/lit/ - Literature

Nail maker wages dropped with the creation of the first nail factories, and former respected tradesmen fell into the status of low-paid factory workers

A Japanese Marxist accidentally disproved the Tendency for the Rate of Profit to fall. Oops.
https://en.wikipedia.org/wiki/Nobuo_Okishio

>>22611920
this only works with the most basic products.

>make music
>it's okay
>use a DAW
>workflow is much easier now
>music is better
>better output
>higher profit

>>22611906
How would automation not lower the value of an individuals labor? How would that make profits fall though? Wouldn’t profits rise as labor costs per unit fell?

>>22611943
>higher profit
post your bandcamp bro
lets see those profits

>>22611906
Because in the long run it’s cheaper to buy a machine than pay humans

If you can automate what would otherwise require labor, then obviously that labor is not as valuable anymore. What is not clear about this?

Economics is a pseudo-science anyway (t. MS in Economics). The great error of Marx was analyzing it. You’ll note that everything Marx is remembered for has little do with economics but are rather mandated by what he considered good ethics.

>>22611983
thats not true
machines are expensive
impoverished children will work for food

>>22611906
>tfw you discover marginal utility

>>22611983
companies are currently finding out that this isn't true at all.
It all depends on the complexity of work and whether maintenance is standardized enough to be largely automated as well (spoiler: it isn't - source: get a robot to fix your sink; this is something "simple", yet it's nearly impossible to automate since there are like a million standards thanks to product variety)

>>22612007
machines are 20x more expensive and have 200x the work output
>>22611906
marxism stopped being cool 100 years ago

Shitalism sucks.

>>22611976
It's also true in layout and design.

>>22612045
I don't think Marx was anticipating robots. The industrial technology that he would have been aware of automated away so many cottage industries that the term cottage industries became antiquated.

>>22612053
>dilates audibly

>>22612055
Marx imagined that at some point in late stage capitalism there will be nothing but a few megalogigafactories on the planet, employing millions of professionals each, and he brought the analogy of big factory hammers replacing many small hand hammers.

>>22612052
>20x more expensive
automation machines like the commonly depicted arms are usually leased for the same cost of renting a nice apartment and still requires high skilled labor to maintain and run the software
impoverished children cost like 10 dollars a week
refugees are even cheaper in some instances and can do complex tasks (see: refugees training openAI)
there are few places where automation is cheaper than humans and they only exist in high skill labor markets or the west where human rights are acknowledged

its not the 1970's anymore. nobody takes marx seriously anymore.

>>22612097
Not at all, you have no fucking idea the profits factories or even small businesses make. You can buy machines for 20-40k for a small-time business, that's the expenses of a single employee for what, 2-4 years? The machine can rake in profit several times greater
In factories it's even greater, and they also get a hefty discount on the stuff they buy. Doesn't require any specific maintenance, there are engineers for that and the software is also easy.
>impoverished children cost like 10 dollars a week
And are incapable of producing the same output. They are simply not worth it

>there are few places where automation is cheaper than humans
Depends on the product. If the product can be automated partially or wholly, it always is because it's worth it

>>22612007
>>22612045
That's assuming the cost of new machinery doesn't decrease and its capabilities remain identical. Hint for professional retards (economists): technology improves.

>>22612114
or the same level of specialization. child labor is good for stuff like growing lettuces, or operating easy machines (=partial automation)

>>22612097
>can do complex tasks (see: refugees training openAI)
Lol do you know what that means? It saying how many dogs are in a picture and shit like that. Tracing a person's face with a mouse. Literal zero skill task.
>there are few places where automation is cheaper than humans and they only exist in high skill labor markets or the west where human rights are acknowledged
Bizarrely out of touch statement. Textiles, tools, cookery, even screws and nails have all been nearly totally automated since the 1800s. I can't even guess the percentage of jobs that would have existed then but don't exist now due to automation and yet we still use those same products. Automation is vastly cheaper than human labor in the majority of industries.

>>22612114
This is a lefty faggot pretending he knows how a machine shop work. You will never be a proletariat you petit bourgeios diletantte.

>>22612124
Retard, I've worked for businesses that had this and also saw firsthand such machinery used in factories. I'm all for it btw
Do you think businesses buy machines because they like to lose money?

>>22612114
Its clear you live a sheltered life and can't see past your own circumstances.

>>22612137
You claimed businesses drop 20-40k on machinery. You are a middle manager that is only trusted with enslaving actual workers. At least the wagies have an excuse for being agressively stupid.

>>22612121
let me know when asia and india automates its slave labor lol

Is it faster to cut down a tree with a chainsaw or an ax? A chainsaw. Now you’ll say: but a chainsaw costs more. Let’s saw a good chainsaw costs 1000 dollars and a good ax 100. Let’s say each tree you cut down you can sell for 50 dollars. If you cut down a tree with a chainsaw 3X faster than the rate you cut down a tree with an ax and in a day of cutting trees with an ax you can fell 2 trees than you can fell 6 with a chainsaw. You make 100 dollars a day with an ax and 300 with a chainsaw. You have the chainsaw paid off after 4 days of work with the chainsaw but 10 with an ax. After some time you hire 9 men to run chainsaws with you and make 3000 dollars a day. Whereas if you were to hire 9 men to run axes with you you would be making 1000 dollars a day. Technological innovation is now causing 2/3rds of the profit of the chainsaw company whereas pure human labor is only contributing a third. The axe company would be better off switching to chainsaws than paying 20 more workers. Let’s assume all workers are paid 50 dollars a day. The chainsaw company pays $450 a day for $3000 in sales for a net $2550. Were the ax company to hire 20 more workers for a sum of 29 they would pay $1450 a day for $3000 in sales. Arguably the ax workers would have to be paid more for the intensity of their labor given that chainsaw companies competing on the free market can offer better working conditions. This is literally basic economics 101. Even right wingers wouldn’t argue with Marx on this point

>>22612146
I was a wagie in the business that did, there was no manager actually
Many businesses will simply buy machinery to start, and will maintain it themselves by reading the manual. Printing is a good example of this. The machines are cheap and do virtually all the work if you feed it the .xps file
Bigger businesses spend much more on automation, and also save much more many while making more greater profit
>>22612145
Humans can simply not compete with machine output. Humans are generally PARTIALLY better at manual field labor which I have also done (the only machine there was the car). Which are also the jobs predicted to not go extinct
all else will be automated inshallah. useless fags who can only do office work will starve and die alhamdulilah total middle class holocaust

>>22612154
>useless fags who can only do office work will starve and die alhamdulilah total middle class holocaust
by this I mean sois and women btw

>>22612151
One of the complaints third world countries have is that manufactured goods are dumped on their markets at prices far cheaper than what local labor can compete with and it destroys the local economy. There is no non-automated cloth industry in India

Can we all agree that mass automation leading to millions dying is good?

This thread is cold garbage with no point or value.
>t. engineer

>>22612115
>technology improves.
That's part of the problem though. Product variety changes through time as well. Standards get replaced, interfaces become obsolete etc
To stick with the example of the sink: The problem with automation in the maintenance sector is that today your sink may be using part A by company X. But company X may go through a bad fiscal year and replace part A with a cheaper part B, which, from now on is built into the entire product line. Any standardized "sink repair robot" must now be aware of all these changes throughout the entire market, which is an impossible maintenance task in itself.
Automated systems have enormous trouble with small changes in complex products and often reconfiguration is simply impossible or extremely expensive. During covid the automated car production in Germany came to a halt because the shipment of a tiny part from china was delayed.

Sure lots of manual labor can be automated, but the machinery is facing conceptual problems on the level of standardization which require creative solutions (for example: ordering an unlicensed part A from china - which is definitely not in the range of automated solutions, especially not on a market scale). Unless product variety shrinks down to a minimum maintenance cannot be standardized.

>>22612194
But automation is already being maintained and steadily growing for the past century or so.
What are you even on about lol

>>22612194
>Unless product variety shrinks down to a minimum maintenance cannot be standardized.
One of the big deals about AI is essentially bespoke engineering on demand. A MacGyver in a box. Not to mention you're focusing on maintenance when the sink and this great variety of parts you're talking about is definitely produced by automation.

>>22612160
>There is no non-automated cloth industry in India
garments are part of that industry and is absolutely not automated

>>22612213
People wear alot of hand stitched clothes in India? I had no idea. All the picture I see are t-shirts and jeans. At least you've acknowledged automation is extremely cheaper than humans in making cloth. No matter how desperate people are willing to work at any wage.

>>22612220
if automation is such a magic bullet why cant it automate garment making?

>>22612252
We do automate garment working. That was what the snark about hand-stitched clothes is about. A sewing machine is automation. When Marx was writing he would have been witness to a large reduction in seamstresses due to the automation provided by a sewing machine. Think of it as the chainsaw of garment making.

>>22612252
if automation is such a magic bullet why cant it produce 10/10 sexbots yet? there, automation is a fad. proven

>>22612252
You don’t need to automate the process of raw material to finished product in order for labor to being automated. If I do 90% of the labor to make a t-shirt, but a machine does the last 10%, that last 10% is automation and it supplanted the labor that would’ve otherwise been necessary to make the shirt.

>>22612198
see
>>22612045
>It all depends on the complexity of work and whether maintenance is standardized enough to be largely automated as well

To claim that everything can just be automated away is a naive point of view since automatization largely depends on standardization.
Any field that has constantly changing standards (like those with high product variety) will, at the very least in their maintenance sector, never see proper automatization, because the overhead of reconfiguration is usually not worth it.

I used the sink example, but you can also see this in software with code generation.
This was the big hype like ten years ago. Then people, especially the good guys over at IBM, realized that code generation for Java is kinda joyless if Oracle plans to push out a new version every year, constantly deprecating their generated architectures.

Automated maintenance of generated code architectures is a matter of model transformation and a huge pain in the ass. It's simply not worth it.

In fact, maintaining legacy code from old programming languages is its own lucrative business model. If you know Fortran well enough you can earn a fortune, because every old banking system is using it.
>>22612211
>One of the big deals about AI is essentially bespoke engineering on demand. A MacGyver in a box.

lol yeah nah, the big problem with AI goes in pretty much the same direction. It's a matter of creative solution finding. The intricacy of engineering problems requires equally intricate inputs for automated solutions. Or, in other words: No matter how great the AI is, the code is always just as good as the problem description that is fed to it.
That is the real bottleneck of AI.
It cannot really maintain or automate anything unless you tell it very strictly what to do. At that point you can just do it yourself.
Perhaps you can scale bigger projects better with AI, but even that's questionable.

>t. software engineer trying to get out of work by giving his work to AI

>>22612284
>Or, in other words: No matter how great the AI is, the code is always just as good as the problem description that is fed to it.
So just feed it all the information you would have fed to the guy doing the maintenance. If he could of done it the AI can too.
>It cannot really maintain or automate anything unless you tell it very strictly what to do
Why not? This is getting into what the eventual capabilities of AI will be.

>>22612100
a majority of social scientists are marxists or heavily influenced by marx

>>22612284
>realized that code generation for Java is kinda joyless
You realize Java is interpreted on a VM right? You literally can't run Java programs without generating Java bytecode. If you're not writing assembly you're relying on code generation to make your program run

>>22611906
It's actually the exact opposite. Technology is one of the main drivers behind increased productivity increasing marginal value product of labor allowing for wages to rise as well (of course this does depend on workers actually getting a payrise which depends more on government/unions)

>>22612185
You are the ugly office closeted fag that gets called in to justify accountings massively inflated bills payable.

>>22612290
>Why not? This is getting into what the eventual capabilities of AI will be.

Because the AI doesn't have a crystal ball, no matter how advanced it is. It cannot guess your intent. You have to give it clear step-by-step instructions if you want proper specialized solutions.
But the thing in programming is, that once the intent of a piece of code is clearly defined, the rest is mostly just typing it out (ignoring side-effect analysis for now).

The difference between the maintenance guy and the AI is that I tell the maintenance guy that I want A and B and he automatically knows implications C and D from that. Either through experience, education or just from having worked with that piece of code for a long period of time. A generalized AI is conceptually not able to do that, because implications C and D may be specialized conditions within the company and not available to whatever training model it used.

To give a more illustrative description: The human mind and creative solution finders can broaden the solution space in a specialized way from a punctual input, whereas the AI must, by definition, always stay within the narrow spectrum of my punctual problem description.

Of course the maintenance guy can still fuck up, but unlike the AI he is not a black box. You can ask him "What the fuck were you thinking here?". If you ask the same question to an AI (if it could explain its "thinking") its answer would be "I narrowed it down from similar examples" - and that fucks up the whole workflow. Forget about debugging that.

And as I said: All of this ignores side-effect analysis.

>>22612312
>You realize Java is interpreted on a VM right? You literally can't run Java programs without generating Java bytecode. If you're not writing assembly you're relying on code generation to make your program run

Hey, nicely read off that disambiguation page.
However, we are not talking about compilation.
We are talking about source code generation, aka automatic programming.

>>22612370
>A generalized AI is conceptually not able to do that, because implications C and D may be specialized conditions within the company and not available to whatever training model it used.
So just train the AI better. Do you think generalized AI will be some monolith that won't vary over companies?

>Hey, nicely read off that disambiguation page.
I have a BS in compsci and have taken a compiler course. Talking about shifting standards making code generation impossible is fucking dumb.

>>22612370
>whereas the AI must, by definition, always stay within the narrow spectrum of my punctual problem description
By what definition? Something you just made up? One of the funnier things about the proto-AIs we have now is their propensity just to make shit up and lie through their teeth. Like you.

>>22612384
>Do you think generalized AI will be some monolith that won't vary over companies?
He thinks it's a linearly programmed thing like he's used to from college: a big if-then tree. He doesn't get how neural networks work.

>>22612305
Social science isn’t real, either.

>>22612384
>Do you think generalized AI will be some monolith that won't vary over companies?
It is questionable if your problem field can provide the necessary training set.
Especially if you only have one product. Or if you work on safety-critical software.
Special conditions are usually the result of interfacing engineering problems as well.
The problems that companies like Boeing are facing in their software department are not problems of software engineering but of aerodynamics (among others). An aerodynamics engineer works that through with them.
The creation and maintenance (!) of a knowledge base that could be used to train an AI in such interface fields is something we cannot even fathom at the moment because we cannot even properly formalize the interfaces between the fields. With evolving standards and evolving technology, it's questionable whether it will ever work. IBM can't even keep their fucking meta-models straight. That's why you have to bring in the experts.

I know this for a very specific reason by the way.

>I have a BS in compsci and have taken a compiler course.

Very cool
I took that one too
ten years ago.
That was a couple years before i wrote a shitty PhD thesis on software engineering
Which is exactly the reason why I know about interfacing engineering fields and why I am telling you this now.

And again:
source code generation has nothing to do with compilation.

>>22612397
>By what definition? Something you just made up?
By the definition of its functionality and its usability within a software engineering workflow.

Because otherwise, as you say, it will give you a bullshit answer.

The bullshit answers are not results of a "broadened solution space" based on the implications read from your input. Instead, they are basically just white noise.

>make shit up and lie through their teeth. Like you.
Fuck, ya got me anon.

If these answers are not satisfying, oh well.

>>22612503
>The creation and maintenance (!) of a knowledge base that could be used to train an AI in such interface fields is something we cannot even fathom at the moment because we cannot even properly formalize the interfaces between the fields
It's literally the same documents the engineers use to communicate between each other. Feed that into the AI. Their is no need to formalize anything that's the promise of AI, especially one that can understand human language.
>PhD thesis on software engineering
Software engineering is up there with HCI in terms of bullshit. It indicates zero special knowledge of machine learning or AI
>source code generation has nothing to do with compilation.
And you shifted to this after being called out for saying Java was proof code generation doesn't work. Which is a retarded statement.
>workflow
Anytime anyone starts talking about "workflow" I tune the fuck out since it's garbage business speak

>>22612538
ignore that moron who paid for a doctorate in SE instead of CS. he unironically thinks dataset availability is the issue and does not understand machine learning at all based on this post >>22612370.

>>22611920
Ok but what does this have to do with rates of profit.

>>22611906
Automation is inherently commodifying, but it isn't the principal cause of the increasing commodification of labor.

>>22612007
I see you are a Victoria 3 player.

>>22611906
Im not even a Marxist but this critique is literally dealt with in the first 5 pages of Contribution to the Critique of Political Economy. It is that basic. Like ffs OP, at least make an attempt

I want money for free lol

>>22611930
That article is so poorly written lol