Why don't indie devs use AI-generated images as art?

[J1M]

That's not a list of tasks. If you list out the tasks it will become clear which parts of the job can be automated.

Since you seem to have trouble believing a lot of what people do in their jobs is just pattern recognition, I suggest you look up Github's Copilot.

 

[The_Sloth_Sleeps]

That's not a list of tasks. If you list out the tasks it will become clear which parts of the job can be automated.

Since you seem to have trouble believing a lot of what people do in their jobs is just pattern recognition, I suggest you look up Github's Copilot.

Ah so is this what you are thinking...?

 

[dbx]

People both under-estimate and over-estimate AI, mostly because they over-estimate both their own human creativity and the complexity of their job while under-estimating the complexity of "lower caste" jobs. And to a lesser extent because of hype.

For the art types I have bad news for you, but AI will replace you.
Yor job is not complex. In fact it's such an easy task that humans have been doing it for thousand of years, we learned to paint long before we learned to even farm. And after numerous millennia bilions of people have already almost exhausted all possible form of artistic expressions (dunno if it's the correct term, I have a hard time trying to translate that concept in english) and AI has access to all that data faster than any human can conceive.
AI will be more creative than most humans. It's not a matter of ifs but of when.

Also, you shouldn't take text-to-image systems as example of the creativity of machine learning as those are by definition guided systems.

 

[V17]

There is no threat to any human job because "AI" is as dumb as a bag of hammers and will always need humans to control it. The wheelbarrow has not replaced the labourer and nor will some dumb procedural art generated algorithms.

It will undoubtedly reduce the number of available jobs in some fields though. Self-driving road vehicles might be a long time away still for example, but you can already buy an autonomous tractor (since that is a more controlled environment, so the AI doesn't need to be very smart). It legally requires a person to check that it's not doing something stupid at all times, but that is done remotely with a tablet and it doesn't need to be one person per tractor. I read some time ago that there are some semi-autonomous harvesters as well, where there is a human driver normally driving the main one and two driverless vehicles following him around, couldn't find a source for that though.


On a different note, I read from some very concerned person on twitter that Stable Diffusion is biased and dangerous because if you use the monkey emoji as a prompt, it generates pictures of black children.
I tested it and was very disappointed to find out that this was in fact a lie, or only the case for the early testing versions. What am I supposed to laugh at now?

 

[The_Sloth_Sleeps]

There is no threat to any human job because "AI" is as dumb as a bag of hammers and will always need humans to control it. The wheelbarrow has not replaced the labourer and nor will some dumb procedural art generated algorithms.

It will undoubtedly reduce the number of available jobs in some fields though. Self-driving road vehicles might be a long time away still for example, but you can already buy an autonomous tractor (since that is a more controlled environment, so the AI doesn't need to be very smart). It legally requires a person to check that it's not doing something stupid at all times, but that is done remotely with a tablet and it doesn't need to be one person per tractor. I read some time ago that there are some semi-autonomous harvesters as well, where there is a human driver normally driving the main one and two driverless vehicles following him around, couldn't find a source for that though.

Lets stop calling automation "AI". "AI" is just human made algorithms, which keep getting written, and thus there is more "AI" than there was a year prior. "AI" is not improving, because it does not exist.

Next, automation has always been a factor in the job market. But it doesn't replace humans. Its incredibly costly to implement automation and humans tend to be cheaper and do a better job in the vast majority of work cases. Where jobs *can* be automated (factory/warehouse, etc) its almost certainly worth doing so.

As you pointed out there are more essentric attempts at automation but these perform rather poorly. I read a case study about automated vehicles i.e. trucks in mines, but these are not really worthwhile as you need a human safety operator, that replaces a driver. Plus you still need all those mechanics and tyre changers etc. Hardly cost effective.

I would be more concerned with immigration, digital currency, outsourcing, if its to do with jobs.

People both under-estimate and over-estimate AI, mostly because they over-estimate both their own human creativity and the complexity of their job while under-estimating the complexity of "lower caste" jobs. And to a lesser extent because of hype.

For the art types I have bad news for you, but AI will replace you.
Yor job is not complex. In fact it's such an easy task that humans have been doing it for thousand of years, we learned to paint long before we learned to even farm. And after numerous millennia bilions of people have already almost exhausted all possible form of artistic expressions (dunno if it's the correct term, I have a hard time trying to translate that concept in english) and AI has access to all that data faster than any human can conceive.
AI will be more creative than most humans. It's not a matter of ifs but of when.

Also, you shouldn't take text-to-image systems as example of the creativity of machine learning as those are by definition guided systems.

Plus there are the technology fetishists like Dexter who doubtless jerk off to this stuff in between booster vaxxines.

Automation only tends to work in very controlled environments. Its the edge cases that are the real stickler.

A great deal of jobs may be really simple, but they are surprisingly difficult, if possible to automate. Even the auto checkouts at supermarkets have failed as a concept. Case in point, at my local supermarket the customers do not prefer them, and are hardly used. They require 2 assistants to stand idle, manning the checkouts as the boop boop machines fail every 10th time to scan items correctly. Plus you have the additional security guy that has to have a presence there so people don't just walk through. Not so simple actually.

The automated carwash - probably the best example of successful "AI". It kind of washes your car so long as you don't hold so high an expectation.

Robotic vacuum cleaners, lawn mowers, also a flop. Yes they tend to do ok in ideal conditions & uniform geometries, but they aren't going to replace human gardeners/cleaners any time soon.

Its a pretty grand extrapolation that algorithms will replace even shitty artists any time soon.

 

[dbx]

Automation only tends to work in very controlled environments. Its the edge cases that are the real stickler.

This was true 5/6 years ago.
Since then, year after year, the "very controlled environment" requirement got increasingly less important (and it's something I have first hand experience as designing industrial automation systems is my daily job).
As for the rest I agree with you, cause is something I see everyday and that I have to explain to clients every time (industrial automation is expensive, the only cheap part is software, but that's like 5% of the job we do).
That said things have changed a lot just in the past 10 years alone, both the cost and the difficulty to automate some jobs have cratered, yet I believe that the remaining blue collar jobs will hardly be automated as it will probably never be cost effective, but for white collar jobs is way different. They were spared in the past because the tech wasn't there, which has led them to believe their jobs were too complex for automation, but it's not true, and they will be replaced faster and easier than in industrial settings as their jobs can be automated by pure software.

 

[The_Sloth_Sleeps]

Automation only tends to work in very controlled environments. Its the edge cases that are the real stickler.

This was true 5/6 years ago.
Since then, year after year, the "very controlled environment" requirement got increasingly less important (and it's something I have first hand experience as designing industrial automation systems is my daily job).
As for the rest I agree with you, cause is something I see everyday and that I have to explain to clients every time (industrial automation is expensive, the only cheap part is software, but that's like 5% of the job we do).
That said things have changed a lot just in the past 10 years alone, both the cost and the difficulty to automate some jobs have cratered, yet I believe that the remaining blue collar jobs will hardly be automated as it will probably never be cost effective, but for white collar jobs is way different. They were spared in the past because the tech wasn't there, which has led them to believe their jobs were too complex for automation, but it's not true, and they will be replaced faster and easier than in industrial settings as their jobs can be automated by pure software.

Specifically what can be less controlled? Having less control implies a less effective system. What jobs have cratered as a result? I can imagine pickers and packers for very very large warehouse operations, but you will always have medium/small warehouses that will need human pickers and packers.

I agree that most white collar jobs can be replaced, but they won't be.
Because we don't even need automation to do away with 40% of jobs. Most white collar jobs are busy work.
However they serve an economic purpose of injecting liquidity into the economy, and for that reason need to be present.
Just because something *can* be done doesn't mean its practical or cost effective to do so.

Example: just getting software A talk to software B, simpler to hire a team of system engineers to run the process themselves.
To be sure, a person in most cases is easier to control and manipulate than a machine or piece of software.

 

[dbx]

Automation only tends to work in very controlled environments. Its the edge cases that are the real stickler.

This was true 5/6 years ago.
Since then, year after year, the "very controlled environment" requirement got increasingly less important (and it's something I have first hand experience as designing industrial automation systems is my daily job).
As for the rest I agree with you, cause is something I see everyday and that I have to explain to clients every time (industrial automation is expensive, the only cheap part is software, but that's like 5% of the job we do).
That said things have changed a lot just in the past 10 years alone, both the cost and the difficulty to automate some jobs have cratered, yet I believe that the remaining blue collar jobs will hardly be automated as it will probably never be cost effective, but for white collar jobs is way different. They were spared in the past because the tech wasn't there, which has led them to believe their jobs were too complex for automation, but it's not true, and they will be replaced faster and easier than in industrial settings as their jobs can be automated by pure software.

Specifically what can be less controlled? Having less control implies a less effective system. What jobs have cratered as a result? I can imagine pickers and packers for very very large warehouse operations, but you will always have medium/small warehouses that will need human pickers and packers.

No, 'I'm talking about the working environment of the automated system.
Well, as an example in the industrial setting, years ago we used to design very precise production lines, in witch parts had to have fixed and precise positions, strictly controlled illumination, etc, which was very costly (reuse of sections between clients was non-existent).
Nowadays most machines just throw stuff on a belt and a robot controlled by computer vision + machine learning pick stuff and place it in the next section, this has reduced the cost to build production lines a lot (as sections can be reused between
clients, machines have more standardized parts and the time to design and build production lines got decimated. It's also cost effective for clients, now their production lines are not "lines" anymore and can be designed as islands that can be reused and be reconfigured between different productions.
This has simplified all aspects of industrial production to the point our products are now not limited to big factories and multinational corporations, we now sell stuff to medium and even some small sized companies.
A real example of something I have worked on was cleaning residual from metal casting of automotive parts (engine heads, connecting rods, etc ). A job that got automated away by computer vision + articulated robots (two articulated robots works in tandem, one pick up the part, the other one has interchangeable heads, first a 3D scanner to scan the part and then a couple of tools to mill away residuals).
This was a system whose first iteration (15 years ago) was accessible by car manufacturer only, last year I've seen similar systems being sold to small carpentry shops for the painting of home furniture parts
Quality control is another big request from clients, right now is mostly human based using specialized tools. It's the hardest part to automate, especially for SMEs that do many JIT productions, but we are working on it and solving one step at a time.

I agree that most white collar jobs can be replaced, but they won't be.
Because we don't even need automation to do away with 40% of jobs. Most white collar jobs are busy work.
However they serve an economic purpose of injecting liquidity into the economy, and for that reason need to be present.
Just because something *can* be done doesn't mean its practical or cost effective to do so.

Example: just getting software A talk to software B, simpler to hire a team of system engineers to run the process themselves.
To be sure, a person in most cases is easier to control and manipulate than a machine or piece of software.

Well, sorry but judging from the requests we get from the market, companies don't care about that and will automate as much as they can.

 

[The_Sloth_Sleeps]

Automation only tends to work in very controlled environments. Its the edge cases that are the real stickler.

This was true 5/6 years ago.
Since then, year after year, the "very controlled environment" requirement got increasingly less important (and it's something I have first hand experience as designing industrial automation systems is my daily job).
As for the rest I agree with you, cause is something I see everyday and that I have to explain to clients every time (industrial automation is expensive, the only cheap part is software, but that's like 5% of the job we do).
That said things have changed a lot just in the past 10 years alone, both the cost and the difficulty to automate some jobs have cratered, yet I believe that the remaining blue collar jobs will hardly be automated as it will probably never be cost effective, but for white collar jobs is way different. They were spared in the past because the tech wasn't there, which has led them to believe their jobs were too complex for automation, but it's not true, and they will be replaced faster and easier than in industrial settings as their jobs can be automated by pure software.

Specifically what can be less controlled? Having less control implies a less effective system. What jobs have cratered as a result? I can imagine pickers and packers for very very large warehouse operations, but you will always have medium/small warehouses that will need human pickers and packers.

No, 'I'm talking about the working environment of the automated system.
Well, as an example in the industrial setting, years ago we used to design very precise production lines, in witch parts had to have fixed and precise positions, strictly controlled illumination, etc, which was very costly (reuse of sections between clients was non-existent).
Nowadays most machines just throw stuff on a belt and a robot controlled by computer vision + machine learning pick stuff and place it in the next section, this has reduced the cost to build production lines a lot (as sections can be reused between
clients, machines have more standardized parts and the time to design and build production lines got decimated. It's also cost effective for clients, now their production lines are not "lines" anymore and can be designed as islands that can be reused and be reconfigured between different productions.
This has simplified all aspects of industrial production to the point our products are now not limited to big factories and multinational corporations, we now sell stuff to medium and even some small sized companies.
A real example of something I have worked on was cleaning residual from metal casting of automotive parts (engine heads, connecting rods, etc ). A job that got automated away by computer vision + articulated robots (two articulated robots works in tandem, one pick up the part, the other one has interchangeable heads, first a 3D scanner to scan the part and then a couple of tools to mill away residuals).
This was a system whose first iteration (15 years ago) was accessible by car manufacturer only, last year I've seen similar systems being sold to small carpentry shops for the painting of home furniture parts
Quality control is another big request from clients, right now is mostly human based using specialized tools. It's the hardest part to automate, especially for SMEs that do many JIT productions, but we are working on it and solving one step at a time.

I agree that most white collar jobs can be replaced, but they won't be.
Because we don't even need automation to do away with 40% of jobs. Most white collar jobs are busy work.
However they serve an economic purpose of injecting liquidity into the economy, and for that reason need to be present.
Just because something *can* be done doesn't mean its practical or cost effective to do so.

Example: just getting software A talk to software B, simpler to hire a team of system engineers to run the process themselves.
To be sure, a person in most cases is easier to control and manipulate than a machine or piece of software.

Well, sorry but judging from the requests we get from the market, companies don't care about that and will automate as much as they can.

Well, it may appear that way with a microscopic engineers outlook but whats actually happening is just re-shuffling.The same thing happens with outsourcing.

Large corps will fire 10k bods in a day after a restructure. But they'll re-hire again and grow and bloat. And the cycle repeats itself.

When I was younger "operational analysts" were hot stuff. Then a few years later the hype was over and the gravy train moved on. Those operational analysts still all have jobs, just different titles. And they are peddling some new bullshit instead.

Bullshit white collar jobs are here to stay and its not something that can be automated away, its the nature of the system.

True redundancy, I'm thinking more like VHS store. Can move onto DVDs, but then it goes online and its not so easy to transform the business.

 

[Davaris]

That's not a list of tasks. If you list out the tasks it will become clear which parts of the job can be automated.

Since you seem to have trouble believing a lot of what people do in their jobs is just pattern recognition, I suggest you look up Github's Copilot.

This is the pattern Github Copilot is learning from:

 

[The_Sloth_Sleeps]

That's not a list of tasks. If you list out the tasks it will become clear which parts of the job can be automated.

Since you seem to have trouble believing a lot of what people do in their jobs is just pattern recognition, I suggest you look up Github's Copilot.

This is the pattern Github Copilot is learning from:

Yah I've not actually tried it yet, but the idea seems...definitely its asking for trouble if you use that for suggestions for code.

Seems like VS2022 actually code assist tries to do something very similar, from what I can tell. Sometimes it gets very simple stuff right as I intended (getters and setters - which I won't use anyway), other times its just weird.

So seems like a glorified snippet tool, which I don't use much anyways.

 

[Viata]

Man, putting a SNES and saying "running every game at 60fps" is a good joke. Sadly, SNES had a lot of games with too much slowdown.

 

[J1M]

You have to look at this stuff differently. The question isn't if the machine can completely replace you, now. It's what percentage of what all the people in your profession do can be replaced?

Maybe today art director + tools can't replace the art department, but it can replace the full-time concept artists on payroll or at the very least push those jobs into contract/outsource roles.

In a few years it may speed up the creation of other art assets to the point where the art team can be smaller and editing AI generated results to make them production quality is the primary task.

Programming may be a little safer because non-programmers can't measure the quality of the work without giving it to users, but there's about to be a lot of unemployed JavaScript developers.

 

[The_Sloth_Sleeps]

You have to look at this stuff differently. The question isn't if the machine can completely replace you, now. It's what percentage of what all the people in your profession do can be replaced?

Maybe today art director + tools can't replace the art department, but it can replace the full-time concept artists on payroll or at the very least push those jobs into contract/outsource roles.

In a few years it may speed up the creation of other art assets to the point where the art team can be smaller and editing AI generated results to make them production quality is the primary task.

Programming may be a little safer because non-programmers can't measure the quality of the work without giving it to users, but there's about to be a lot of unemployed JavaScript developers.

Hrm, the way it will pan out will be far more underwhelming than you imagine. I'd say its more that these are just tools that some people like and some hate. It probably won't have much of an impact either ways. We may see more shovelware as a result. Art is always a bit of a tough & competitive field, and doesn't pay that well, nor are there many jobs in it.

The Copilot project is especially unimpressive. Looks gimmicky. Its functionally just a code snippet generator. You still need a code monkey to sit there and tell it what to do/not to do so I don't think any javascript programmers will be too threatened by it.

*If* you are ever replaced, you won't be replaced by some romantically fancy AI. No the reality will be much more mundane. If you get replaced it will be by Indians, immigrants or outsourced. You won't afford to work for $10 an hour. If you get replaced by a machine it will be a 10 line javascript bot that can trade crypto futures quicker than you.

 

[Dexter]

Can we get back to cool AI art or at least Twatter artist spicy existential butthurt instead of this "Sloth" faggot's boring and borderline retarded page-long personal rant butthurt? Here I'll help:

You gotta wonder what is worse: being replaced by an AI or an indian.

What about an AI made by an Indian? https://twitter.com/EMostaque

 

[J1M]

You have to look at this stuff differently. The question isn't if the machine can completely replace you, now. It's what percentage of what all the people in your profession do can be replaced?

Maybe today art director + tools can't replace the art department, but it can replace the full-time concept artists on payroll or at the very least push those jobs into contract/outsource roles.

In a few years it may speed up the creation of other art assets to the point where the art team can be smaller and editing AI generated results to make them production quality is the primary task.

Programming may be a little safer because non-programmers can't measure the quality of the work without giving it to users, but there's about to be a lot of unemployed JavaScript developers.

Hrm, the way it will pan out will be far more underwhelming than you imagine. I'd say its more that these are just tools that some people like and some hate. It probably won't have much of an impact either ways. We may see more shovelware as a result. Art is always a bit of a tough & competitive field, and doesn't pay that well, nor are there many jobs in it.

The Copilot project is especially unimpressive. Looks gimmicky. Its functionally just a code snippet generator. You still need a code monkey to sit there and tell it what to do/not to do so I don't think any javascript programmers will be too threatened by it.

*If* you are ever replaced, you won't be replaced by some romantically fancy AI. No the reality will be much more mundane. If you get replaced it will be by Indians, immigrants or outsourced. You won't afford to work for $10 an hour. If you get replaced by a machine it will be a 10 line javascript bot that can trade crypto futures quicker than you.

I completely disagree. There are already shovelware games populated with asset store artwork. These tools are going to increase the quality of shovelware art to the point where consumers will need to modify their hureistic for identifying shovelware.

If it was possible to replace Occidental programmers with Oriental ones, it would have already happened. People have been trying to do that for decades. Might be a concern a generation after a poorer country adopted English as a first language.

 

[Lutte]

There is no threat to any human job because "AI" is as dumb as a bag of hammers and will always need humans to control it. The wheelbarrow has not replaced the labourer and nor will some dumb procedural art generated algorithms.

It will undoubtedly reduce the number of available jobs in some fields though. Self-driving road vehicles might be a long time away still for example, but you can already buy an autonomous tractor (since that is a more controlled environment, so the AI doesn't need to be very smart). It legally requires a person to check that it's not doing something stupid at all times, but that is done remotely with a tablet and it doesn't need to be one person per tractor. I read some time ago that there are some semi-autonomous harvesters as well, where there is a human driver normally driving the main one and two driverless vehicles following him around, couldn't find a source for that though.


On a different note, I read from some very concerned person on twitter that Stable Diffusion is biased and dangerous because if you use the monkey emoji as a prompt, it generates pictures of black children.
I tested it and was very disappointed to find out that this was in fact a lie, or only the case for the early testing versions. What am I supposed to laugh at now?

Tractors on farms are far from being the only thing that are putting people out of jobs right now.






The future looks pretty bleak. You don't need to replace 100% of the humans either to have a deleterious effect. Just replacing most of them is bad enough.

Machines aren't going to replace your local plumber, mason or electrician anytime soon. But how many plumbers, masons and electricians are needed ? all the people being replaced right now and that will be replaced in the future won't be able to replace their jobs with one of the more irreplaceable ones.

It's only a matter of time at this point before all the burger flippers of the world get replaced too. You think they're the types of people you could give a bullshit white collar job to?

We have a growing mass of mostly incompetent humans who only ever were capable of doing things like moving boxes and flipping burgers and they're going to be out of anything to do in their life, forever.

 

[J1M]

Dexter





I'd say "learn to code", but we've already covered that...

 

[The_Sloth_Sleeps]

There is no threat to any human job because "AI" is as dumb as a bag of hammers and will always need humans to control it. The wheelbarrow has not replaced the labourer and nor will some dumb procedural art generated algorithms.

It will undoubtedly reduce the number of available jobs in some fields though. Self-driving road vehicles might be a long time away still for example, but you can already buy an autonomous tractor (since that is a more controlled environment, so the AI doesn't need to be very smart). It legally requires a person to check that it's not doing something stupid at all times, but that is done remotely with a tablet and it doesn't need to be one person per tractor. I read some time ago that there are some semi-autonomous harvesters as well, where there is a human driver normally driving the main one and two driverless vehicles following him around, couldn't find a source for that though.


On a different note, I read from some very concerned person on twitter that Stable Diffusion is biased and dangerous because if you use the monkey emoji as a prompt, it generates pictures of black children.
I tested it and was very disappointed to find out that this was in fact a lie, or only the case for the early testing versions. What am I supposed to laugh at now?

Tractors on farms are far from being the only thing that are putting people out of jobs right now.






The future looks pretty bleak. You don't need to replace 100% of the humans either to have a deleterious effect. Just replacing most of them is bad enough.

Machines aren't going to replace your local plumber, mason or electrician anytime soon. But how many plumbers, masons and electricians are needed ? all the people being replaced right now and that will be replaced in the future won't be able to replace their jobs with one of the more irreplaceable ones.

It's only a matter of time at this point before all the burger flippers of the world get replaced too. You think they're the types of people you could give a bullshit white collar job to?

We have a growing mass of mostly incompetent humans who only ever were capable of doing things like moving boxes and flipping burgers and they're going to be out of anything to do in their life, forever.

I do get your point, but this has been "progress" since the industrial revolution.
Anything that can be automated, will be. No doubt.
Automation could wipe out most available loading jobs in that port, for example. Then overpaid white collar types who service the port trucks will move into the area and drive up the prices.
Eventually workers would move on, and onto some other form of manual labour.

 

[Viata]

I find it funny when people thinking robots and AI replacing them is good because then "they can focus on what they want". It's like money will just disappear if there is only robot working in the factory.
If I have no job, there is no way I'll have money to pursue my dreams or buy things someone following their dreams did.

 

[J1M]

Am I talking to an AI? Nobody thinks truck maintenance is white collar and none of the loading jobs at a major port are manual labor today. It's all heavy equipment operators.

 

[The_Sloth_Sleeps]

Am I talking to an AI? Nobody thinks truck maintenance is white collar and none of the loading jobs at a major port are manual labor today. It's all heavy equipment operators.

It used to be men, before they were replaced by cranes.
The containers still get packed by hand, or forklift truck of course.

Maintaining the automated trucks/systems is sort of white/blue collar work.

 

[Galdred]

I tried generating some art for Zodiac Legion with Midjourney and Dall-e 2, but I found that the AI had big troubles painting unusual things (which makes sense since these would have much less available references to use in the algorithms).
I wanted to make some "star magic" scenes, or a battle scene with Zodiac heraldry, and zodiac themed helmets, but the AI seemed at a loss at what a crab/cancer helm should look like.

It also seems that it actively tries to ignore copyrighted material from the training (as when asking for 40K Space Marines, you get something that doesn't look like it at all from Dall-e 2, and something that looks like alternate totally not 40K knock offs from Midjourney).

I'll give it another shot, but I think it is better to illustrate "regular fantasy/medieval scenes" than unusual ones.
As for game assets, I don't think we are there yet at all, even though some have used it to generate tilesets.

 

[dbx]

There is no threat to any human job because "AI" is as dumb as a bag of hammers and will always need humans to control it. The wheelbarrow has not replaced the labourer and nor will some dumb procedural art generated algorithms.

It will undoubtedly reduce the number of available jobs in some fields though. Self-driving road vehicles might be a long time away still for example, but you can already buy an autonomous tractor (since that is a more controlled environment, so the AI doesn't need to be very smart). It legally requires a person to check that it's not doing something stupid at all times, but that is done remotely with a tablet and it doesn't need to be one person per tractor. I read some time ago that there are some semi-autonomous harvesters as well, where there is a human driver normally driving the main one and two driverless vehicles following him around, couldn't find a source for that though.


On a different note, I read from some very concerned person on twitter that Stable Diffusion is biased and dangerous because if you use the monkey emoji as a prompt, it generates pictures of black children.
I tested it and was very disappointed to find out that this was in fact a lie, or only the case for the early testing versions. What am I supposed to laugh at now?

Tractors on farms are far from being the only thing that are putting people out of jobs right now.






The future looks pretty bleak. You don't need to replace 100% of the humans either to have a deleterious effect. Just replacing most of them is bad enough.

Machines aren't going to replace your local plumber, mason or electrician anytime soon. But how many plumbers, masons and electricians are needed ? all the people being replaced right now and that will be replaced in the future won't be able to replace their jobs with one of the more irreplaceable ones.

It's only a matter of time at this point before all the burger flippers of the world get replaced too. You think they're the types of people you could give a bullshit white collar job to?

We have a growing mass of mostly incompetent humans who only ever were capable of doing things like moving boxes and flipping burgers and they're going to be out of anything to do in their life, forever.

Meh, it all depends on how much the economy will grow (or shrink) and how much demographic collapse we will have.
Past 15 years extreme economic growth (due to credit expansion) has mitigated things a lot. Collapse of the workforce (which will likely decrease both unenployment and working population) will probably balanced things out for the next automation revolution.

 

[Machocruz]

Art is a broad field. Concept art will definitely be ruled by the machines, judging by what AI is already capable of, which is on par with a lot of work in that field. On the other hand, things like doing a mural in a rich person's home requires physical labor, so you'll need the AI to not only get to the level of human creative power and technical craft, then it will need to be operating in a dexterous machine.

I've been saying that artists (meaning people who like to create images, not who want put their fee fees out there) to wear many hats for a long time. Not just painting, not just character design, not just still images, etc. Traditional, digital, photography, filmmaking, graphic design, 3D printing, home decoration, etc. The days where you could just be an oil paint guy or a pen and ink guy or an airbrush guy or just do book covers or just editorial cartoons have been over for a long time. Yeah you won't become Michelangelo of one or two things, but it beats crying about the lack of opportunities out there because you were too invested in one single direction

I can't speak on other industries (although the notion of " job security" in any field seems like a luxury to me when viewed against the sum of human history. Life itself isn't secure). The ogliarchs don't want the common people working or having any kind of self sufficiency, they want everyone who isn't in their echelon to be replaced with machines and either put on their BUI system or fucking off and dying. The ensuing civil disobedience will be taken care of by the Boston Dynamics hunter killer robots, biowarfare, HAARP technology, etc.