Why decentralisation and federation are the future of online media
Over the last decade or so, I’ve been commenting in fits and starts around a theme:
Centralisation of internet infrastructure and services is anti-democratic.
Starting with Iran’s internet blackout following the 2009 presidential elections, I have been less-than-methodically reviewing the impact of centralised networks on the internet’s ability to ‘treat censorship as damage, and route around it’, as Internet pioneer John Gilmore famously said.
The idea of a neutral, borderless network of networks where any point could connect to any other was a brilliant design. We can legitimately call it revolutionary. In its best case, decentralisation is democratisation.
The protocols that distribute our cat pics and right swipes are indiscriminate and able to go literally everywhere there’s a pipe. But who owns the pipes? When it’s a government, a state-controlled entity or a large-scale corporate player—or some combination of that—then the tendency will inevitably be toward control, whether for profit or political reasons.
Back in 2009, we didn’t really have the means to cope with that, and in Iran’s case, it provided the government with the ability to control the entire nation’s communications through a mere five control points.
At the time, Farhad Manjoo used this bitter experience to argue that this meant that the revolution will not be digitised. I responded that he was conflating the physical networks and management tools with the protocols themselves, which are actually far more permissive. In the end, it seemed to me that the best—and worst—we could expect was a mixed bag.
So no surprise then, that back then I was strongly anti-regulation.
I wasn’t against it because I didn’t see any merit in the idea. I was against it because when large companies are regulated by governments, what usually ensues is either regulatory capture (as happened in the US telecoms market) or state cooption (as witnessed in China, Iran, Ethiopia and countless other autocratic nations).
And that’s why I’m against most regulation of social media companies.
Most, but not all.
I think we would all benefit from a little monopoly-busting. Well, commoditisation, at least. Facebook, Twitter, Google & Youtube all skate around their fundamental responsibility not to break society to bits because they claim to be as borderless as the network protocols that underpin their service.
Walled gardens are dangerous. They create a disparity of information, in which the platform knows everything about us, and we know next to nothing about the platform, or how it works.
Allow me to quote at length from the 3000-word screed that inaugurated this substack:
In 2013, Neil M Richards wrote about The Dangers of Surveillance. At the heart of his principled cry for what he described as ‘intellectual privacy’ is the contention that it’s not data per se that is dangerous to an individual, but the disparity of access to it.
Our problem, the argument runs, is not that Facebook knows everything about us, but that we know next to nothing about Facebook. What exactly does it know? Who else is it telling? Why is this lie/ad/scam/rumour being shown to me right now? What information about me makes their algo think I’ll be particularly receptive to this so-called ‘sponsored content’?
If we knew the answers to these questions, it would at least make the task of unraveling the untruth a trifle more possible.
The contention that big data, the algorithms that shepherd it and the profits that derive are all private is bullshit.
The contention that the resources required to acquire data, put some shape on it, and make sense of it all spring up sui generis like mushrooms in Mario World is equally bullshit.
There is absolutely nothing in the world, bar a few private islands worth of wealth, that stops corporations from building their algos publicly, sharing exactly how they were built, down to the slightest minutiae, and still making a very large bundle of dosh. Better still, that bundle of dosh can be amicably and equitably shared among all the contributors to this wealth creation project in amounts commensurate to their contribution.
Socialism! Yes. Just don’t say it like it’s a bad thing.
The ‘how’ of big data algorithms is devilishly hard to understand, and hella expensive. The ‘what’, however is not. And people have a right to know it.
This is not the Secret Sauce. Or maybe it is, but who cares? It’s a secret sauce that can only be spread on a Very Large Dataset using Very Large Resources.
And it’s a Secret Sauce made for use on Us.
If algos and their impacts were public, companies would only have the size and composition of their dataset to compete with. And they would retain the allegiance of that data set (data is people!) by making it worth the data’s while.
That’s how it works right now, only people’s data is kept secret from them. The public don’t have any way to choose who to trust it with, and for what purpose. Those secrets must be revealed. It’s not the end of the world. It won’t bring the data giants crashing down.
It does mean they’d need to compete more, and compete on things we all can see, and measure.
That might be convenience. That might be services tailor-made to order, okay Google? That might be by demonstrably and consistently telling the truth (which would be pretty cool). That may mean straight up profit sharing. Who cares? We’re not prescribing, remember?
What I’m trying to say here is that the role of aggregator of data can be—and should be—separate from the role of content service provider. The aggregators need to be utterly transparent. The service providers need to be utterly responsible.
But you can’t be both. Because being both the one who gathers the data and the one who serves it up creates the situation we’ve got today, where straight-up ratfuckery and deadly collusion happen behind an utterly opaque and impenetrable barrier. Facebook has proven that it’s only willing to be as good as it has to be. And practically speaking, it profits more from being not good.
Right now, faced with the prospect of being forced to remunerate news agencies from which it derives immense revenues, Facebook’s best play is to pull a Murdoch. Sidle up to the Coalition government, remind them which side their bread could be buttered on, and leave those nasty Australian (and global) media to die on the vine.
The potential power that this represents is a fundamental threat to democracy and a rules-based order.
And there are few if any scenarios in which regulation doesn’t end up being coopted in one way or the other.
Unless that regulation required them either to localise their businesses and operate as a media organisation, or to contract the provision of their data on a neutral basis.
In the best of all possible worlds, your social media data would be as commodified, and as private, as your email. In other words highly imperfectly. But owned by and controllable by the parties who contributed to it.
It’s not impossible to achieve this. It’s just really really hard, because you’d be militating against an entire generation of Tech Money, and returning the internet to its hippy roots* and make the services that run on the internet look more like the decentralised protocols that made it all possible in the first place.
* The truth is that its decentralised design was the US Defence Department’s plan to make communications networks that could survive a nuclear ‘beheading attack’. They didn’t envision cooption from within. Which is why centralised services like Facebook, Youtube and Twitter and the rest are actually antithetical to the internet that made them possible.