete nonsense from beginning to end.

In fact, only the 1.9% or so of people earning minimum wage or less are in the kind of crisis crunch you're depicting, and most people manage to start earning higher wages once they've got some experience and track record behind them in life. Before our society became as dysfunctional as it has in the last 20 years, almost all of those minimum wage earners would have been young adults just getting started in life, with a near-certainty of moving on to better paying jobs, even if they never went higher than assistant manager at a fast food place, it'd still put them decently better off than minimum wage.

Implementing more redistributive policies, and institutionalizing further entitlement mentality in our culture, will make things worse, not better, because these are a big part of what has made things so messed up in the first place.

Bear Ribs said:

But that's meaningless. By your own standard there, a human also isn't doing the boxing, a forklift is. So what's your point?

That makes everything else you're arguing is irrelevant, the cost of the forklift is meaningless because the company has to buy a forklift either way. The difference, then, is if they have to pay a human an hourly wage to operate or have a copy of an AI operate it for free.

For every physical task that isn't purely muscle-powered you need a machine anyway, so it can be automated by putting an AI in charge of that machine instead of a human. Self propelled lawnmower? Put an AI in it and don't pay the landscaper anymore. Actually, we already have a wide range of robotic lawn mowers for sale. Trim hedges? Add a hedgetrimmer attachment to the lawnmower. Dig ditches? We already use tractors for that, put an AI in it and you don't need pay a human anymore. Clear a Drainpipe? I linked to multiple plumbing robots above. Plan a lawn? They're training machine learning systems to do that right now. Negotiate pricing? They're already using machine learning systems for that.

machine learning - Landscape Architecture Magazine

landscapearchitecturemagazine.org

Retailers Pursuing AI to Help Negotiate Prices, Enhance Operations - AI Trends

Retailers expanding use of AI beyond online shopping to contract negotiations with suppliers and achieving more efficient operations.

www.aitrends.com

Click to expand...

Okay, its possible, as I've already said. What do you think your proving with this?

It all comes down to the cost of automating the person out of the system, vs continuing to use labor. This kind of automation tends to be, well, expensive. To use your robotic lawn mower, it costs $1,500 looking at the list as an averagish price. A push electric mower is about $250 bucks. Still, not hugely expensive, sure. At $15 dollars an hour, you only need to automate about 83 hours of work. If you were using mexican or teenagers, depending on the stereotype at $5 an hour, 250 hours of mowing work.

But, if your a rental company, you need something unload and load the mower. So you need a machine to unload something from the truck. That's probably a not insubstantial piece of equipment, compared to someone picking up a lawn mower and placing it down. Automating that is probably going to be about as expensive as the automated lawn mower. So, another $1,500 dollar piece of equipment to install on the truck.

Then you need charging. What I can do with my current electric equipment is swap charged batteries with exhausted batteries. And swap batteries between other equipment as necessary. Battery swapping automation is another big expensive system, otherwise you have the things that self plug and charge. Which means you need at least an hour of downtime at least for every hour of cutting. This halves the rate one can cut grass with a single mower. Being reduced to 3 hours of mowing a day instead of, say, 6 hours, charged at $20 an hour, reduces daily income from $120 per day to $60 dollars per day. Very similar problems with gas refilling. So, that's another bit of expensive equipment, or your buying two or three automated lawnmowers to keep the throughput high. So, say another $3,000 dollars of equipment.

And of course the truck itself needs to be automated. We don't know how much extra this will cost per vehicle. It will probably be more. Lets say $5,000 dollars.

So, very basic attempts to automate the person out of the lawn mowing adds something like $10,000 in extra equipment. And the person really isn't automated out of the above really. We haven't covered washes, repairs, mapping out the area to be mowed, dealing with a stick or other such thing being left on the yard, getting the mower over the curb or up a stair, someone picking it up if it does break, and a whole bunch of other things. Automating the person out of it gets increasingly expensive.

And of course, if I am doing this work myself, I have no real reason to spend all this money vs just doing it myself and pocketing the $10,000 difference for myself. If I'm hiring someone full time anyways, it might make much more sense to keep the driver, and have him go house to house, drop off the electric mowers, and then pick them up once their done. Also allows us to work with people who want to pay in cash which needs a person there to pick up and count. So, pay an extra $1,000 to have automated mower, then save the rest on the worker. Maybe pay them an extra $2 an hour and pocket the $5,000 difference.

And remember this is all regards to just mowing the lawn, which is one small part of the overall job.

The question isn't if it can be automated, its if its cost efficient to do so. And, well, as stated my sense of recent automation is its being done to replace expensive labor, not cheap labor. Or, the automation will make sense in New York City, it will not make sense in Alabama. Or Brazil.