[ / / / / / / / / / / / / / ] [ dir / choroy / dempart / klpmm / mde / randamu / tingles / trueb / vril ]

/leftypol/ - Leftist Politically Incorrect

A collective of people engaged in pretty much what the name suggests
Winner of the 77nd Attention-Hungry Games
/x/ - Paranormal Phenomena and The RCP Authority

April 2019 - 8chan Transparency Report
Name
Email
Subject
Comment *
File
Password (Randomized for file and post deletion; you may also set your own.)
* = required field[▶ Show post options & limits]
Confused? See the FAQ.
Flag
Options
dicesidesmodifier

Allowed file types:jpg, jpeg, gif, png, webm, mp4, pdf
Max filesize is 16 MB.
Max image dimensions are 15000 x 15000.
You may upload 5 per post.


File: dac9398d156822c⋯.jpg (58.85 KB, 500x500, 1:1, Cybernetic Empire.jpg)

 No.2891312

Once again, it's the thread about computer-aided economic planning and direct democracy and general Cockshott news I guess (old thread: https://8ch.net/leftypol/res/2781898.html).

His latest blogpost: https://paulcockshott.wordpress.com/2019/05/14/input-output-or-harmony-planning/

>Kantorovich held that his Objective Valuations, would in the long term converge on labour values, but that in the short run they were useful to ensure that the best use was made of currently available plant and equipment.

>I arrived at the Harmony approach whilst thinking over how to come up with a better algorithm for solving Kantorovich’s problem. As it happens I was in Budapest in 89 looking to find a Hungarian publisher/translator for the work the later became Towards a New Socialism and came across a Hungarian planning text in English and thinking about it whilst in a concert, I thought of an improved algorithm…

Read "Towards a New Socialism" by Paul Cockshott and Allin Cottrell, if you haven't already: ricardo.ecn.wfu.edu/~cottrell/socialism_book

 No.2891331

File: a2de67bd0a8bb73⋯.jpg (121.85 KB, 1280x640, 2:1, 1280px-Whole_world_-_land_….jpg)

So which in countries can you implement any of these ideas, taking into account their political realities?


 No.2891332

>>2891331

North Korea, Venezuela, Cuba, China, Bolivia, Vietnam? Roj*ava?

These are all the most "socialist" countries out there, I think those have the best chances. Obviously it isn't in the interest of the bourgeoise to do this.


 No.2891346

>>2891312

Can we also do a meme dump of all cockshott/cybernetics memes I’m especially looking for the vapor wave ones


 No.2891415

here's another archive

https://archive.is/j942c


 No.2891417

>>2891346

pdfs too


 No.2891506

>>2891312

The better question is, what programming language will the econ planning system be written in? When we did a poll on it back in 2017, the answer was python, oddly enough. Thought I'm not sure python would be the best language for a production level numerical computing program. Probably 95%+ of it can be written in python(backend logic and APIs)and the intense numerical computing parts written in C++.

TBH im not sure about the whole apis thing since REST is now being phased out in favor of things like GRAPHQL. Instead of having a front end app call restful backend services, it makes one graphql query to a graphql api server, which then in turn calls the microservices that provide the info in a serialized format, after which point its returned to the front end app (written in whatever js framework is flavor of the month, or likely in the future as a web assembly).


 No.2891529

>>2891506

>Probably 95%+ of it can be written in python(backend logic and APIs)and the intense numerical computing parts written in C++

Python has no multithreading, making it terrible for performant servers. As well, it has dynamic typing and lots of runtime errors, meaning it is not really secure.

Slow interpreted languages (IE, all of them aside from Lisp) should be phased out in communism. Compiled languages can be just as convenient and easy. As well, strong type systems are very important.

Go:

- Simple (friendly to workers of varying skill levels), nice multithreading, reasonable speed, compiled, typed. But kind of lame in some aspects. Rejects bourgeois OOP ideology.

Rust:

- Very fast, pretty easy, but complex. Fantastic safety guarantees, type system, etc, and more modern than ugly old ADA. Also rejects bourgeois OOP.

Java:

- Current dominant corporate language. Familiar to all worldwide. Pretty simple. Has fast modern web library (vertx). Awful OOP shit. Cockblast's POC program is written in it.


 No.2891538

>>2891529

Lol, I'm using ada.


 No.2891542

>>2891529

I'm a brainlet, why is OOP shit?


 No.2891547

>>2891538

More power to you. It's great for what it is.

>>2891542

Virtual function table causing lots of cache misses

"everything is a reference" causing lots of cache misses (issue does not exist to same degree in c++)

encourages packing loads of fields into one struct, polluting cache

huge class hierarchies, most data hiding is just obfuscation

hacky design patterns that aren't needed in other paradigms

outdated type systems

literally Platonic idealism

promoted by managers because they think it makes programmers more replaceable

other stuff too. Look up "Data oriented design."


 No.2891583

>>2891547

>>2891529

I tend to agree with you, and that OOP is generally shit. Functional all the way and procedural for the low level bits that need optimization.

>dynamic typing

agree it sucks, the only reason it became popular is because you didnt have to write the type name in front of variables making development faster. Now with static compiled languages having type inference (C++'s auto and C#'s/Java10+'s 'var' for examples) there really is not much reason to use dynamic typing.

Lisp is an exception to dynamic languages being shit because its uniquely simple syntactic structure makes it have unique meta-programming properties i.e. macros.

just a quick pedantic note: not a golang programmer, but was under the impression that go coroutines were not "muh troo multithreading".

Rust: I like what I see, is likely to slow down development though as you have a very strict type system. Still, good choice for long term production systems.

JAVA: HELL to the NO. I know cockshott likes it but he was a programmer back in the 70s/80s/90s and his whole shtick was a version of Pascal he made, and if youve seen pascal/delphi syntax it looks sortof like a procedural version of java. The JVM is good, but java itself is a dogshit language. eww eww ewww ewwww. Its the COBOL of the 21st century. Maybe a JVM based language though like Scala or Clojure.

One thing I havent heard mentioned is DOTNET CORE. with dotnet 5 coming out in 2020 (supposedly) all of the current disparate dotnet platforms for mobile, desktop, game engines, etc. will be unified under a single, open sourced and free platform of dotnet, phasing out the old proprietary asp.

dotnet advantages:

C# is better than java.

has native functional programming in terms of F#

mainstream language so has all kinds of libraries/frameworks to choose from.


 No.2891592

File: c12d6e30781ad96⋯.gif (133.55 KB, 340x340, 1:1, lauging.gif)

>>2891529

>bourgeois OOP

8/10 bait

>>2891547

>muh cache misses

not real problem

>data hiding

>hacky design patterns

aka "I can't into proper basic architecture"


 No.2891627

>>2891592

>muh OOP is good

ok if you think thats BS what about difficulties with parallel programming due to mutable state


 No.2891662

>>2891592

ok the bourgeois OOP is obviously a shitpost, but you have to admit that the way OOP has been interpreted by mainstream corporate languages like JAVA is terrible and has lead to bloated terrible software.

Design patterns A LA gang of four arent really necessary in FP, plus immutability makes unit testing easier as you only have to test the function and not also take into account stateful variables. Plus you can more easily parallelize operations


 No.2891704

>>2891627

>>2891662

ok, I understand your point, except for "terrible" OOP interpretation of JAVA, with which I have absolutely no experience as primary .net developer. Every time I have to work with lot's of data I push all(most of) business logic onto database engine which do this kind of heavy lifting very efficiently.


 No.2891754

File: 80a80270487edd1⋯.jpg (38.52 KB, 671x701, 671:701, _20190515_124047.JPG)

>>2891506

Ada, Haskell, some Lisp.


 No.2891762

>>2891583

>not a golang programmer, but was under the impression that go coroutines were not "muh troo multithreading".

It's an async IO/thread pool model. A mix of userspace threads and OS threads. Also called "green threads." There will usually be fewer OS threads than goroutines, since the blocked goroutines can be moved off the OS thread and replaced with a routine that needs to run while it waits for an IO signal.

>Rust: I like what I see, is likely to slow down development though as you have a very strict type system.

Well, isn't that precisely a bourgeois mindset? Languages without Algebraic Data Types and memory safety assurances slow development down immensely once the codebase gets big enough, because of all the debugging headaches. But from the capitalist POV, it's just getting venture capital and shipping a half-finished product that counts.

As well, Rust code can be very abstracted and clean. It has all of the nice quality of life stuff people want– iterators, map/filter/reduce/etc, lambdas, type inference… Once you understand the "everything is a move" management model and a few other things, there's no reason it couldn't be great for scripting.

>JAVA: HELL to the NO. I know cockshott likes it but he was a programmer back in the 70s/80s/90s and his whole shtick was a version of Pascal he made, and if youve seen pascal/delphi syntax it looks sortof like a procedural version of java. The JVM is good, but java itself is a dogshit language. eww eww ewww ewwww. Its the COBOL of the 21st century. Maybe a JVM based language though like Scala or Clojure.

I agree, but I mentioned it out of consideration for the Java-programming masses. Really, they should be re-trained to better languages (some of which, like Go, are simpler to use).

>One thing I havent heard mentioned is DOTNET CORE. with dotnet 5 coming out in 2020 (supposedly) all of the current disparate dotnet platforms for mobile, desktop, game engines, etc. will be unified under a single, open sourced and free platform of dotnet, phasing out the old proprietary asp.

I've heard it's nice, but a lot of C#/DOTNET stuff is still not Linux-friendly. You are right, F# is good, though still has some jank with interop.


 No.2891765

>>2891592

>>muh cache misses

>not real problem

Cache misses are making most programs five to twenty or even hundreds of times slower than they should be. L1 cache takes like 2 cycles to move into registers, main memory takes upwards of 300 in many cases.

>aka "I can't into proper basic architecture"

read

http://www.dataorienteddesign.com/dodbook/


 No.2891774

i like the armchair programming experts ITT who act like they know what's what but haven't contributed a goddamn thing to any meaningful open source project


 No.2891829

>>2891506

It's neural linguistic.


 No.2891833

Parallel neural cybernetic networks made from dimorphic radiotropic fungi have THZ+ bandwidth.


 No.2891871

>>2891529

>>2891506

>>2891583

>>2891754

C/C++ or nothing. There would be no free software without the GNU Compiler Collection.

>>2891592

My comrade right here.


 No.2891910

>>2891871

>Upholds C, yet simultaneously upholds OOP

wtf

your entire precious GNU system is written with barely any function/data coupling at all, and with no type inheritance.

In any case, C is good for what it is, but "everything is a nullable pointer" is one of history's greatest mistakes, it has UB, c-strings are not good, and it is just plain old.

C++ is a travesty, you only have to see one std template compile error barf to understand that.


 No.2891926

>>2891910

ok, go rewrite linux in Go or some other meme language since you know what's best


 No.2891928

>>2891506

BBC BASIC


 No.2891932

File: 617a6ca3c3b8eff⋯.jpg (63.89 KB, 1200x630, 40:21, talk-is-cheap.jpg)


 No.2891960

How are you going to secure the planning algorithm and handivote from hacking and sabotage?


 No.2891972

>>2891774

I’m a programmer but I haven’t made any open source contributions because I’m too busy writing code for rent money in my 9-5 and porky owns all the code I’ve been making


 No.2891998

>>2891960

In my view, the planning algorithm should be backed up by multiple independent agencies checking each other's calculations. The voting system would require advanced cryptography.


 No.2892004

>>2891932

yeah sure linus, not like actually discussing language choice and architectural designs beforehand is good or anything, better to just start shitting out unmaintainable code in whatever language/platform/framework i feel like and stick the rest of the people who will work on it with those decisions for the rest of time amirite

>>2891960

Good question. The answer is pretty complicated, because there are so many levels of security - network level, physical access, vulnerabilities in the code. Basically the codebase would need to undergo periodic audits, as well as the networks/servers hosting the system which would need security/pen testing.

>>2891754

Rust is the new Ada. Lisp = good for scripting

Haskell = language from 1000 years from the future that i cant understand because i dont have a phd in category theory. Really though haskell is great but every time i try, i just cant into it

>>2891762

>goroutines

sounds good

>rust vs rails

I agree that startup style hipster development, just shoving an inefficient half baked minimal viable product out the door as fast as possible would not be the goal of an open sourced project. You are really rustpilling me here

>dotnet

yes, we will have to wait to see how dotnet 5 does in 2020. If all goes well we will have a performant native dotnet implementation on all 3 major OS's


 No.2892010

the bigger question is hosting though, obviously cloud hosting such as AWS, Google Cloud, Azure also digital ocean, ibm, oracle, vultr, and countless others is out of the question as almost all those companies are under the jurisdiction of the US government which means if you have a third world country that tries this they would need to have their own servers and network, 90s style. The only cloud provider even remotely considerable is alibaba cloud because its chinese.

Perhaps the planners would have to buy a bunch of servers/network equipment and implement their own private cloud using something like openstack. This itself is not outside the range of possibility as major corporations (Target, etc.) already do this


 No.2892013

>>2891910

>>2891871

C/C++ would be great for the performant bits but youre not thinking broadly enough. What about the UI ? wouldnt it be better as a web app and why write that in C++?


 No.2892015

>>2891704

>make the database do it

yeah its the same at my work too, though in oracle stack rather than microsoft. Push all the logic into the database in the form of stored procedures etc. The problem is this is an old enterprise architectural style, in a major corporation for example you have not just multiple testing environments of the same app, but multiple apps (for HR, accounting, /whatever) on the same really really buff oracle or sql server because hey, the management shilled out a shit ton of money for it, so every app becomes dependent on it. Then you need a high paid database admin to babysit it, because if something goes wrong, large parts of the whole business critical functions stop working.

Nowdays things are shifting to a MICROSERVICES architecture:

https://microservices.io/

where every service = a business function and has its own internal data store


 No.2892018

File: ad4da813538037a⋯.jpg (97.67 KB, 1024x742, 512:371, D1L9EgIUwAAabBY.jpg)

> web app


 No.2892023

>>2891932

You can download this html doc with javascript and run in your browser: https://pastebin.com/bPyr7Vau The purpose of this is to divide up a given amount of free identical cookies among individuals based on what they wish for without giving them an incentive to exaggerate (like would happen with scaling down all stated amounts by the same factor to make the sum equal to what's available). Just a few more features *cough* and we'll have a complete computer-communism package.


 No.2892027

>>2891506

Is this even a relevant question? The way I'm imagining it the architecture of the planning software would be defined in human language, and there would be space for some diversity in how this schematic is implemented. What matters is how the different pieces of the planning system fit together and produce a result. Each node is a black box, what's going on inside doesn't matter as much as its external behaviour.

I'm not a programmer though, so maybe someone could enlighten me.

>Probably 95%+ of it can be written in python(backend logic and APIs)and the intense numerical computing parts written in C++.

Something like this is probably how it'd end up working, I think.


 No.2892070

>>2892027

Yeah, we don't live in the 80s anymore, so something particularly well-optimized isn't needed. The hardware cost really isn't an issue. Every component of the whole thing could be implemented in seven different ways, with different outputs by components dedicated to the same goal decided by a virtual vote among them.


 No.2892132

>>2892070

Ideally the different implementations wouldn't all be running. Resources aren't as big of an issue as they used to be, but they still matter. What I imagine is different planning agencies regularly checking each other's calculations with their own software in order to prevent fraud.


 No.2892162

>>2891926

Linus himself has said Rust is pretty good and solves important problems, but also that the core kernel code is, in his view, a solved problem. They have coding standards, a history, and procedures that make the core kernel pretty damn safe.

However, the core kernel is only like 200k LOC, meaning it wouldn't be wildly impossible to rewrite it. And as for kernel drivers, many of which are not up to the same standards as the core, Rust is very viable.

Go is GCed so it is not viable for any perf-critical code like a kernel.

>>2891960

The planning system and data (barring military secrets) should be open source so everyone can verify the plans on their own PCs or servers. As for general security stuff, same as any bank or server system. Actually, most banks have awful security from the 90s. The labor credit system would be much more secure. Aside from that, communism would heavily disincentivize illegal hacking, and there could be a great firewall.

>>2892010

If a country has a communist revolution, it can host its planning and banking systems in private servers, not the cloud. Cockshott's system is much more efficient than the modern banking system, as it gets rid of huge amounts of speculation, exchange between producers, service fees, and so on. The cloud is not some kind of miracle, it really sucks in many ways. If you're just making a demo app, the cloud is probably sufficient.

>>2892013

I think we were talking about backend. Frontend definitely should be a functional reactive framework like Elm, ReasonML, Elmish etc. Zero runtime errors. Preferably on WebASM, or we could do as god intended and switch browser scripting to a Lisp interpreter.

Wouldn't hurt to have a standalone desktop app either, so everyone just has it in their package manager. But you could do that in QT, or we could have a new desktop GUI framework that uses FRP and data oriented programming.

>>2892018

I think it just depends on which component. The labor credit banking should obviously have a webapp. The numbers-crunching parts of the program are more suited to a desktop or CLI app.

>>2892070

>Yeah, we don't live in the 80s anymore, so something particularly well-optimized isn't needed. The hardware cost really isn't an issue.

Absolutely bourgeois. This mentality is why our programs keep getting slower in spite of hardware speedups. As well, we are at like 7nm transistors now. The age of ever-increasing CPU speeds is over. We need to preserve the environment by making faster, more efficient programs and selling slower, less power-hungry hardware that lasts a long time.

Look how much slower Python, PHP, etc are compared to Go, vertx, actix, etc. Some of the best modern backend systems are literally 50x faster than Django or Rails. That means, if you have a socialist cloud service, Python users are taking up a massive amount of resources other people need.


 No.2892178

>>2892132

I was more thinking along the lines of one big agency doing the same calculations in several places. You need to have redundancy in that stuff – and not just in software, but hardware, since even when humans do everything right there is always the risk of cosmic radiation flipping a bit; my inspiration is NASA/NGE. Not really a fan of having regional planning agencies.


 No.2892260

>>2892162

>it can host its planning and banking systems in private servers, not the cloud.

yeah by making its own private cloud using the computers and networking equipment they have on hand using open sourced software like openstack

https://www.openstack.org/

the whole reason for this is to conserve computing resources by using virtual machines and lightweight function as a service. if you have to make each service on an actual hardware then when its not being utilized 100% its basically just sitting there using up power. It saves energy, compute cycles, and is more environmental to make more efficient use of computers/servers by having them as the basis of a private cloud system under open stack and implementing the necessary services on VM's/ FAAS on that. I forsee each socialist country in the future having its own cloud, basically.

>WEBASM/LISP

switching browser to lisp (probably scheme in that case) sounds good but it would mean having to roll our own browser or browser plugin to parse lisp and execute it… unless youre talking about clojurescript or something.

>le moores law means i dont have to write performant code anymore

correct. Although i will point out that PHP7+ actually has way faster execution that python.

>>2892178

>le byzantine generals problem

This is more applicable to embedded software than apps running on normal servers. Unless you plan to be running the software on a satellite in orbit or something I don't think this will be a problem


 No.2892289

>>2892260

>the whole reason for this is to conserve computing resources by using virtual machines and lightweight function as a service. if you have to make each service on an actual hardware then when its not being utilized 100% its basically just sitting there using up power. It saves energy, compute cycles, and is more environmental to make more efficient use of computers/servers by having them as the basis of a private cloud system under open stack and implementing the necessary services on VM's/ FAAS on that. I forsee each socialist country in the future having its own cloud, basically.

Thank you for pointing this out.

>switching browser to lisp (probably scheme in that case) sounds good but it would mean having to roll our own browser or browser plugin to parse lisp and execute it… unless youre talking about clojurescript or something.

For a government project, it would be relatively easy to fork Firefox and give it a Lisp interpreter alongside its JS interpreter. JS was originally just supposed to be a flavor of Scheme anyway. Scheme would be nice for FRP, and a good transpilation target. However, WASM does make transpiling a bit redundant.

>Although i will point out that PHP7+ actually has way faster execution that python.

I have heard this about PHP7, but it is still PHP.


 No.2892295

Gonna share two posts from /marx/ regarding Cybernetics:

Cybernetics was most popular in the USSR in the Khrushchev period, when the CPSU was claiming that Soviet citizens would be experiencing communism in 20 years' time, colonizing planets, etc. But supporters of cybernetics had opposition from both "the establishment" (government ministries didn't want to replace existing planning methods and were concerned about issues like unemployment if so much human labor was no longer necessary) and from supporters of a more market-based approach (who argued that cybernetics was a waste of time and resources which would only result in a further centralization of the economy to the detriment of its efficiency.)

There's a book about the subject titled From Newspeak to Cyberspeak: A History of Soviet Cybernetics, which can be found here: http://b-ok.xyz/book/899907/0135f4

Here's an excerpt from page 273:

>Economic cyberneticians quickly realized that it was impossible to centralize all economic decision making in Moscow: the mathematical optimization of a large-scale system was simply not feasible. CEMI researchers estimated that complete optimization of the Soviet economy required solving a gargantuan system of equations with 50 million variables and 5 million constraints. They admitted that even a computer performing 1 million operations per second, which was much faster than any available Soviet computers, would require one month to solve a system a billionth as large. Besides, economic cyberneticians realized that there were some serious conceptual difficulties: linear programming was suited for the problem of resource distribution, but it did not work well for prospective planning, there were different views on what constituted an economic optimum, and it was difficult to agree on a single criterion for optimization. In 1967 Fedorenko unequivocally stated that "the full formalization of the functioning of an economic system and the creation of a fully automated centralized system of planning and management of the economy is unwarranted.

With issues like that, it isn't surprising that cybernetics was criticized as impractical. Soviet planners and Party leaders did constantly talk about the need to implement the "scientific and technological revolution," but they envisioned it as assisting current methods to plan the economy rather than anything drastic.


 No.2892296

>>2892295

The second one is less related to Cybernetics but more about problems socialist countries have faced:

The Chinese and Cubans are still trying to figure that out, and there's no shortage of authors in the West with their own views (e.g. Paul Cockshott.) However, I can quote Parenti for some examples of problems the socialist countries faced. From his "Blackshirts and Reds," pp. 60-61:

>1. Managers were little inclined to pursue technological paths that might lead to their own obsolescence. Many of them were not competent in the new technologies and should have been replaced.

>2. Managers received no rewards for taking risks. They maintained their positions regardless of whether innovative technology was developed, as was true of their superiors and central planners.

>3. Supplies needed for technological change were not readily available. Since inputs were fixed by the plan and all materials and labor were fully committed, it was difficult to divert resources to innovative production. In addition, experimentation increased the risks of fail- ing to meet one's quotas.

>4. There was no incentive to produce better machines for other enterprises since that brought no rewards to one's own firm. Quite the contrary, under the pressure to get quantitative results, managers often cut corners on quality.

>5. There was a scarcity of replacement parts both for industrial production and for durable-use consumer goods. Because top planners set such artificially low prices for spare parts, it was seldom cost-efficient for factories to produce them.

>6. Because producers did not pay real-value prices for raw materials, fuel, and other things, enterprises often used them inefficiently.

>7. Productive capacity was under-utilized. Problems of distribution led to excessive unused inventory. Because of irregular shipments, there was a tendency to hoard more than could be put into production, further adding to shortages.

>8. Improvements in production would lead only to an increase in one's production quota. In effect, well-run factories were punished with greater work loads. Poor performing ones were rewarded with lower quotas and state subsidies.

>Managerial irresponsibility was a problem in agriculture as well as industry. One Vietnamese farm organizers comment could describe the situation in most other communist countries: "The painful lesson of [farm] cooperatization was that management was not motivated to succeed or produce." If anything, farm management was often motivated to provide a poor product. For instance, since state buyers of meat paid attention to quantity rather than quality, collective farmers maximized profits by producing fatter animals. Consumers might not care to eat fatty meat but that was their problem. Only a foolish or saintly farmer would work harder to produce better quality meat for the privilege of getting paid less.


 No.2892306

>>2892260

>Unless you plan to be running the software on a satellite in orbit or something I don't think this will be a problem

What's the effect of a bug? This is not about software for a factory, it is about software for a continent. Its output does not get an extra check by humans going over it and having a discussion what to do with it, the updates are too frequent for that; so even though it's humans down the line that order other humans and not the Computergod blackmailing you with leaking your nudes, the frequency and quantity of the program output won't leave room for anything but turning the recommendations into directly binding orders. The human control is only changing the parameters for the program, which happens at much lower frequency, between these updates the program plays with itself and generates thousands of short-term plans. How can you be sure a program is bug-free unless it isn't much more complex than Tic-tac-toe? This project isn't like making a smartphone game, it is more serious than writing software for hospitals. That's why there needs to be redundancy with several implementations.


 No.2892322

>>2892306

well, the core calculation algorithms are very short pieces of code. easy to verify and write tests for. of course some kinds of bugs could creep up in the db, interface, network, and elsewhere. However, I think you are overstating the precision of the issue here. The plans are clearly very precise compared to capitalism or soviet socialism, but at the end of the day, we will need to build tolerances for human error and random stuff happening into the way the system is enforced. make a "see something, say something" campaign for people to notice weird things going on in the plan. don't sweat minor cases of overproduction or underproduction. stuff like that.


 No.2892473

>>2892306

>What's the effect of a bug? This is not about software for a factory, it is about software for a continent.

the whole byzantine generals issue is because it is embedded software running in a hostile environment, not an air conditioned server in a warehouse datacenter not being hit by massive solar radiation every day. Hardware fault tolerance has been a solved issue for decades now except in quantum computers, the problems like that would have been corrected at a lower level of abstraction, as application level programmers we can largely trust the computer unless you want to make it as an embedded program on custom hardware/plcs designed by electrical engineers specifically


 No.2892494

>>2892473

What about getting hit by an earthquake or NATO bombing?


 No.2892503

>>2892494

sure there should be escrow backup sites, but that is for redundancy and data backup purposes, being geographically distributed. nothing to do with errors in calculation. If the server is bombed it wont give a wrong answer it will just stop functioning altogehter


 No.2892508

>>2891932

I'll start with basics:

foreach (Citizen comrade in citizens.FindAll(c => c.ReadingList.Contains("Capital") && c.Bunker.Built == false)){
comrade.Bunker.Priority++;
}


 No.2892510

interested in computer aided direct democracy, what's a good book to start?


 No.2892511

>>2892510

Towards new socialism by cockshott

>>>/gnussr/545


 No.2893353

File: fed3e4fe925ae19⋯.jpg (89.43 KB, 736x924, 184:231, bear7.jpg)

>I used to record the bedtime stories I told my own children. This is one I did for my niece Ruby. The stories were made up as I told them, and hava a strong local Scottish character. This is a first attempt to put one out on YouTube. Not sure if the visual effects are needed, or whether simply a voice is best with the pictures left to the child's imagination.

https://www.youtube.com/watch?v=TM-tjs6XSuA

This is just adorable.


 No.2893378

>>2892296

>>Managers and innovation

I think managers don't have that much to do with innovation, actually. There's R&D, which managers are usually to dumb for unless they are also engineers, and when a manager who is also an engineer does develop a useful engineering idea I'd say it's mostly not due to management-specific knowledge. There are many ideas from the shop floor. The manager can listen to that and get a fat bonus, the person on the shop floor gets a t-shirt.

>>under the pressure to get quantitative results, managers often cut corners on quality

Measuring quantity instead of quality is a general problem, not a socialist one. How can you systematically improve and not just by luck if you don't even have a measure? Better scoring rules were developed – the score for the big goal has to be made of of scores for distinct goals which have required minimum thresholds, and what's above that is weighted and aggregated not using arithmetic mean, but geometric or harmonic mean (as proposed by Miroslav Toms and others) – but this is very annoying to do with pencil and paper and using snail mail or phone call. With spreadsheets on a computer reluctance to use something like that is much lower and plants can now also easily send the details about the composition of what goes into these aggregate scores. There is a huge change here towards higher efficiency for socialism without need for a new big-brained theory.

>>excessive unused inventory

A lot of that can be avoided by a change in the accounting rules. Just like with computers and the internet, we can "steal" from developments that have already happened in capitalism. When you take the historical cost of unused inventory as a given, it appears as wealth in your part of the system, whether it is over-supplied or outdated stuff, you then have some serious extra thinking to do to fix that, and a lot of that is going to be speculation about the future importance of the stuff. If instead you focus on throughput, inventory just looks like 'locked up money (or locked up labor time) which you try to transform into throughput. Inventory is a cost.

>>Improvements in production would lead only to an increase in one's production quota. In effect, well-run factories were punished with greater work loads.

This is muddled phrasing (probably also muddled thinking). A factory isn't a person. Which individual should be punished or rewarded for what? What are you responsible for? If there are good tools and bad tools and you work with good tools, you should be rewarded or punished based on what one can do with the good tools UNLESS choosing/creating those tools was your achievement. And if you were given these superior tools, why shouldn't you be expected to do better? Of course, places with better technology (whether the responsibility for having them lies there or not) should get higher production quotas, don't call that punishment. To the extent the higher productivity is due to decisions and hard work at that place, we can still reward that with promotion and very quick feedback with a remuneration bonus.

>>Poor performing ones were rewarded with lower quotas and state subsidies.

Lower production quota is sensible here. If the subsidies are earmarked for obtaining better technology, if the difference in productivity lies in that, that is sensible. This is not a kafkaesque situation. Remember we still got the tool of punishment by lowering salaries/demoting managers and workers, insofar as this lower productivity is due to them and not factors outside their responsibilities.


 No.2893379

>>2893378

>managers are usually too dumb

>_<


 No.2893393

>>2892178

My intention with multiple agencies isn't that different regions will do independent planning, but rather that there are multiple independent observers making sure planners are following the proper instructions. These observers can represent different regional or demographic interests and such, so that everyone is able to trust that the plan represents them as advertised.

If we tie central planning to a single state apparatus, I don't see how the state will ever be able to just whither away. For that to happen planning needs to exist as a free agreement between workers. They negotiate a set of abstract principles that constitute planning, and all these "planning agencies" then do is figure out what the concrete results of these principles are.

In other words, planning should be scientific. If we have just one planning agency we run the risk of it introducing all kinds of arbitrary determinations that do not properly represent the wishes of the working class. But if we have multiple agencies checking on each other, there will be a single equilibrium in which interests of the workers are best represented. Basically it's the equivalent of peer review.

I hope that makes sense. To me this seems like a very important point, although I have difficulty finding the right words for it.


 No.2893613

>>2893379

They are, you’ve obviously never worked as an engineer for a non technical manager


 No.2893661

>>2893613

It was a comment on my own typo.


 No.2893702

>>2893379

Management typically has more to do with being capable of handling people than anything else, unsurprisingly. That's not even close to what it takes for innovation and development.


 No.2893869

>>2891960

Looks like Cockshott just made a video on this https://www.youtube.com/watch?v=Mw_WQRPIBNU


 No.2893881

>>2891312

I dont want to live under the tyranny of statistics. Men should rule themselves, never bow before gods or machines


 No.2893888

>>2893881

>Not making man both the God and the machine

Lame


 No.2893905

>>2893888

Transhumanists get hunted and gathered.

Fear the anprim tribe


 No.2893940

File: c84eed291ccfae9⋯.png (22.37 KB, 890x628, 445:314, leftypol shitposter on yt.png)

I wonder how much Paul enjoys/detests some of the shitposts left as comments under his video. I'm sure he's glad us yung'uns are taking to socialism, but he's also an older man, and our rambunctiousness just might rub him the wrong way. All I'm saying is that pic related is definitely from our little board, possibly on this thread right now, and perhaps not best representing the young communist movement to older, more serious minded ones.


 No.2893941

>>2893905

enjoy dying at 30


 No.2893943

>>2893940

Yeah that comment looks fucking horrible there.


 No.2893949

>>2893940

Nice profile pic


 No.2893951

>>2893940

>I wonder how much Paul enjoys/detests some of the shitposts left as comments under his video

Detests. All of them. How is this even a question?

>he's also an older man

I'm sure he would have seen it the same way at 20. The big problem is how to grow the radical left. Doing subcultural injokes outside of subculture gatherings is a really bad idea, unless you really only care about your subculture and delight in how people in the mainstream (and even others of the far left) don't get you.


 No.2893967

>>2893949

Thanks, I made it myself

>>2893951

>>2893943

Well I'm glad yall agree with me. I'm sure there's dozens, maybe a hundred leftypolers watching his videos, and most of them stay quiet, but between comments like this (though I must say Install Gentoo has overall been respectful and fairly well composed, exception being this one comment, and thank god Paul probably doesn't know wtf a trap gf is), but shit like this sticks out like a sore thumb.

I think to some extent, he must understand that the younger generation is undoubtedly going to have some cultural and behavioral differences from his own generation, and there's something to be said for social misfits and weirdos being more likely to adopt a very unpopular political position, but there's no need to go over the top. So it's one thing having a profile pic of an anime boy, and it's another calling yourself "The Hunter x Hunter 2011 Dickriding Association". Must say though, THxH2011DRA doesn't post trashy comments, from what I recall, but one commenter on the early vids did invite Paul to /leftypol/ to, and I quote "shitpost with us". And finally, maybe it's time for me to put away that stirner/cheetos-cat OC for my pic, I made that a long time ago and am no longer particularly interested in Stirner.

-self crit over


 No.2893970

>>2893951

>Detests. All of them. How is this even a question?

Rhetorical question, but yeah.

>I'm sure he would have seen it the same way at 20. The big problem is how to grow the radical left. Doing subcultural injokes outside of subculture gatherings is a really bad idea, unless you really only care about your subculture and delight in how people in the mainstream (and even others of the far left) don't get you.

Totally agree, and few of our jokes aren't even designed as bait to gain interest in socialism. traps and nazbols are just shitpost material. Fuck, man. We need a new website bad. I really fucking appreciate the existence of /leftypol/ but we need to advance beyond being one board on a nazi/weeaboo/trap-lover's forum. It's good, in my mind, to keep this board here as long as possible, as a thorn in the side of the aut-right, and as a place for leftists to talk shit without ruining their image, but for constructive discussion and organization, a new format is required.


 No.2894071

>>2893940

It's always that one retard who is shitposting.


 No.2894648

>>2893941

'cave men' didn't die at 30.

what happened in ancient societies was high infant mortality, which drove life expectancy down.

https://aeon.co/ideas/think-everyone-died-young-in-ancient-societies-think-again


 No.2894663

>>2894648

Well, in that case, I am now an prim. Fuck you, babies.


 No.2894684

File: 1f1fd56bc834d60⋯.png (153.37 KB, 200x347, 200:347, e65.png)

>>2894071

Shitposting is infused into the DNA of my soul.


 No.2894686

>>2893941

I think most sane an-prims would go to the nearest city full of trans(sub)humanists for medical care beyond what our knowledge of the forest could provide.

I dont think anyone really believes the whole world will be on board with an an-prim lifestyle, even if it's the only way to save the earth. You can put computers in your brain and live in a robot skin suit, but I and at least a few others would like to live in the natural state that society has deprived us of for so long. Go out an colonize the galaxy in your tin cans and brain interfaces, but I want to live and die here as a natural human excelling at what I'm evolved to do. Ambition will not be my downfall




[Return][Go to top][Catalog][Nerve Center][Cancer][Post a Reply]
Delete Post [ ]
[]
[ / / / / / / / / / / / / / ] [ dir / choroy / dempart / klpmm / mde / randamu / tingles / trueb / vril ]