πŸ”­ Cosmos Goes Live β€” Issue No. 42


This week marked the mainnet launch of the Cosmos Hub, the first piece of the broader Cosmos ecosystem its creators hope to enable. This launch comes after years of work by the team that created it, and a high profile ICO in 2017 that raised more than $16M. The Hub itself is essentially a Proof-of-Stake blockchain powered by a new consensus algorithm called Tendermint (more on this soon). Unlike other chains, though, Cosmos' raison d'Γͺtre is not limited to its own network-- the goal of Cosmos is to enable interoperability between all blockchains. Link.

The Cosmos team has designed a novel architecture that enables blockchains to communicate. Essentially, each child chain's node software also contains a light client implementation for the Hub chain. This means the Hub can verify transfers of arbitrary tokens from the chains connected to it and vice versa. To scale beyond a small number of chains, Hubs themselves can also communicate, allowing activity across "Zones" through multiple transaction hops. What about existing blockchains that don't implement the Cosmos spec? Chain specific "peg chains" can be developed for these. The peg chain communicates with the hub, and with a smart contract deployed on the other blockchain. When funds are locked in the smart contract, the peg chain releases them on the Hub. Pegs for Ethereum and Bitcoin are already implemented. Link.

Just in case connecting all the world's blockchains wasn't ambitious enough, the Cosmos project is also innovating along another axis: consensus algorithms. The Cosmos Hub uses a new Proof-of-Stake algorithm called Tendermint. Under Tedermint, regular users can delegate their stake to validators whose block creating power is proportional to their total stake. This might sound similar to the Delegated-Proof-of-Stake system used by networks like EOS, but it's actually quite different. Unlike EOS, validators also stake their own funds, and validators can have their funds slashed if they behave badly. The algorithm builds on decades of academic research on distributed systems and uses slashing to layer an incentive mechanism on top. In doing so, it aims to create a system that is resistant to both the "nothing at stake" problem (seen on chains with naive PoS implementations) and the problem of validator collusion (which plagues DPoS systems like EOS). Link.

All the research in the world can't predict how the system will hold up in use, though, and I will personally remain a healthy skeptic of all PoS systems until one proves truly robust. The Cosmos team has tried to de-risk this new algorithm by hosting a year long "Game of Stakes", wherein participants were encouraged to try to collude with each other on semi-private test networks in exchange for rewards. This is a smart idea, and it uncovered some bugs which were fixed. Even if the system works as designed, though, it still has real tradeoffs compared to Proof-of-Work. For one, it only takes 33% of stake to attack the network, and 33% attacks can actually halt the network completely for long periods of time. Additionally, the system will have only one hundred validators. If a third of them go offline, the network also halts. This feels much less robust to a coordinated attack than PoW mining. Link.

I can't help but be impressed by the incredible ambition of the Cosmos team and the fact they've largely delivered. The sheer volume of high quality code produced is remarkable, as is the fundamental research that undergirds it. Still, while the system seems promising and the team legitimately talented, there's no guarantee it will succeed. It may fail due to unseen flaws, or it may just be too late -- or too early -- to capture the market's imagination. Regardless, it's a noteworthy example of the incredibly intelligent and talented people who are resolutely investing their careers in this space.

I also want to make a meta point along these lines. Because I publish this newsletter, run a meetup, and provide software consulting around crypto, I try my hardest to stay on top of the major technical developments. I have to tell you: it's just impossible. There is so much incredible activity going on all the time. Literally every week, I discover a new project or tool I could spend a month diving into. There aren't enough months!

I know at times it can be disheartening to see the scams, the overhyped use-cases, and the way "blockchain" has become yet another meaningless corporate buzzword. And all this while so many fundamental challenges are yet to be resolved! It's easy to wonder whether anything real is ever going to emerge from all this. Believe me when I tell you: crypto is teeming with smart, creative people who are heads down building amazing things. Far below the canopy, in the deep shade of the rainforest floor, there is a vibrant ecosystem of passionate builders and tinkerers. I wouldn't bet against them.


3. The minimum number of validators who, at the time of this writing, could combine to have enough staking power to halt the Cosmos network. I've said a lot of nice things about Cosmos in this edition, but that statistic is worrying. For Cosmos to be viable in the long run, I think we'll have to see the staking power become far more distributed. Link.