The new wave of decentralisation: a look at Anytype

Don't mention the crypto

If the Computing inbox is anything to go by, we are witnessing a new wave of decentralised applications and networks.

Previous waves have been closely bound to crypto prices. First there was bitcoin and blockchain, 99.9% of whose breathlessly promoted use cases failed to float. Even the serious ones floundered. Witness Maersk's recent shelving of its pioneering supply chain ledger TradeLens.

After that came smart contracts, and second layer networks, which speeded up the process of transacting tokens - and indeed lifted the market interest in those tokens—but again the applications that were meant to usher in a new and better web never arrived.

Hacks and scandals came and went until the next big wave arose. Topped by the froth of NFTs, hyped by hip hop stars, pumped on prime-time TV, this one crashed in spectacular fashion, taking a hefty chunk of the crypto market with it, with the collapse of FTX and Sam Bankman-Fried's arrest on fraud charges. Sotherby's and Paris Hilton are currently being sued by Bored Apes buyers who never were admitted into that fabled yacht club. So it goes.

Given all that, are we really ready for another wave?

The answer is a qualified yes. Potentially, at least, decentralised infrastructure is uniquely placed to tackle some rather urgent problems. Local control of data, security, privacy, self sovereign identity, censorship resistance and global redundancy are more pertinent than ever, particularly in view of big-tech's AI-driven scraping frenzy. If knowledge is power, that power is being centralised in a few very large models held in even fewer hands. For an illustration of how the concentration of power can play out, 31-year-old Bankman-Fried was planning to buy a country. Had he succeeded, he may have avoided arrest.

Token? What token

So, tackling tech centralisation issue is a worthy and important cause. The problem is, it's very hard to do. The likes of Tim Berners-Lee have been preoccupied with getting the technology to work for a decade or more.

Another issue is crypto itself. Once a handy way to raise funds unbeholden to VC timescales, cryptocurrencies soon became catnip to exactly the sort of moggies you really don't want hanging around your door. Countless projects were derailed through misaligned incentives; many were nothing but hype to start with. Then the regulators started circling, the US SEC in particular, judging anything not a pure utility token to be a security and thus subject to financial laws.

It's interesting that recent announcements in the space, those that get through Computing's spam filter anyway, have not mentioned the C-word at all. No doubt there is a desire to stay under the regulators' radar, but many projects just seem keen to disassociate themselves from the craziness of the lambo-moon crowd.

Veild, a privacy-oriented peer-to-peer decentralised network unveiled at DefCon this month, is a case in point.

"Does Veilid have a cryptocurrency?" Poses its FAQ. "Heck no. Veilid does not have a cryptocurrency."

However, some sort of cryptocurrency may be essential for a fully decentralised network - one not controlled by any company or cartel. It will be interesting to see how the team behind Veilid provide an incentive for node operators and a disincentive for would-be attackers in the absence of a token.

Anytype - application first

Anton Pronkin, Berlin-based cofounder of Anytype, concedes that at some stage his project will need to consider incentives.

"I think it's hard to build a purely decentralised network without a token," he said.

However, for now the project, which has just landed $13.4 million in VC funding, is focused on the missing pieces of the jigsaw - the applications and user experience that have so far proved elusive.

People use products, not protocols - Anton Pronkin, Anytype

In its blurb, Anytype is described as an "everything app", a multi-platform, local-first, no-code, drag-and-drop, web design, knowledge management, planner, wiki... you name it.

"People use products, not protocols," said Pronkin. "And if you want to bring this mission, to empower sustainable cooperation by building tools for freedom and trust, we need to build a killer app."

Currently in beta, the Anytype app is based on a graph database, allowing connections to be made between objects, which could be tasks, notes, ideas, documents, workflows or other things. It allows them to be searchable in different ways, and visualised as tables, Kanban or clustering.

The app is single-user only for the moment, and free to use. However, the lofty long term goal is "to enable large-scale collaboration to create a global repository of interconnected knowledge," by connecting devices together to form an end-to-end encrypted network with collaboration capabilities. At that stage it will move to a subscription model.

Anytype's peer-to-peer networking protocol Anysync does not use a blockchain. It is based on InterPlanetary Linked Data (IPLD), part of the protocol developed by IPFS, but with some modifications. It also uses dRPC, a decentralised modification to gRPC created by the Storj project for internode communications, and mDNS for local peer discovery. Global peer discovery is a work in progress.

Data on local machines can be synched with IPFS or Storj or potentially any other content-addressable back end for replication, but Pronkin insists the new network should be faster, as those systems are not designed for real-time data transfer. "It will have cloud parity in terms of the speed," he claimed.

However, it will require a "trust layer" to deal with spam and malicious or poorly performing nodes. This will be a ledger which different providers of storage space or other services can use to connect to different "spaces" (nodes controlled by end users).

The onus will be on the end users to choose reliable providers based on their track record or reputation, and on the providers to prove their reliability and performance. Punishment for bad behaviour? Almost certainly a loss of staked tokens, but this is still on the drawing board.

Currently, said Pronkin, the closest competitors are probably Urbit or Ceramic Network, but his team has its eye on more mainstream applications like Discourse and Reddit, allowing communities to own their shared space while still benefiting from the network effect.

"We are going towards the direction of community management and knowledge management in communities. This is our niche. We want to create the best community software in the world with great user experience and real ownership."

Investors, which include Balderton Capital, Inflection and Square One, agree with this vision and are in it for the long haul, he said. "They expect us to be the market leader in this new, emerging, Web 3 ecosystem of applications for users who don't care about the token price."

The Anytype app is certainly very slick and well-designed, a nice change from the usual clunky fare in this space. But let's face it, the front end is the easy bit.

Whether this wave of decentralisation delivers will depend on whether the back end can be made fast, flexible, reliable and secure; whether the internal economy delivers the right balance of incentives and protections without being sidetracked by coin price; how it is viewed by regulators; and how adaptable it proves in the face of changing demands.

That's the hard bit.

We watch this space with renewed interest.