Kosambi Circle meeting 9 notes

10 May 2020

1100 – 1320

Meeting notes

This was the ninth meeting of the KRAC done online via zoom, 104 people attended. This meeting started our May theme of discussing technology and its interactions with workers and society from the lens of socialist theory.

From the large list of readings to be covered in May, the ones discussed in this meeting were Marx's Wage Labour and Capital (as an understanding of wage labour is fundamental to understand why production of science and technology in society is waged work), then some pages from the notes of Grundrisse of Marx, (pages 690-698) on Fixed capital, Means of labour, Machine where Marx is talking about how humans and machines are used together. This was followed by discussing Kosambi's essay Science and Freedom where he talks about how science is a product of the class society it comes from and what it means to liberate it. Then we discussed Capitalism’s New Clothes by Evgeny Morozov which is a critique of Soshana Zuboff's category of “surveillance capitalism” and makes a strong argument that surveillance capitalism is not being prodded by some new logic of totalising surveillance but is coming out of the old capitalist logic of profit as always. Morozov's critique in itself is also a very well written description of the political economic debates around the topic and these were discussed in detail. After which we discussed Anupam Guha's Artificial Scapegoat which makes the argument that the blame of algorithmic bias and then the attempt to optimise bias reduction are both red herrings and shift the lens away from human actors in capitalism which makes algorithms used in the first place. And finally we discussed in parts Platform as a Factory by Moritz Altenried, scholarly work which goes into piece wages, crowd work, and microwork, their history, how the platform economy is amplifying precarity in capitalism etc.

We could not finish discussing this piece in detail on this meeting and continued the discussion on the next meeting on 22nd along with the pieces not covered in this meeting from the May reading list.

Before the meeting, some voice notes had been used to explain some of the material which can be found here: Notes on Technology

We started the meeting by discussing what are social relationships and how in Marxist analysis politics and economics are not separate and why. Then, we first went into Marx's Wage Labour and Capital, which were a series of lectures delivered by Marx to the German Workingmen's Club of Brussels in 1847. These were never really completed because the rebellions had started by that time, and Engels had to revise bits of it. One major correction was Engels making the distinction between labour and labour-power (the latter being the thing the capitalist hires out).

We went into the discussion into the distinctions here, what is labour? When you work and create something. The worker is not being paid for their labour, they're getting paid for their labour power. If you're a coder, the sum total of your labour is the creation of the program. But you're not being paid to create a program – you're being paid for a set of tasks that create a product for a certain period of time. To work for a certain amount of labour for a certain amount of time – and the additional value you created is what is generating profit which you are not being paid for. Thus, fundamentally to the Marxist framework, wage which results in profit is theft, profit is the unpaid bit.

The labour done in the time labour power is hired out for is more than the cost incurred by the capitalist to hire it. And this is the source of all value. So the piece begins with what are wages? Wages are the cost the capitalist incurs to buy labour power for a certain period of time. The exchange value of the commodity estimated in money is called its price. Wages are not a share of the worker in the commodity produced (this nuance is important), wages are only the means to buy labour power (which itself is a commodity in capitalism, it wasn't always because labour was not waged labour in feudalism or slave society). How is the price of the commodity determined? Over a period of time it is the cost of its production (in short term it is above or below). Now these fluctuations are important as the wages fluctuate as well, proportional to the cost of production of workers. Production of workers are dependent on social relations which change when the means of production change. Class society thus is the totality of the relations of production.

Capital thus is means of subsistence of workers + instruments of labour + raw materials + exchange value, all of this transformed via labour power

To obtain this labour power, the existence of a class of people which can provide it and only it, possessing nothing else but the ability to work, becomes necessary for capital. So profit has not risen because wages have fallen, the falling of wages is because of existence of profit. We also discussed how this system while increasing the division of labour also simplifies it till workers become replaceable. Marx also talks about the act of making workers superfluous protects this system and these workers empirically do not find new means of employment. Even in the most developed economies, the idea that the 'new economy' will create new jobs is not true. We discussed how thus the actual material distress of the working class when these job losses happen is mischaracterised as anxiety and technophobia, how the Luddites were mischaracterised etc. We also discussed here how a lot of this means social reforms are not possible unless you're in a social republic (which is an attempt to build a socialist republic) simply because the very logic of capitalist social relations are exploitative.

Here the discussed veered into how the means of labour modify under technology and the few pages of the Grundrisse which go into that were discussed. In short that bit talked about “objectified labour” which is the accumulation of knowledge and skill, the general productive forces of the social brain absorbed into capital as opposed to labour. Also the observation is made that hence, machinery which has manged t objectify labour does not help labour but reduces labour to a convenient number for the capitalist to negotiate with. We discussed who is actually doing the labour in tech monopolies, how workers are assisting in their own obsolescence and the dilution of collective bargaining, and how this implies that science has to be emancipated from capitalism for these productive forces to be used in benefit of the working class and not to trammel it. On that note we moved the discussion to Kosambi's essay, Science and Freedom.

Kosambi talks about that the scientist is not some lone innovator whose relation to the society is how “free” they are (which is the dominant discourse, namely freedom of thought) but is essentially an unfree worker as dependent on the wage system as any other worker (and in past scientists have been thralls of their class societies and any attempt to go opposite to the dominant class interests by scientists have resulted in retribution). The scientist depends on salary, stipend, grants etc which capitalists control. Kosambi goes into what exactly freedom is, freedom is the ability to solve problems, but capitalism constraints it to be what it considers useful (to capital). So in a capitalist society the grant will be given to scientists that serve the needs of the capitalist society. Thus Kosambi makes the argument that the freedom of wallet is as important as freedom of thought. And that the two are connected. Also we went into how this implies it is not the “language” of science which is elite and inaccessible, its that capitalism forces it away from you, you cannot spend your free time learning astrophysics because you have to pay rent and need a job.

At this point we started discussing Capitalism’s New Clothes by Evgeny Morozov. This work is a critique of Zuboff's book which presents her theory of Surveillance Capitalism, a theory Morozov contests. The theory goes like this, Zuboff argues that tech firms are not doing surveillance etc as a moral problem or malice etc. but are following a rational imperative, determined by trying to make a “behavioral surplus” to manipulate customer behaviour, and thus contesting that via privacy violation litigation is not enough. These companies use behavioral surplus to modify your behaviour in the hopes of selling you things – the famous line that you're the product is not enough, but you are the dead carcass after the information has been extracted. Thus Zuboff creates this category of surveillance capitalism wherein information is essentially the thing used for control and the customer becomes the fundamental unit of what the capitalist is interested in. Morozov's critique of this is quite extensive but boils down to this, that the class relations are still the same, this is still capitalism, and the unit the capitalist is exploiting is still the worker. The power these companies have is that they have newer ways to exploit labour while pretending they don't have labour. Zuboff's critique of technology lacks the complete angle of class, and his counter-argument goes thus. Throughout history, people have tried to explain capitalism in various ways – such as Chandler's managerial class, which revolutionalized the way capitalism works, similar to Zuboff's surveillance capitalism. The hegemony is not a consequence of surveillance, but the surveillance is the consequence of hegemony.

Thus Morozov is critiquing what essentially is an ethnographic approach and not a political economic one, which has ignored the logic of profit. He also compares these with past frameworks (like Negri's autonomists) which had similar premises but different conclusions, like the autonomists thought that the capitalists have stopped being capitalist and becomea renter parasite class, all workers need to do is reunite, and overthrow capitalists – which has also not happened because as Morozov argues the capitalism still remains, the workers are still being exploited. The surveillance is not the point, but profit.

We then quickly discussed The Artificial Scapegoat which takes the problem of technosolutionism and turns it on its head and tries to refocus the lens on human actors operating under capitalism, actual agents.

Technosolutionism is trying to solve everything with tech which have social origins but a simultaneous problem is blaming the problems on technology an its makers rather than political economic pressures as to why tech gets used (and at times giving the platforms more power in the name of “accountability”). Problems like censorship on social media, bigoted decisions etc are being attributed to algorithmic bias etc. This isn't completely true. Neither is it absolutely correct that the bias in machine earning systems are primarily because of epistemological frames of those who make it. Neither can calls for technological transparency solve it. The simpler and more correct lens is to look at where did the data came from, who paid for its collection, which context should it have been used in and why was technology used in a particular problem anyway. The answer is often not malice but capitalism, profit, convenience.

We discussed that instead of arguing on the standards of a piece of AI, should it really exist for a particular problem? Should facial recognition exist in policing? No. Its very existence is oppressive. Its existence means that someone's freedom is determined by non-deterministic machine learning, which is both a violation of rights and surrendering of democratic power to companies. Why it exists is because finance capital pushed that tech into that field and the State finds it cheap to use it, and convenient to slyly erode rights. We then discussed the current frameworks of AI governance using FAT (Fairness, Accountability, Transparency) and why that is completely lacking because it does not take into account Capital and its motivations, let alone trying to govern it in any meaningful way. Asking companies to make political decisions (in the name of accountability and moderation) is giving them power, which shouldn't be done. And this lack of looking at how money is not limited to technosolutionism, we also discussed legal solutionism as well, that you can solve policy issues by making legislation on artefacts, by enforcing standards, without making legislations on power of companies, without looking at finance capital. The class nature of the state is ultimately something you cannot run away from. At this point we had started talking about data collection etc for AI which was a pivot into the paper by Moritz Altenried, Platform as a Factory.

We quickly discussed the main points as we were running out of time. The work is about crowdwork, microwork, and piece wages. About how digital Taylorism (Taylorism is about scientific management of labour, making it as mechanised, atomised, and as quantified as possible) is standardising, surveilling, and controlling the crowdworkers. About how work is being increasingly atomised, crowdworkers replacable, and we are reaching that point of capitalism Marx was talking about in the Grundrisse. About, the erasure of what is work, to the extent where it isn't popularly considered work anymore. We decided to discuss this piece in detail in the next meeting. We discussed a lot of issues here (like is consumption work, the consensus was no, also is creation of data itself work?, how will scaling and innovation happen in socialistic systems when surplus is not being accrued?, questions on liability, about concrete problems like autonomous cars and researchers trying to assign agency and morality to machines, and other questions). Here we run out of time, the meeting ends, and we decide to discuss from here the next time.