[ / / / / / / / / / ] [ dir / asmr / egy / hrvatska / hwndu / kc / newbrit / polk / wai ]

# /leftypol/ - Leftist Politically Incorrect

A collective of people engaged in pretty much what the name suggests

Catalog

Name Email Select/drop/paste files here * = required field [▶ Show post options & limits]Confused? See the FAQ.
Flag
Options
 dice sides modifier
Password (For file and post deletion.)

Allowed file types:jpg, jpeg, gif, png, webm, mp4, pdf
Max filesize is 12 MB.
Max image dimensions are 10000 x 10000.
You may upload 5 per post.

File: 957c5cd327a7d41⋯.jpg (313.58 KB, 1148x813, 1148:813, da zdravstvuiet stallman!.jpg)

No.1477348

It's recently come to my attention recently that there are people here interested in the attempts of various computer scientists to introduce cybernetic computerized planning (a model acceptable to most Marxists and anarchists) to the USSR, to improve efficiency and put control in the hands of the workers themselves. We can learn a lot from these pioneers of computerized socialism, seeing as our own society is basically run on the basis of computerized markets - it wouldn't be too big of a leap for most people to imagine democratic planning via computer networks. Besides, even FALC will need some basis for resource allocation.

The problem for Western readers, however, is simply the scarcity of the material in English. Slava Gerovich has published most articles on the matter, but beyond him, details are scarce.

With Google's recent implementation of deep-learning techniques in its translation service between Russian and English, however, it's now feasible to translate whole articles for the most part.

My actual ability to speak Russian is fairly limited in terms of vocabulary, but I've studied a lot of the grammar and verb prepositions. If there's a sentence which you think isn't translating right, I can try to figure it out for you. No guarantees, though.

Here's a good article from Slava, let's start digging up the sources and posting links to the original Russian ones if possible: http://web.mit.edu/slava/homepage/articles/Gerovitch-InterNyet.pdf

No.1477370

http://www.au.af.mil/au/afri/aspj/airchronicles/aureview/1967/mar-apr/sleeper.html

We should consider starting a project on this, this is how we take on the world guys.

No.1477404

Here's the first Russian source cited by Gerovich:

Berg, Aksel’. ‘Ekonomicheskaia kibernetika: vchera i segodnia.’ Voprosy ekonomiki no. 12 (1967)

Its title translates to "economic cybernetics: yesterday and today", from something called "Problems Of Economics"

I found their site, but the archive only goes back to 1992: http://www.vopreco.ru/eng/index.html

I looked up Aksel Berg, and did find this article: http://www.unz.org/Pub/AngloSovietJ-1964q1-00013

It's not the one that I wanted, but still useful as a perspective of someone relevant.

I also found this list of sources from a declassified CIA document, not sure how useful it is: http://www.dtic.mil/dtic/tr/fulltext/u2/831939.pdf

What we need are technical details, what the planners specifically thought were relevant problems. I want to see what I can code in Python.

Paul Cockshott's linear optimization of input-output tables is a fine core, what we need are peripheral programs to feed it the right info.

No.1477886

I'm interested, it's really amazing how much capitalism limits people's imaginations. So even if it's really old, there could be a lot of unusual or radical theory in this Soviet cybernetics.

No.1477917

File: c5e4efba3436920⋯.pdf (7.99 MB, [Oskar_Lange_(Auth.)]_Intr….pdf)

File: bcde318d8704cee⋯.jpg (62.11 KB, 850x400, 17:8, quote-when-education-and-r….jpg)

Ever since I stumbled up on the topic of cybernetics, it felt like

my education has cheated me from getting to know this multidisciplinary science.

So I presume that anyone interested has read introductory books on cybernetics.

Ross W. Ashby or Norbert Wiener, Stafford Beer and many others. Oskar Lange was a pioneer of applying cybernetics on the problems of economics.

Regarding specific parts, I have found Czech scientific journal, that has most of the articles in english. And also special issues detailing some specific problems of cybernetics theory are also written in english.

http://www.kybernetika.cz/content.html

A study group would be good. Also I could help with the translations from russian to english.

>>1477886

Every introductory book on topics of cybernetics written by an author living in Eastern Bloc has an introductory chapter saying how humans have been using labour to secure a living to themselves, but their information processing abilities were sufficient even in the era of the industrial revolution.

Essentially a historical materialistic view on how the topic of information processing and information itself came to be investigated by humans.

Also they extend the materialistic notion of matter in that all matter also has an ordering in space, whether it is random or ordered, and thus it has information. And all information is bound to matter or energy.

So they say that information and matter exist together.

No.1478212

>>1477917

What would be the best on-line tool to create a study group?

No.1478215

And for the programming we can create a Github repository.

No.1478223

Cool shit man

Will be checking it out.

t. CS student

No.1478233

So I would be up for using discord for study group, or a github (although I would need to make another git account to prevent doxing).

But I know lots of security fags here are paranoid about le CIA, even though you can't outrun them no matter how many measures you take, and refuse to use them.

So you tell me what would be the best tool to use.

Github allows for colaborative programming and writing simple pieces, and discord can be used to discuss while also saving the discourse for anyone who isnt online at that time. IRC isnt that great in that way, since it doesnt save what is typed, so if youre offline, tough luck.

No.1478288

>>1478215

>>1478233

Why not something like GitLab or Gogs? I mean, the only valuable thing about github is the social framework it offers, but I don't assume that that's what we're in for.

No.1478297

Stupid question, but could we help the Kurds with this as a real world application?

No.1478301

>>1478288

We could do that.

>>1478297

Kurdistan is a fucking warzone, they dont have good information infrastructure and arent really in the mood or position to build it.

No.1478460

File: 2640627c0955945⋯.pdf (7 MB, Introduction to economic c….pdf)

>>1477917

Cropped the book for easier use on ereaders

No.1478479

>>1478301

Surely there's a smartphone per commune, smartphones are absolutely fucking everywhere.

No.1478481

>>1478479

A smarthpone really isnt enough to centrally plan an economy. You need to have all production connected as well.

But we could try.

No.1478484

I'm very afraid about socio cybernetics so I'm not sure how I should feel about cybernetics in the context of the economy .

Cybernetics is a powerfull tool to create order in society. This also makes it easy to abuse it.

If you establish a chatroom or a mailing list to discuss this topic further and maybe to do some planning, please post about it here. I would be very interested in it .

No.1478485

>>1477886

> I'm interested, it's really amazing how much capitalism limits people's imaginations. So even if it's really old, there could be a lot of unusual or radical theory in this Soviet cybernetics.

One of weirdest things I've seen so far (second only to theremin) was Soviet hydraulic computer from before WWII.

http://www.digitaljournal.com/article/338106

No.1478500

>>1478233

>discord

we can create board here on 8ch with similar functionality, otherwise I'm for >>1478288

No.1478504

>>1478500

The problem with an 8chan board would be that it can be raided and shit, its not invite only.

Anyway, im down for this. I think this is one of the first things on this board that really grabbed my attention and made me want to study the theory behind it and shit, in 3 years.

No.1478505

>>1478481

Well it would function like this with a computer anyway from my limited understanding about the subject. Someone fills in a printout on a clipboard, then everyone hands in their results to an office, which then enters it into the server using a computer. A smartphone would be more than capable of that, and the more smartphones a commune has the less people would be using clipboards. Smartphones are incredibly abundant in the third world though because they get all the older models we had for cheap second hand.

Although we would introduce a problem of potential economic sabotage by having it stored on the internet, and this will likely be unavoidable since I imagine Kurdistan is serviced with cell towers rather than fixed lines, which would cost them $200/km just for cabling. Russia might be able to give them a hefty discount though. Just spitballing ideas though, obviously we would need to get it made first and then send the idea off to our YPG contacts to see if any of the councils are interested rather than asking them if they're interested with nothing to show for it and then potentially not getting it done. >>1478485 Fuck! No.1478509 >>1478505 I think while cybernetics for kurdistan would be cool, we shouldn't focus our small study group on that alone. No.1478513 >>1478509 Well they may not accept it like another poster said, but if you get to the stage you have a proof of concept what I'm trying to say is that the level of Kurdistan infrastructure will match the level we produce it to right off the bat. No.1478515 >>1478504 >8chan board would be that it can be raided and shit, its not invite only. small boards don't get raided. look at /marx/, even with triggering name like this they are doing ok + occasional recruitment of /tech/ autists would be much easier. No.1478518 >>1478513 Call me a negative nancy but I severely doubt the ability of 7-10 odd programmers on a chan to be able to produce such a project succesfully. But I am open to trying it. No.1478519 >>1478515 Fair enough. No.1478531 >>1478515 > small boards don't get raided. look at /marx/, even with triggering name like this they are doing ok + occasional recruitment of /tech/ autists would be much easier. Actually, /marx/ got raided not too long ago - this wiped practically all old threads (except for the stickied ones). No.1478537 >>1478531 >sticky all the threads >add max numbers of pages >create huge post limit >disable deleting old posts at post limit >back up everything No.1478539 >>1478537 At this point you might as well look for a more suitable framework. No.1478543 >>1478539 The point of changing five options in the board settings? No.1478547 >>1478531 I know, I lurk it occasionally, this is not really problem. Of course there are autists who would want to raid le cultural /marx/, but cybernetics is not so triggering simple /pol/ minds, it would be closer to /freedu/. My main point is that if we go full sekrit club we can't have lurkers and we need some means for (very slow) growth, so number of users can at least stagnate, otherwise it is doomed to fail since most people will quit. No.1478550 >>1478547 So, name it something incredibly boring? Would a board/github combo be ideal? No.1478554 >>1478550 >So, name it something incredibly boring? Anything which is not associated with left in american pop culture. tbh even /cybernet/ would work. >Would a board/github combo be ideal? I believe so. No.1478557 >>1478550 Ideally : /leftypol/ + discord + repository No.1478561 >>1478557 >discord what is so great about discord apart from satisfying redditors need of massive tripfag circlejerk? No.1478562 >>1478539 Mailing list, IRC and cgit? Public, but invite only. Most people wouldn't even bother reading through such sites, if it doesn't have 5MB css frameworks and every single html5/tag, so there would be some protection though elitism. But at the same time, anyone could use the technology to cooperate. If the linux kernel can work with this kind of stuff, so can we, tbh. No.1478567 >>1478561 le emojis! 👌😂💯💯💯 No.1478571 best thread on leftypol by a wide margin. Currently reading the PDF, will read Paul Cockshott after. >>1478561 What about Riot? No.1478575 >>1478562 Sounds good. No.1478576 >>1478571 why not some plain old irc? shit just works No.1478578 >>1478567 yeah bruv 👌👌👌👌👌👌😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯💯 No.1478579 >>1478562 >Mailing list I don't think you can convince people here to give somebody their email addresses + who would manage it? but it sounds good apart from this. >>1478571 do you mean riot.im? I don't know it, but it looks interesting. No.1478581 >>1478562 Wait is cgit a host or just an api? Would we need to host our own server? No.1478585 >>1478579 One could either use something like cock.li to create anonymous addresses, or, since I don't trust VC, one could run a mail server on a server. Downside would be the required maintenance. >>1478581 >Wait is cgit a host or just an api? It's a cgi program, that displays local git repositories hosted on a server. Suckless have their own here: http://git.suckless.org/ >Would we need to host our own server? Yeah, but I'd suppose a$5 could server (eg. DigitalOcean, Linode) would do.

No.1478590

>>1478576

IRC doesn't save conversation that take place when you're logged off.

>>1478579

I like it. It runs on matrix, wich is free software, while also being easy and shiny. It's also not locked down to third party clients, unlike discord.

No.1478599

>>1478590

>IRC doesn't save conversation that take place when you're logged off.

Server side logs?

No.1478613

>>1478599

If you're he'll bend on hosting everything yourself, sure

No.1478626

File: 6b6e0d926bf5031⋯.pdf (2.88 MB, how-to-not-network-nation.pdf)

This PDF might also help.

>>1478590

IRC shouldn't save the log publicly, it's instant and linear messengering. If more important stuff is being discussed or something is being announces, it better to use something like a mailing list, with parallel conversations and threads. I personally always hate having to read through chat logs to see if anything important happened while I was off, and then not being able reply to a message without interrupting the current conversation.

>>1478613

???

No.1478636

>>1478626

I think he's trying to say if I'm hell bent on storing a text file myself, I can do so at my own peril. Just put it on the server you goofs.

No.1478647

>>1478561

>>1478571

M A T R I X is the best option

No.1478653

>>1478636

>>1478626

No I just don't get why you are so hell bend on hosting a server yourself when you can use free publicly available alternatives. Just setup abother email if you don't want o be doxed, use one of the many irc severs there are and just use github or some other existing git site. I don't see why yu would want to host a server.

No.1478654

>>1478626

There's a matrix-irc-bridge not unlike XMPP transports which you can use with any IRC client. Also you don't have to read chatlogs, but you can if you wish.

I agree with you in principle, but there's a reason why most people prefer accessible, shiny interfaces like Discord and Slack and so on.

No.1478668

>>1478647

>>1478626

>>1478613

>>1478599

>>1478590

>>1478585

>>1478581

>>1478576

>>1478571

>>1478562

>>1478561

>>1478557

>>1478554

>>1478550

>>1478288

>>1478233

>>1478215

Ultimately, it's mostly irrelevant what we use. Shouldn't this thread be used to discuss theory and cybernetics? I mean, we can use IRC and matrix, if we want to. If it becomes more serious, we can switch to more serious solutions.

>>1478653

Control? But as I already mentioned, we shouldn't focus on such meta-issues, at least for now.

No.1478702

I'm not knowledgeable about this at all, just dropping by to say this looks really cool and that you should check out decentralized web technologies (saw this a while ago and figured it might come in handy sometime) it figures you will have to look into this kind of thing if you intend for your network to sustain an attack by Porky.

https://blog.ethereum.org/2014/08/18/building-decentralized-web/

(warning: this one page is heavy loaded with videos so it will take a while to load)

No.1478710

>>1478668

>not endlessly arguing over the chat client until interest tapers off

Get a load of this guy.

No.1478734

>>1478626

>How not to Network a Nation by Benjamin Peters

I've read 80 % of that so far and unless there is a miracle at the end, the book is shit. S H I T. Here is my review of the first 80 %.

If you want to know anything about economic planning, keep on searching, because this "work" does not describe a single algorithm nor a single criterion for judging algorithms related to planning. (It only mentions Pareto efficiency in passing without giving a definition.) You don't need to be a programming wizard to figure out the author is weak when it comes to computers and math because of just how weak he is. Being an interested lay person should be enough to see right through the blahblahblah.

Look at this waffling:

>Because computational methods do scale, the economic cyberneticists enthused that maybe the principal question for economic reform (who should control the command economy and how?) might be resolved without either the price of politics of the politics of price. It might, the cyberneticists reasoned, be solved with computers.

This looks a bit strange because of a typo: "either the price of politics or the politics of price". So, with the typo fixed, let's read it again. Do you feel you know more now? "Because computational methods do scale…" Whether some method scales well or not certainly depends on the method. You can't say something scales well just because some small-scale version happens to work on a computer. Who writes like that? Somebody who doesn't know what he is talking about. It looks like the entire book is like that. Do you want more waffleshit quotes?

No.1478750

>>1477917

Is it just me or is the first bit of this book really simple? im on 30 right now and the book is explaining highschool tier feedback loops to me.

No.1478764

>>1478734

He needed to reach his word count and it was one day before deadline, no bully.

No.1478772

>>1478750

I'm on a stone-age connection and didn't download that, but I did read a book on cybernetics by Lange. So if that is the same, I can tell you that it is really dull and simplistic, and it just tells you about the Pareto criterion and that if you are risk-averse and don't know about probabilities of different scenarios, you should choose based on performance in the worst-case scenario.

No.1478782

>>1478772

So that entire 200 page book doesnt really delve into math or anything? Its just 200 pages of introductory stuff to how feedback loops work and how you can use to to approximate optimality?

No.1478787

>>1478734

I only skimmed parts of the book myself, but I thought that it might offer helpful insights as to what they did wrong, and future projects should avoid.

No.1478821

>>1478787

It is written by a liberal guy who cannot think mathematically. He talks about linear programming, but then fails to give a proper example of it. He is only talking about word games, like how translators of American thinkers fiddled with specific terms to avoid being accused of being full of capitalist ideology and stuff like that, and how different bureaucrats dissed each other.

Here, have more waffleshit:

>Unlike market behavior, any deviations from the plan could send culpable ripple effects down or up the chain…

Different producers are dependent on each other in real life, independence and responsibility for you own success is a capitalist legal fiction.

>Kharkevich’s article is remembered among some technologists today not for proposing the ESS but for formulating what became known as Kharkevich’s law. This law holds that the quantity of information in a country grows proportionally to the square of the industrial potential of the country (N^2 ). The original formulation of his observation in the article is perhaps less elegant than information technologists might remember: “Given a large number of factories, the number of paired links between them is approximately equal to half of the square of the number of factories.” The law, in effect, prophesies a power law connection at the macro level between an industrial society and an information society.

I thought a power law is a relation that is independent of size? I don't know if the original article is like that or whether the translation is screwed, but this law just seems to describe the number of different ways of making a pair when you have N things: N*(N-1)/2

>These sibling laws (Moore’s 2^N and Kharkevich’s N^2 ) diverge interestingly in complex systems (when N is larger than 4). They also backlight their micro and macro focuses—Moore on microscopic industrial production and Kharkevich on informational industrial society. The result is different framings of the national network as a sort of central processor.

I'm pretty sure that the 2^N and N^2 already diverge in case of N = 1, mate. Waffle, waffle, waffle. It causes me physical pain when people write like that.

On Viktor Glushkov:

>His intense persistence of mind rendered possible his mathematical achievements, and his vision was shortsighted since youth due to his voracious reading habits. It is not known whether this contributed to his protracted struggle with a fatal brain tumor.

Can you get brain tumors from thinking too much?! Well, I guess if that's a problem, it's one the author of this book doesn't need to worry about.

>The hidden networks governing the Soviet state were far too complex and heterarchical to have had any single cause. (Most multiactor networks involve complexities that are impossible to express in linear form.)

Eureka, MULTIACTOR NETWORKS, you guys! Today I learned that the social networks and relationships that formed USSR society had more than one person in it!

No.1478953

>>1477917

So appearantly this introductory book is far too simple.

What else is interesting?

No.1479201

No.1480037

So it I see it correctly, one first has to learn Russian, to study anything worthwhile regarding Cybernetics?

>where to learn russian

No.1480078

>>1480037

I think the point of this thread is to have google translate it with their new translator while rusfriends help us with edge cases.

OP needs to post though.

No.1480241

>>1480078

You can find them too, just go into the link that I posted in the OP, look down into his bibliography, and start digging.

Finally found something in Russian on one of the researchers mentioned, Liapunov

http://www.nsc.ru/win/sbras/dates/Lyap/lyap1.html

It mentions that his main work is "Теоретические проблемы кибернетики", or "Theoretical Problems Of Cybernetics"

Looking for the book, I did find this pdf:

http://elib.ict.nsc.ru/jspui/bitstream/ICT/1154/1/Zhuravlev2.pdf

No.1480275

File: 4d0767df5f8edd5⋯.pdf (3.9 MB, Pioneers Of Soviet Computi….pdf)

Here's a full-length book on the topic in English, no translation required.

No.1480537

>>1479201

OK, this is good shit!

No.1480576

File: ffbc550c8549d90⋯.pdf (97.02 KB, Upravleniye Nauchno-Tekhni….pdf)

Management With Scientific-Technical Progress

This is an article actually written by the man Viktor Glushkov himself, the guy most often mentioned in conjunction with "Soviet Cybernetics".

Google translate has a function of translating pdfs: https://translate.google.com/?tr=f&hl=en

Download pdf and translate it. If there's a problem, just post the original phrase in Russian here.

No.1480739

File: acf83691d656ac1⋯.jpg (16.73 KB, 615x410, 3:2, robot-social-Robots-turn-h….jpg)

>>1477348

What was the soviet stance and your stance on robowaifus? This will determine if I help you or not

No.1480867

>>1480739

Does it matter? Robowaifus will presumably not be banned in the instance of anarchy. I don't know if anyone in the USSR ever considered it.

If there were Russian hikikimori though, they probably would've provided state-appointed wives to get them back to work. From that, we can extrapolate the possibility of a Soviet robowaifu project.

No.1481016

File: 6be62b77ec21d61⋯.jpg (256.44 KB, 585x707, 585:707, 6be62b77ec21d612973162dbd4….jpg)

Heya, I guess I posted with you in a thread a few days ago.

I honestly think that in the next cycle of leftist revolutions and governments, this "decentralized plan" by way of iterations in a real-time computer network will inevitably be attempted by at least one nation. It has the potential to be the X factor on making socialism stick for good.

It recently hit me that, before reading up on cybernetics and AI, it might be better to start by reading up on how the socialist economies did come up with their plans. The amount of data involved, the time and manpower it took, the ability to change data on the fly or not, and of course, all the practical details, because seriously, how the fuck do you even start planning out the economy of a fucking continent? I'll start making a list for this material. I suppose that, unsurprisingly, the juicy stuff is in Vodka, but the material in English should be more voluminous than the one about cybernetics.

Incidentally, there already exists something akin to this "decentralized plan" we're dreaming off. Electrical grids are monitored in terms of flow of power between a number of nodes, some generating electricity, most of them consuming. It's like a very simplified version of what we have in mind, but it's a standard technology used the world over for decades. Some key terms for googling are "power-flow study" and "one-line diagrams".

>>1480241

I don't think there should be much problem finding material in English on Lyapunov. His work was very, very important for control systems, to the point of being basic teaching material in the West even during the Cold War.

No.1481098

File: 8bcf8efa1146285⋯.png (329.04 KB, 540x360, 3:2, make america green again.png)

>>1481016

Was waiting for you to show up! Really excited to see what you have on hand!

>Some key terms for googling are "power-flow study" and "one-line diagrams".

Cool, will google those things.

>His work was very, very important for control systems, to the point of being basic teaching material in the West even during the Cold War.

So I can look up the names of his books in English?

> it might be better to start by reading up on how the socialist economies did come up with their plans.

IIRC, they used something called "material balances".

> in the next cycle of leftist revolutions and governments

With climate change putting a hard cap on how long humanity can survive with capitalism, we need to do this now.

No.1481445

Paul Cockshott is a computer scientist at the university of Glasgow, one of his primary research topics is cybernetically planned economy.

http://www.dcs.gla.ac.uk/%7Ewpc/reports/index.html#books

"Towards a new socialism" is downloadable for free and seems to be quite relevant.

No.1481466

>>1481445

Decent book, already read it. He has considerable ideological baggage, spending a not-insignificant portion of the book trying to rehabilitate the USSR as truly socialist without even knowing the proper definition of the word (an economic system characterized by social ownership of the means of production)

His essay "Calculation In Natura: From Neurath To Kantorovich" has the relevant details of rationally planning an economy via computers, without the baggage. The first 6 pages are milquetoast and irrelevant (talking about the death of socialism in the West), but the remainder of the 41 are nothing short of epic. They truly convinced me that computerized planning is the way forward for socialists.

No.1481505

What would be the goal in research of cybernetics, aside from of course having a proposal? I think it is the case that proposals essentially already exist, though obviously the real infrastructure and application likely wouldn't even be considered under current capitalist conditions.

Is there any value in creating simulations in order to indicate the way the system could work? At least as an experiment or proof of concept that might drum up attention among lay people, allowing them to not only conceptualize, but actually see the system in action?

No.1481551

File: 11869d8083239b3⋯.pdf (256.84 KB, Calculation In Natura - Fr….pdf)

>>1481505

Scroll down to Pg 25 to see example 2, Cockshott gives a good example of how this planning system works mathematically.

The main reason why I'm so interested in Soviet cybernetics specifically is the peripheral aspect of the whole thing - the algorithm for a program is relatively easy to develop (it's just linear optimization of input-output tables), the questions are "how do we get there from raw info", "how do firms input this info efficiently", etc.

I want to know what they thought were relevant issues with regard to the actual implementation, in short.

No.1481951

>>1481551

The "internet of things" should be useful in the peripheral aspect so we should be researching that area too.

https://en.m.wikipedia.org/wiki/Internet_of_things

No.1481952

>>1481949

The internet of things is a buzzword. All those toasters and fridges connected to the internet is not going to give us usefull information, its only going to give privacy issues.

No.1481960

>>1481952

"The Internet of things (IoT) is the inter-networking of physical devices, vehicles (also referred to as "connected devices" and "smart devices"), buildings, and other items—embedded with electronics, software, sensors, actuators, and network connectivity that enable these objects to collect and exchange data."

I think it is perfectly useful, and of course security might be the hardest part.

No.1481968

>>1481960

Its litterally just linking shit to the internet. The word itself is a buzzword, just like synergy.

Everything has to be "smart" but 90% of the shit they make smart will not produce any useful data and are just gimmicky. It will die down eventually. Being able to turn on your toaster downstairs with your phone is the definition of gimmicky.

No doubt that some things, like cars, can help in traffic control, but all of industry is already connected to the internet. Consumer electronics are not going to give a socialist state any useful information regarding industrial logistics and demand.

No.1481971

>>1481968

It would help tuning up the models of society in real time having the most accurate information on the status ( position and any parameter) of everything.

No.1481981

>>1481971

Do you have any idea how much computing power that would take?

Also what the fuck is the use of tracking where every fridge, toaster or phone is? We want economic planning, not fucking NSA cranked up to 11. It has no benefit to the planning of the economy and the cost to create such a massive model of society would be immense.

No.1481987

>>1481981

We need NSA cranked up to 11 In The Service of Communism.

No.1481990

>>1481987

Very funny, stalin.

No.1481993

>>1481990

Cyber Marxism-Leninism will win :P

No.1482339

>>1481981

>where they are

You're right, that could probably be discerned more generally by a kind of labor token price mechanism or something similar, like has been stated before. Though I already forgot what the paper said about that, because prices don't rise in a vacuum. Somebody raises the price, and if there is no profit to be had why raise it to check if there is still a demand for it? It seems like it would just stay at it's fixed price of cost and run out.

But either way, I think something can be devised to discern what people want in a region at the point of distribution. Even if it were as simple as having a sort of request system for anybody who couldn't find what they wanted at a store. That is useful data about demand.

But ultimately it seems like having more information is almost always good for planning. You can definitely be inefficient about how you are collecting information, since there is cost to it, but we already do collect a shitload of information that just happens to be isolated by different companies for their use.

No.1482566

File: 5bac4c371be218c⋯.png (106.66 KB, 1530x863, 1530:863, AASSTV1.PNG)

My autistic screeching, tell me watch you think tovarishes.

No.1482590

>>1481993

>:P

fuck off

No.1482699

>>1482566 checked

I don't understand it

No.1482704

No.1482739

>>1477348

If you're interested on the Western version of cybernetics, Operations Research is what you have to look for. Although resource planning in particular is not the main focus of research, actually more complex results were found, such as ways to solve non-linear systems.

No.1483117

>>1481016

>decentralized plan

>Electrical grids

try again

>Electrical grids are monitored in terms of flow of power between a number of nodes, some generating electricity, most of them consuming.

wew, you mean like it was monitored already in the fifties?

Soviets were monitoring reactive power coefficient on the level of the whole country

periodic reduction of this coefficient was one of the rare instances of planning actually working

electric power industry was one of the first to be significantly automated

mechanic relay systems did the job even before your "wew internet dude, new age!" mentality

personnel sits on their ass most of the time, and only when relay defense goes off they identify and fix the problem

>>1481551

>it's just linear optimization of input-output tables

great, you answered question of how to produce shit

but the real question is what to produce and what should be the structure of the end product

it is in society's interest to keep prices of goods as close as possible to their values

it is in enterprise's interest to keep prices of its goods as far away from their values as possible

No.1483419

>>1483117

>it is in society's interest to keep prices of goods as close as possible to their values

>it is in enterprise's interest to keep prices of its goods as far away from their values as possible

>enterprise's interest

You are projecting capitalist relations. If short-term some products are under-supplied relative to demand when priced according to labor costs, the price is increased (consider that people have similar income, so this roughly does ration according to need), but this increase doesn't land in the pockets of the people in the unit producing that thing. It doesn't land in the pockets of anybody.

No.1483509

File: 414fd9203606211⋯.gif (1017.8 KB, 325x203, 325:203, you will never know unless….gif)

>>1481098

>So I can look up the names of his books in English?

AFAIK he wrote no books, only academic articles, and extremely technical ones at that. But you could literally fill libraries with literature that teaches or expands upon his discoveries. I think the theories that would most concern our magical decentralized plan would be the functions and the stability theories that have his name. It's quite fascinating stuff, rendering real-life systems into numbers then checking if they're stable or will collapse given X time and how you might fix that and whatnot. Mind you, Lyapunov's material is not for novices, it's the sort of thing they start teaching midway through engineering degrees. And I figure that if someone does manage to get a prototype for the decentralized plan going, Lyapunov and other control systems material wouldn't be needed anytime soon, because there would be a lot of work in software design, modelling, data acquisition etc. that needs to come first, even if an existing power-flow study software was used as a basis.

>"material balances"

Hmm, apparently it's effectively what our algorithm would do: reduce all commodities to numbers (usually weight), and then measure its flow. I like that word. It's possible those early planners were using a cruder version of what we call control systems now. Cruder in method, but on an absurdly large scale.

>With climate change putting a hard cap on how long humanity can survive with capitalism, we need to do this now.

Well shit. Look man, I know it's great we're brainstorming and sharing very interesting texts and all, but this is a very long-term project for a real-life group/task force/superhero squadron, working full time, and more importantly, with a lot of academic background. Maybe one of us will manage to kickstart a pre-pre-pre-alpha version as a postgrad project, God willing. Maybe a hundred years from now, a curious soul begins to research the history of MARX Mx.II the AI that allowed mankind to end poverty despite still having scarcity, and he will find traces of a Chinese cartoon board where plucky dreamers dared dream of a better future long before it became reality.

No.1483524

File: 475a70073103e20⋯.jpg (17.75 KB, 600x342, 100:57, ea07_periodic_genius.jpg)

>>1482566

So we basically run simulations of Bolshevik Parties, until we reach the optmized Party?

Fucking brilliant. Someone give this man a Stalin Lenin Prize.

No.1483527

>>1483117

You have shitposted before in other cybernetics threads, haven't you? Your style feels really familiar.

No.1483533

>>1483509

I will try my best if I ever make it to university.

No.1484773

File: 11869d8083239b3⋯.pdf (256.84 KB, 11869d8083239b3cad10bc9c85….pdf)

File: 33ccc7a6fa62b5a⋯.jpg (302.83 KB, 800x800, 1:1, 33ccc7a6fa62b5ad8445101baf….jpg)

Bump

Are you guys doing your homeworks? I'm trying to create a very small program in Perl with this, page 25 example 1:

No.1486082

>>1484773

this looks fun

No.1486926

I have finished the book by Benjamin Peters and can now tell you that it is indeed shit all the way to the end. He also promotes the horseshoe model:

>the ruinous rise of socialism (whether Soviet, German national, or other form)

That's not to say that one should avoid all conservative or liberal writers (Neumann was a conservative, but his insights are so general that they are useful in general), I just want to say that this book is the perfect turd. It really has nothing going for it.

I'll check out Mikhail Botvinnik next, USSR chess guy that researched computer chess and apparently found something in that for a more general purpose, including economic planning.

And BUMP.

No.1487018

>>1481952

>fridges connected to the internet is not going to give us usefull information

Have a good think about this statement and then realize why you are wrong.

No.1488175

File: 448f0cf24b738e2⋯.jpg (49.21 KB, 640x480, 4:3, C0Lpio4WEAAF586.jpg)

No.1488231

File: bb73d50fb4af2f0⋯.jpg (12.1 KB, 250x263, 250:263, lenin-photo19.jpg)

>>1484773

God damn it, so much stuff to read it's ridiculous, even without this cybernetics binge now.

Some guy in another forum mentioned he discovered a Soviet mathematician with interesting ideas about rendering processes into mathematical models. I misplaced his name and will have to ask again tho. Maybe it's someone we already mentioned here, maybe not.

Oh, and it seems no one posted these, but here are a couple of great links.

http://bookzz.org/ for a shitload of pirated books

http://sci-hub.cc/ so you can pirate virtually any academic paper

No.1488243

So out of all this shit, which book would you recommend to a second year CS student who already knows whats in this paper

>>1477917

??

No.1488254

>>1488243

Suicide.

No.1488257

>>1488254

Fuck off cunt

No.1488303

>>1488257

Nigga this is "Conjure the level of data the Soviet Union had for the sake of wanking off in a 🍀🍀🍀study group🍀🍀🍀" not ????

Weasels have already shit all over this project by rejecting real life applications and opting for unusable "testing". A project like this can only exist in the real wold with real applications, made possible by real data from a real, willing, able, body which requires our help.

There was potential at the start when people were thinking big but it is now a mockery of all the previous ideas. People did their best to think as peabrained as possible and shit all over this as much as they could as early as they could and instead opted to go back to shuffling papers at the J. Edgar Hoover building making greentext frogthreads. What I'm trying to say is, for example, automating Kurdistan is far more likely than creating a system based off the internet of things which shovels hamburgers in your mouth. And there is a reason that has happened. One benefits a poor state in the ME, another is prime data. Wake the fuck up.

No.1488321

>>1488303

Jesus christ, take a fucking chillpill, dickhead. I just asked what, out of all the large papers and pdf's posted here, are most appropriate for my level of understanding. Don't project your anger against "muh internet of things" faggots onto me. I just asked for a fucking recommendation, if you can't even handle that you should consider killing yourself.

No.1488340

>>1488321

Wasn't trying to bully fam, was just making clear how much people hate this project.

No.1488343

>>1488340

So why would you direct that anger at me? Do you expect me, a fucking scrub who just manager to program a silly half-broken app together, to "just go to kurdistan and automate everything" or "just go to kurdistan and do big data". I haven't even learned the basics of fucking program structure because my school is retarded. Thats why I asked for a recommendation on this topic because it is interesting and I want to learn what smart people wrote about it, and how they might have done it.

No.1488389

>>1488343

Because I am drunk and angery, angery at everyone who doubts us in the thread.

Saturday night leftypol fever bitch!

No.1488453

File: 900e60c96d614e9⋯.jpg (80.78 KB, 950x960, 95:96, homer jubilant russian wor….jpg)

People please don't fight, it's highly unoptmizing. Also you know what they say about arguing with drunkards.

>>1488243

Honestly man, we're all learning as we go along too, reading a lot of shit before we can even put it all in a proper order – that is, even if we can read it all, because this is some arcane shit. So I can't say for sure what's the best introduction yet, but my best guess would be >>1481551

No.1488468

>>1488389

Fuck yeah dude I'm drunk too

No.1489209

File: 5e9753fc9a7f616⋯.jpg (289.43 KB, 1280x1792, 5:7, feets37.jpg)

>>1483419

>It doesn't land in the pockets of anybody

on the balance sheet of enterprise this increase in price in accordance to the level of demand will result in the profit

so you have two ways

1 center takes those profits from enterprises

2 enterprises pay rent to the state for the use of the MOP, and keep the rest of the profits, reinvesting them as they see fit

Cockshott is not very clear on the issue

his main concern is manipulation of prices around their values and calculation problem

anyway, there's no third way around this

as long as prices fluctuate around values, you have to somehow deal with profits

and as soviet example shows, enterprises very much don't like when the state takes their profits

and they will fight tooth and nail to keep them

>>1486926

>it is indeed shit all the way to the end

his book has some valid points

ARPANET was a centralized effort at its core

Soviet military and ministerial networks were scattered and disorganized

more so, they were incompatible on the hardware and software level

proposals for the shared military-civilian network and standardized unified data transfer protocols were shoved under the desk

this is facts, you can't just ignore them

why ministries were so isolated from one another? why are they often tried to sabotage each other?

only answer I can think of, is that because they had different antagonistic interests

>>1488243

don't bother with the fancy cybernetics

do yourself a favor and read some solid book on operations research and solve all problems in it

also another interesting book by the same author http://gen.lib.rus.ec/book/index.php?md5=E258CE40E4E969BDA57D9B44A71CC882

No.1489256

>>1488389

>>1488468

No.1489583

>>1489209

>>1488453

Thanks friends

No.1489601

>>1489209

>feets37

Post the rest? :3

No.1489836

Non-slide bump.

No.1489891

>>1488243

Read Towards a New Socialism by Paul Cockshott and Allin Cottrell, if you haven't yet.

http://ricardo.ecn.wfu.edu/~cottrell/socialism_book/

>>1489209

>as long as prices fluctuate around values, you have to somehow deal with profits

Do you picture the income spent on products as physical bills and coins that go to the shop, and then wander from the shop to the factory and from the factory to the state? Same recommendation for you as above.

No.1490268

Here is one simple example of linear programming in python

http://www.vision.ime.usp.br/~igor/articles/optimization-linprog.html

I would suggest using Jupyter Notebook for python, you can even use it interactively online

https://try.jupyter.org/

just create a new python notebook, paste code inside and play with numbers

No.1490657

>>1478485

The west had analog computers as well. Britain had a simple economic model made visible with water flows called the MONIAC, Monetary National Income Analogue Computer. It was more of a didactic gadget than a device for planning.

No.1492181

Bump

No.1492489

Bumpx2

No.1493472

So I've started reading Botvinnik's chess program book ("Computers, Chess and Long-Range Planning" which is the English translation of "An Algorithm for Chess") and got an idea useful in designing a program for chess and other board games, and I don't even play chess myself. The hubris, eh? But it's a simple and general idea, that I believe applies to economic planning as well: The medium-run goal of the program shouldn't just be to get nearer to what the specified end goal is, but it should aim towards creating situations it can comprehend well. Indeed I believe this is such a simple idea that many other people must have had it and I wonder whether there is a specific word for it.

No.1495445

So how are we going to get the data to model this?

No.1496614

>>1495445

Good question.

No.1496738

>>1495445

Depends on what exactly you want to test. best thing I can test is small web app, not worldwide economic simulation.

No.1497064

File: 48090111065607a⋯.jpg (73.94 KB, 600x600, 1:1, 1348010282557.jpg)

>>1489209

Holy fuck I cant even solve the problem with calculating the maximised surface of the square in the first problem set.

Am I retarded? How can you calculate the derivative? The function for the surface is

>S = a*b

and this is also true for the total length of sides L

>L = 2(a+b)

But if you substitute, say, a in function for S for

>a = 1/2*L - b

Then you get

>S = (1/2*L - b )*b

>S = -b^2 + 1/2*L*b

Which doesn't help you any further at all.

No.1497092

>>1497064

>Am I retarded?

I am. Holy shit.

No.1497793

>>1493472

Finished the book. Despite what the introduction promises, it really does not go beyond being about chess programming, there is a vague claim that it does and some very vague words, but no example. Not recommended.

No.1499705

>>1489209

Thanks these books look interesting

No.1499824

>>1496738

Let's start small then. How about simulating the national economy of Nauru? Climate change will make it cease to exist, further simpifying the simulation.

No.1500221

>>1497064

>Am I retarded

No, that's just maths for you. At first you don't get it and it seems completely impossible, and then at some point it starts making sense and it's easy as pie.

I'm not sure, but I think you're on the right track. Remember that L is a constant, even if it's not a defined value, so you can work with it. So:

>S'=1/2*L-2b=1/4*L-b

What you can see is that S'=0 when b=1/4L. S' is a linear function, and a falling one, so you know that S will be rising in value until b=1/4L and then start falling again, which means that S has its highest value when b=1/4L. So:

>b=1/4L

>L=2(a+b)=2(a+1/4L)=2a+1/2L

>2a=L-1/2L=1/2L

>a=1/4L

hence

>a=b

So the square has the highest surface when the length and width are equal

No.1500238

>>1500221

To make the middle part more clear:

>b=1/4*L

>L=2(a+b)=2(a+1/4*L)=2*a+1/2*L

>2a=L-1/2*L=1/2*L

>a=1/4*L

No.1500271

>>1500221

Yea someone explained to me that L was a constant. Its been 2 and a half years since I actually used derivatives to I forgot you can do that.

No.1501446

>>1499824

>Climate change will make it cease to exist, further simpifying the simulation.

Thank god, without the need of calculating weather we need one supercomputer less to accomplish this simple task!

No.1504479

Nice and easy bump.

No.1506355

File: 674316555f33c32⋯.png (158.84 KB, 1024x768, 4:3, OGAS.PNG)

Bump.

No.1508243

Bump

No.1508638

>>1477370

haha I love the cold war propaganda tone of that article

No.1508800

Read some essays by Stafford Beer. These were good for getting hyped about cybernetics (Beer must have been amazing at motivating people in person), but I didn't learn a damn thing about it. Perhaps it's better to just read up on logistics and voting theory / game theory to get an idea how to get not too dishonest data from self-interested humans for the logistics.

No.1509085

File: 5dcc45964146be9⋯.jpg (67.78 KB, 500x339, 500:339, cyb.jpg)

>>1508800

Beer is the shit. Project Cybersyn, yo. Fuckin' Pinochet crashed that shit. It was gonna crack too. I think if Allende had gotten a chance to launch it, it would have triggered mass revolution. https://en.wikipedia.org/wiki/Project_Cybersyn

No.1509618

File: b4dec8511cdd587⋯.png (49.42 KB, 672x760, 84:95, b4dec8511cdd587bb9004a07c3….png)

>>1509085

>the closest the world ever got to democratic socialism and computer-aided planning

>of course both had to be ruined by an Porky ball-licker

We truly live in the worst timeline.

No.1509783

>>1509618

Is in our hands to make it real comrade.

No.1512125

>>1509783

problem is that nobody here have a clue what we need to create.

Should we create some (almost) game AI which would try to create some town/country/commune/…? Or some protocol for consumer-producer communication, maybe even with self-destructive currency? I don't know…

No.1513001

Where could an aspiring anon go to study this subject in more detail? Are there any universities still doing research into cybernetics?

No.1513289

>>1513001

The meaning of "cybernetics" is different here and now than it was then and there. Currently, it's just anything that involves hi-tech crap, but originally it meant something like studying processes and things as systems, and how they could be interfaced and controlled. The key thing that has been (possibly) lost here is that these systems used to include virtually any kinda of systems instead of just electronic gizmos, including biological, psychological etc. and more importantly for us, economical systems. (Of course, the brunt of the operational aspects, the "under the hood" parts, always was STEM, but its applications on fields other than STEM are what has decreased.)

Of course, I can't tell for sure if there are or not any universities out in this big wide world that still teach this broader cybernetics. But I can tell you that, "mainstream-education-wise", the remit of old cybernetics has been kind of parcelled between various STEM disciplines (which again demonstrates the putative abandonment of biological, economical etc. areas).

Control systems is an important field within electrical and automation engineerings, among similar courses. Some places have the more specialized course of control engineering and, of course, just plain cybernetics. All of these courses are commonly found. I imagine that the amount of non-STEM applications involved varies a lot by university, and there might be plenty of programmes that teach oldskool cybernetics, but the trick would be finding those from amongst the "mainstream" ones. I suppose you'd have to ask in specialized communities of enthusiasts, engineers and such about which places might have the programme you're looking for.

No.1513946

>>1478668

So is there going to be anything done about this? Who would be interested in starting something?

No.1514553

>>1513946

>Who would be interested in starting something?

lot's of us are, but there needs to be a few ideas what exactly is "something"

No.1514815

>>1512125

>some protocol for consumer-producer communication, maybe even with self-destructive currency?

Something like that. And before doing that, start with reading about voting-system research and auctions to get inspiration for criteria or directly copy them.

No.1514914

File: 7f807d848dac31c⋯.jpg (100.79 KB, 600x840, 5:7, tumblr_lh8x3rTN2J1qz6f9yo1….jpg)

>>1513946

>>1514553

>>1512125

I don't want to be a wet towel, but let me point to my 3rd paragraph here: >>1483509

Even just a proof of concept would be a massive undertaking, not the sort of thing started by a group of ancoms who met on the internet. This AI/network/algorithm/Deus Ex Machina might one day bless mankind with unerring economic guidance, but I'm afraid its genesis won't be at our dear /leftypol/. Altho if you want to start a group so we can keep sharing information, I'm all for it. Or maybe we could have periodic threads here on /leftypol/? Once every 2 weeks or something?

For now, I'd say inform yourself and read as much on the topic as possible, and seek academic guidance on the topic. Not becoming a postgrad student or anything, but just talking to professors and their students about further reading and whatnot.

But of course, if you want to learn programming and start simulating a town or something, don't let me stop you. You'll learn a lot with the experience. In fact, any projects that deal with the many topics of our super-algorithm (flow study, programming in general, power grid diagrams, simulations, control systems, computer networks, software design etc.) is the sort of thing that makes for a cultured individual.

Oh yeah, was about to post when I remembered something that would be extremely valuable: translations from Russian, or acting as an interpreter between us and Russian academics. Like mentioned before, the vast majority of the material on Soviet cybernetics hasn't been translated yet, with >>1479201 being a perfect example. Of course, translating books and articles is no simple task either, but it's more doable by humble anons on the internet.

And to finish things up, some more links:

http://www.rand.org/pubs/reports/R0700z1.html

http://nautil.us/issue/23/dominoes/how-the-computer-got-its-revenge-on-the-soviet-union

There was at least one American periodical of Soviet cybernetics: Soviet Cybernetics Review. Also another general periodical called Joint Publications Research Service Reports occasionally had articles on the topic. Unfortunately, both of them are behind paywalls. Fucking Porky standing in the way of muh knowledge REEEEEEE An interesting sidenote is that the idea of a Soviet economy-optimizing national computer network had the CIA shitting its pants in the 60s.

No.1514934

>>1514914

We should start translating all those articles by date and putting them somewhere all together classified the best we can. A collective dropbox or something.

No.1515110

File: b4fd6d45d129d90⋯.jpg (33.59 KB, 600x338, 300:169, your brain on reformism.jpg)

>>1488175

Where is this from? I'm very interested.

>>1509085

>>1509618

The Soviet GOSPLAN already did computer-assisted economic planning in the 60s and 70s, but due to the limited power of the processors that were then available it could only calculate a few thousand key products of the economy. I imagine the Chilean Cybersyn would've fared even worse, only improving efficiency in some few sectors of the economy, and certainly not "triggering mass revolution". See Elab. 2.1 (pp.7-8) of: http://www.dcs.gla.ac.uk/~wpc/reports/quito.pdf (another great text by Cockshott and Cottrell, elaborating on the thesis of socialism based on computerized in-natura calculation)

All in all, the mistake of Allende and the Chilean socialists back then, was their naive trust in the bourgeois political system and their unwillingness to let the proles rise in arms. They pretty much demonstrated that the pretensions of a "democratic way to socialism" are a meme and forever will be. The correct line was held by the MIR all along; by '73 it reached thousands of militants in size, led the seizure of farms and factories all over the country (much to the dismay of the reformists), and managed to build a primitive military structure: https://en.wikipedia.org/wiki/Revolutionary_Left_Movement_(Chile)

No.1515185

>>1488231

>Some guy in another forum mentioned he discovered a Soviet mathematician with interesting ideas about rendering processes into mathematical models. I misplaced his name and will have to ask again tho. Maybe it's someone we already mentioned here, maybe not.

Leonid Kantorovich?

No.1515190

File: ffc1154cf497095⋯.jpg (184.09 KB, 640x629, 640:629, ffc1154cf497095f55d28887af….jpg)

>>1514934

Speaking from experience, a good translation is more time-consuming and tiring than you think. It's the sort of freelance work that ends up being better remunerated than most people imagine. We would be extremely lucky if we found a Russian translator willing to do it for free.

>>1515110

As I understand it, the GOSPLAN computer assistance was effectively just a souped-up calculator, not an actual simulator like what we're talking here. My notion is that the apparatchiks limited the use of computers for the same reason they pulled the plug on cybernetics as a whole: they would lose power.

Also

>mfw he could have proclaimed socialist Chile 3 months earlier thanks to the Tanquetazo, but still took the high road

No.1515237

>>1514914

>ancoms

Hey now, not all of us are retarded.

No.1517679

>>1514553

At least a chat (IRC or Discord) or a periodic thread on the topic?

No.1517705

>>1517679

let's go with periodic thread. let's make it… every even week's saturday? IRC/discord is circlejerk.

No.1517726

>>1517679

Correction: wanted to say cyclical thread, like the /rojava/ one.

No.1517799

>>1514815

The allocation system should react in a monotonic way to inputs. So, with a single-winner voting system, comparing two situations that are only different in that in the second situation some thing X is ranked higher on some ballots, in the second situation the winning probability of X shouldn't be lower (the criterion is called mono-raise in voting system analysis).

Likewise, in a system for allocating multiple resources to people who bid on it, comparing two situations that are only different in that in the second situation one actor bids higher on some item X, in the second situation the probability of that actor obtaining X shouldn't be lower. Now, where does the actor got that extra "money" to bid more in the second situation? If the allocation procedure uses some point system that has no usage outside of it, and no points are saved up for later usage inside of it, it's just all used up and when the system is used again at a later date everybody gets a new endowment of points for bidding, that means the actor got these points by reducing their other bids. We can split this criterion into two variants: One where the other bids by the actor are shrunk proportionally by some factor that frees up points for increasing that one bid compared to the first situation, and a stronger more general criterion where the other bids by the actor are just any combination of same or lower amounts as in the first situation to free up some points for that one higher bid.

Further variations are possible: instead of a bid for one unit of a thing, it can be a bid for multiple units of the same thing, and the guarantee the criterion asks for is that the higher bid doesn't result in getting less.

No.1517950

Continuing from >>1517799:

Another import from voting system analysis is clones. Clones are similar candidates. A voting system is robust with respect to clones if shrinking or increasing the size of a clone set has no effect on whether the winner comes from the clone set or not; and furthermore, if the winner is outside the clone set, it shouldn't change either. How to properly define clones, then?

For rating ballots, clones can be defined as candidates lumped together in the same rating on every ballot: So, if we do a poll on the best video game ever, and everybody who rates Virtua Street Kombat at three stars also rates Mortal Samurai Fight at three, and everybody who rates VSK at four stars also rates MSF at four and so on, we call these rating clones. Average ratings is robust with respect to rating clones. This is also true for median ratings. For ranked ballots, clones are defined as sets of candidates that are always ranked together without another candidate between them. This is quite different from rating clones in that subsets of clone sets are not necessarily clone sets themselves.

Both variants leave something to be desired in that the clone definition is so narrow that you are unlikely to encounter them in the wild (adding a single ballot not following the pattern destroys these clones), but it's a starting point. Rating methods that aren't even robust with respect to rating clones should be avoided and likewise with ranking methods.

Inspired by this for our analysis of systems handling allocation requests, let's call ratio clones things that are requested in fixed proportions. So, suppose Alice wants 30 X and 20 Y and Bob wants 12 X and 8 Y and everybody else wants these two things in this 3 : 2 ratio, then X and Y are a ratio clone set. A second situation that is only different in that now X and Y are only available in a package called Z in the aforementioned ratio, with people making their requests as before, should leave the allocation basically unaffected.

No.1519463

cont. from >>1517950

The participation criterion states that adding ballots shouldn't make the results worse from the point of view of what is stated on the ballots. So, if your ranking is A>B>C>D, then adding your ballot to the pile shouldn't change the result from B to C , for instance. I doubt it's a good idea to insist on it, as it is a very restrictive criterion. Many voting methods, like instant-runoff voting and all Condorcet methods, fail this criterion.

Median ratings fails here as well, and that's the method I would like to see for councils determining prices and quantities (and for juries determining penalties). I'm particularly partial towards a version of this with deliberation. Instead of instantly setting the median, there would be two voting rounds. In the first round, people are asked to specify an upper and lower limit (you can put in the same value twice if you are already sure), then the upper limit is set by deleting the highest quarter (rounded down) of stated values and taking the highest one from the remaining ones, and likewise the lower limit is set by deleting the lowest quarter (rounded down) of stated values and taking the lowest from the remaining ones. The final value has to be within these limits. There is a discussion, and then there is a median vote for what the value is going to be.

Giving a median rating to a single thing doesn't fail participation, but when rating several things it can happen that in the aggregate X is rated higher than Y, you add some ballots also rating X higher than Y, and the aggregate changes so that now Y is rated higher than X. This problem can arise when the resolution of the ballots is higher than three grades.

No.1519594

What if we create a website? And make it a virtual political party with communist cybernetics as it's constitution?

We could create and manage it from here in the /leftypol/ cyclical threads.