The web31 discourse has been floating around in tech circles for about two months and I’ve noticed it spilling out into broader discussion, usually accompanied by the questions ‘what is this thing and why are people talking about it?’
Another less useful answer is that this is a way of marking epochs. Web 1 lasted from the mid-90s to around 2006, Web 2 took over from then until the late 2010s, now the web is moving to a new historical phase and need a label to mark the change.
Beyond that, web3 is a concept, a new web which embraces crypto technologies and promises ato be truly decentralised. There’s a lot of enthusiasm for people doing cool things with networks, and there’s also a lot of hype from people who think this is an opportunity to make some money. Web3 causes significant confusion as it’s sometimes difficult to distinguish between excited hackers and excited investors.
Various people have picked up on the dual character of web3. On the bright side, web3 promises to fix a lot of the problems with the modern web, and on the other side we could bring about a web “littered with tokenised microtransactions seeking rent on all possible user actions”. Between these competing visions, the coherency of the concept has been brought into question. Maybe underneath the hype there’s just nothing there.
If I were prone to making grand statements about the future of the web, my most likely prediction is that the World Wide Web will continue to exist in its current form and nothing will fundamentally change in the next few years.
Nonetheless, web3 exists and it has nurtured a movement which is exciting enough to get people talking about it. To begin with, there is a widespread feeling that the web has gone wrong somewhere.
The internet is broken
Many essays on web3 start from an acknowledgement that the modern web is riddled with problems. By sketching out an alternative, web3 has created a space for comparison where all frustrations come to the surface.
- Not enough links between sites.
- Content is disrupted by excessive advertising.
- Little respect for user privacy.
- Websites are slow and bloated.
- Search algorithms don’t highlight quality content.
- Websites are boring and impersonal.
- Content is hidden by paywalls and login barriers.
- Social media makes people unhappy.
Not all of these things are automatically fixed by decentralisation, but taken as a whole all these problems make web3 more attractive as a potential alternative.
Dead Internet theory
While carefully stepping through the minefield of deeply concerning nonsense associated with this conspiracy theory, the following passage nonetheless rings true:
The Internet feels empty and devoid of people. It is also devoid of content. Compared to the Internet of say 2007 (and beyond) the Internet of today is entirely sterile. There is nowhere to go and nothing to do, see, read or experience anymore. […] Yes, the Internet may seem gigantic, but it’s like a hot air balloon with nothing inside.
The web is a gigantic ‘library of knowledge’, it holds millions of pages all linked together. And yet, surfing the net today often involves just clicking back and forth between the same handful of big sites. Almost everything I see is in english. Against this homogeneity, the advocates of decentralisation promise a return to the network.
The web is supposed to be resistant to censorship, which is mostly true, barring a few exceptions which could be solved by decentralisation. I’m not interested in going into cancel culture here, or the content moderation exercised by social media platforms. If you got banned from instagram for sending people photos of your penis that’s your problem.
The exceptions which do interest me are the worst-case scenarios, where your adversary could be a state with control over all the physical infrastructure in the country. For example, when a country with a big protest movement suddenly experiences an internet blackout. This kind of scenario might not be a credible problem in Britain, but it happens often enough in other situations that it’s now a pattern.
What about running an illegal website, such as Wikileaks. Your domain can be blocked by ISPs, removed from Google, or suspended by the registrar. In the case of The Pirate Bay, police just turned up and took away their servers. What do you do if the state wants to completely erase you from the web?
Besides shenanigans from state security agencies, sometimes thousands of websites are taken offline by accidents. Last year there was a ‘global internet outage’ when a big CDN provider went down. The island of Tonga was taken offline due to a volcanic eruption AWS also had a big outage last December. All of this is to say that the web is far more fragile than you might expect.
Decentralisation can give sites a better chance of surviving in these most extreme cases. Communications projects such as Scuttlebutt and Briar are designed to work without a central server, they account for situations where the internet isn’t available. The project to decentralise Sci-Hub and Library Genesis relied on torrents and IPFS to secure the pirate archive, and now it’s much more difficult to take offline. Of course putting everything on a blockchain comes with its own legitimate privacy concerns, but if you did need to keep a site online at all costs, the decentralised web provides a solution.
Let a thousand protocols bloom
There’s a separate debate about whether it’s even worth pursuing new net protocols. If the intention is to create an alternative to the modern web, is it necessary to start again from scratch?
I recently set up a version of this site on gemini, along with a friend who created a personal gemsite of his own. Gemini is something we both find very exciting, but my friend ended up wavering on whether it was entirely necessary to set up a new protocol. Why write a whole new server and client architecture when you can already send simple markdown files over HTTP.
On the other hand, new protocols have their own strengths and tradeoffs. Here are some exciting protocol combinations to consider:
- Gopher or Gemini served through Tor or i2p.2
- IPFS or torrents distributed through mesh networks.
- Entirely offline anonymous file sharing through dead drops.
While I don’t use Brave browser, I admire its commitment to protocol plurality, that’s the direction browsers should be going for. HTTP may be the dominant method of downloading web pages, but it shouldn’t be the only method.
A server in every shed
The last argument around decentralisation entails the practical difficulty of getting more people to run servers. This problem is simple: in order to successfully decentralise, you need to add more nodes to the network.3
The cost associated with running a server used to be prohibitively expensive, particularly the point at which people were investing in multiple racks. In the last decade, we’ve seen the introduction of single board computers; these are cheap, small, power-efficient, and can run most web applications. The cost of running a server is no longer a problem.
This leads to another question, if the cost is reasonable, why aren’t more people running their own servers? Moxie points out the difficult answer, that most people actually don’t want to be involved in server maintenance. Moxie is correct in one sense; given a choice between hosting your own web services and just getting an account on some platform, most people would go with the convenient option.
Yet, despite the technical hassle involved, there is still a small minority of people who really want to run their own servers. I count myself in this odd group, speaking personally I do like having a server to run my own services on, and to some extent server admin becomes a hobby of its own.
For the rest, people have their own motivations. I know someone who posts photos on instagram, I don’t have an instagram account, but I still want to see their photos. Thankfully, there are several dedicated individuals who generously host public bibliogram instances to solve this exact problem. The web I love is kept going by these kinds of people who readily dedicate their own time and money to free initiatives.
Some platforms have taken a keen interest in the cryptocurrency potential of web3, without the same interest in decentralisation. Reddit likes web3 because they want you to invest in their special currency, they don’t want you to be able to run your own Reddit mirror. Twitter is happy to let you buy NFT profile pictures; they’re not going to help you to run your own distributed microblogging service unless it somehow benefits them.
There is significant excitement around web3 coming from companies whose business models are built on monopoly control of a platform. This inconvenient contradiction leads back to the contrasting ideas around what web3 is really about.
I signed the web0 manifesto, pitched as a counter-proposal to web3.
Web0 is web3 without all the corporate right-libertarian Silicon Valley bullshit.
The decentralised web cannot be both a financial instrument and a tool for breaking up the corporate web.
Here’s a binary table4 for comparison:
|crypto speculation||sharing culture|
|NFT trading||digital sovereignty|
|distributed apps||personal websites|
|fintech opportunities||fun experiments|
Web zero retains the driving idea that the web can (should) be improved, while maintaining a critique of systems based on financial speculation. The web is made for people, it’s not there to make money for investors.
Taking a more holistic view, the pitfalls of web3 just continue an existing trend for corporate control. The web today is dominated by US tech giants worth billions (trillions?) of dollars. It’s difficult to imagine a future without Facebook,
The decentralised web is already technologically feasible in terms of software and infrastructure. The more fundamental problem which remains to tackle is down to ‘culture, not protocols’. Or in the words of Peter Sunde:
Stop trying to fix the web instead of fixing society.
Or w3b, pronounced ‘weethreebee’. ↩
Has anyone ever tried this? It seems like a cool experiment. ↩
This isn’t easy to study, but are there any surveys of how many people run personal servers today compared to, say, 2004? ↩
Keeping in mind these are broad stereotypes, not hard definitions one way or another. ↩