The Pathfinders Column from the December 2017 issue of the Socialist Standard
"Blockchains could change our world as much over the next two decades as the internet has over the last two" - source: zdnet.com - https://tinyurl.com/mp372v5
The thing about revolutions is that it's not always obvious when you're in one. They only reveal themselves in hindsight. Now some are saying there's a new revolution on the horizon that will be as huge as the internet. They claim it will change capitalism, see banks disappear, even abolish global poverty. When we last mentioned this development (Pathfinders, October 2015) investment in it stood at around $360 million. Today it's close to $2 billion, and this may be only the trickle before the torrent (ft.com - tinyurl.com/yb5zpdeh).
This is the world of the blockchain, and it has implications for socialists too. To understand it though, it's worth understanding something about networks.
Computer networks have in the past followed a centralised model where clients are individual computers communicating via a central server. This client-server structure dates from the early days when computers were the size of basements and operated via 'dumb' terminals capable only of basic input and screen display. Even when terminals got smarter and became PCs, this structure was inherited, and many businesses still use server systems today.
But there are problems with cost and scalability. The network can only be as big and as multitasking as the server can handle. The bigger the network, the bigger the server, the bigger the costs, and the bigger the risks of catastrophic breakdown if something goes wrong. And while the server does the heavy lifting, today's smart PCs are still behaving essentially like dumb terminals.
Consider for a moment an obvious analogy with the state, and centralised state institutions, or any centralised organisational structure. If one applies a top-down exploded view, every hierarchy looks like this. Such client-server structures are historical legacies which remain universal in the capitalist mindset, yet many of the same problems of cost, scalability and risk apply. In addition, these structures are monolithic and unadaptable, and despite massive social and educational advances, smart workers are still required to behave essentially like dumb terminals.
In computing, a new kind of structure, the peer-to-peer (P2P) network, harnesses the power of modern PCs by taking the central server out of the equation. Instead files or bits of files are held on multiple distributed computers, or nodes, and can be disseminated directly to any other node independently of other network operations. Having multiple nodes means parallel processing with no bottlenecks, and it's harder to break, because if a node fails alternative routes exist. P2P is therefore faster, cheaper, more scalable and more robust than client-server systems. But is it more error-prone?
In P2P file-sharing networks, multiple copies of the same data are an advantage. But P2P is now running crypto-currencies like Bitcoin, increasingly popular because unlike bank-mediated digital money the transactions are untrackable. Clearly, multiple copies of the same money cannot be allowed (the so-called 'double-spend' problem), so with no central control or validation, and in an anonymous public network where trust cannot be assumed, what prevents Bitcoin inflating and collapsing in chaos?
Enter the blockchain. Strictly speaking, 'blockchain' is the specific Bitcoin application of a thing called Distributed Ledger Technology (DLT), but as Hoover came to mean 'vacuum cleaner', blockchain is now being used to describe any DLT application.
When you make a Bitcoin transaction, the details are distributed across the entire network. To be sure the transaction is unique (ie not a 'double spend') it must be validated. To do this, the system triggers a competition in which freelance 'miners', acting somewhat like accountants, race to validate the transaction in return for a diminishing new-issue Bitcoin payment, which also helps to grow the currency at a controlled rate. Once validated, the transaction is then written into an encrypted public ledger as a permanent record or 'block'. This block is linked to previous blocks and in turn becomes the anchor or link to the next created block, forming an unbroken chain. Any subsequent attempt to tamper with an individual block disturbs the whole chain and results in a network-wide alert. 51 percent of the network, acting in concert, is enough to prevent interference. In plain terms, you can't buy product X on Wednesday and then pretend you didn't buy it on Thursday, at least not unless a network majority allows you to. System integrity is thus maintained, not by central state or bank control but by what could be called a distributed democracy. Barring a direct and unprecedented hack of the block-creating code itself, it's hard to see a weak point in the system. The strong point is that it offers to cut out all the financial middlemen in capitalist commerce. Business gains would be spectacular, which is why investors are throwing money at this.
DLT can in theory be applied to any field where data validation, transparency and integrity are important. Think big and local government, supply lines, transport systems, food quality and provenance, voting procedures, carbon trading, maybe even accreditation of news stories to prevent fake news.
The truth is, nobody is really sure what it can do, and this has provoked some reckless hyperbole. For example, the claim about abolishing global poverty is patently ridiculous. As tends to happen with emergent technologies, DLT is at the centre of a hype storm while still barely developed and little understood even by its investors. Aside from Bitcoin, no blockchain system has progressed beyond pilots and beta tests, although there are more than 400 start-ups. Small wonder some pundits are now saying that it has already reached the peak of expectations and is about to freefall into the 'trough of disillusionment' with its investment bubble bursting.
Yet there remain implications for socialists. DLT suggests a mechanism for a dynamic and decentralised model of socialist democracy and production which avoids the 'double-give' problem in distribution while offering a flexibility and adaptability not associated with traditional centre-periphery structures. More immediately, it could change how workers today understand the word 'organisation'. In the same way that a future subscription-based capitalism could alter mindsets over the need for money (Pathfinders, October), DLT shows how you can decouple regulatory oversight from centralised authority. If you don't need a state to ensure that things work properly, but can utilise the 'power of crowds', then an important prop in capitalist ideology is kicked away. Leaders and centralised elite structures engender cronyism, corruption and monolithic thinking yet many workers remain wedded to the supposed need for them, convinced that anything else would result in chaos. DLT may make them think again by showing them the power and flexibility of distributed democracy in action, and not just in socialist theory.
PJS
The IT industry has another example that might be of interest to socialists; the Agile development methodology. Instead of the "waterfall" model, where a senior person decrees that a product be developed, hands down work to developers who code it, who then hand it down when finished to be tested and finally deployed, Agile (sometimes also known as 'Scrum') relies on self-organisation.
ReplyDeleteIn the Agile model, a 'Product owner' commissions the product. The team decides for itself which features need to be developed - this becomes a 'product backlog'. They then select which items to work on, develop and test them, and then deliver finished items which over time assemble into a whole. There is no manager as such; a "scrum master" facilitates the work of the team, intervenes when external impediments crop up, and generally fixes problems leaving the team to do the development and testing work.
Work is divided into "sprints". The team decides for itself which items from the backlog can be developed during the Sprint to make a stand-alone deliverable. Progress is monitored daily towards delivery at the end of the Sprint (usually two or four weeks). At the end of the sprint, what has been completed is demonstrated to the Product Owner, signed off and then the team sits down to decide what they will do next. They continue in this cyclical way until the project is completed.
Agile encourages collaborative working; teams' self-management can embrace personal development as well as delivery. It is (in its ultimate form) self-directed, autonomous and non-hierarchical. To me, the parallels are obvious; yet I'm surprised that no-one seems to have noticed this.