Cosmos History — Part I — Inception to PreLaunch

The constellation of core Cosmos and Interchain development organizations has a complex history. Its details are scattered across the internet, in blog posts, news articles, tweets, github repositories, and even blockchains. What has emerged out of this history is a truly remarkable collection of smart, passionate, and thoughtful contributors that have together made manifest the vision of the Internet of Blockchains. This post outlines the origin story of Cosmos up until just before the launch of the Cosmos Hub. A subsequent post will outline the history since then. As the Cosmos ecosystem continues to expand and flourish, we believe it’s important to reflect on how we got here to remind us of where we want to go and how we bring the learning from the past into the opportunities of the future.

Pre History

One could say that blockchains became too big to ignore once Bitcoin crossed the $1B valuation mark in early 2013 during the Cypriot financial crisis. By the end of the year, Bitcoin’s market cap had 10x’d again, before famously crashing and collapsing the Mt Gox exchange shortly thereafter.

There were already adaptations of Bitcoin — like Litecoin, which launched in 2011, and Dogecoin, which launched by the end of 2013¹ — and there was a significant amount of protocol design activity and experiments. There were hundreds of cryptocurrencies, many of which never reached widespread adoption. But it was difficult, expensive, and unsustainable to continuously secure new networks using Proof of Work, and the early crypto community was researching alternatives. By 2012 and 2013 the idea of Proof of Stake was floating around, but the early Proof of Stake protocols — like those used by Peercoin and Nxt — had quite limited security guarantees because there was no cost to re-writing history. This came to be known as the “Nothing at Stake” attack and was considered a fundamental problem for Proof of Stake. The core problem remains, though it is now handled in practice by adjusting the security model.²

2014: Jae Kwon invents Tendermint

2014 was a formative year for Proof of Stake and blockchain more generally. By early 2014, Vitalik had invented Ethereum and the first implementations were being worked on. He was also publishing a lot about protocol design, and conceived an approach to handling the Nothing at Stake problem using security deposits — what he called “slasher”. In this version of Proof of Stake, validators would “bond” their tokens in such a way that they could be “slashed” if they misbehaved (ie. tried to re-write history illegally).

Inspired by the ongoing research on Proof of Stake, by the spring of 2014, Jae Kwon invented Tendermint. It was unique in that it was the first blockchain system to:

• Use a classical Byzantine Fault Tolerant (BFT) consensus algorithm
• Implement Security-Deposit-Based Proof-of-Stake (“slasher”)

Both of these components are now staples of nearly every modern Proof of Stake system.

The original PoS systems, before Tendermint, didn’t actually solve the consensus problem securely. Tendermint was the first system to build on the decades of academic research in Byzantine Fault Tolerant (BFT) consensus, in particular by developing a variant of the 1988 DLS algorithm for partial synchrony that “upgraded” it to use blockchain data structures and a stake weighted voting power. By combining this approach to BFT with Vitalik’s “slasher” idea, Tendermint came to both epitomize a new model of Proof-of-Stake and provide the first serious implementation you could really start to build with.

By the end of 2014, a growing community of cryptocurrency researchers had come together online in the form of the CryptoCurrency Research Group (CCRG, now defunct)³, largely convinced that Proof of Stake cryptocurrencies were the future of blockchains. They set off to build that future. Among them were Vitalik Buterin, Vlad Zamfir, Ethan Buchman, Jae Kwon, Zaki Manian, Dominic Williams, Arthur Brietman, and others. Vitalik had invented Ethereum, and Vlad had been working as a leading researcher there. Ethan worked closely with Vlad and was contributing to go-ethereum. Jae Kwon had invented Tendermint. Dominic was working on the ideas that would become Dfinity. Arthur had invented Tezos. Zaki was running a supply-chain focused blockchain startup called SkuChain at the time, and was otherwise already everywhere.

Jae Kwon had begun an initial implementation of the Tendermint software in Go around spring 2014 under his company All in Bits Inc, which did business as Tendermint Inc. He chose Go because the language was gaining popularity, especially for distributed systems projects and blockchains. What would become the main Ethereum software was also being written in Go. At the same time, Ethan was working for Eris Industries (now Monax) as the lead blockchain developer, with a mandate to bring Ethereum to enterprise, and was looking for a Proof of Stake solution to integrate with. Tendermint was the leading candidate.

2015: Ethan Buchman Joins

In early 2015, the CCRG held a conference in Silicon Valley called CryptoEconomicon where much of this early R&D community attended. It was here that Ethan and Jae, the founders of Cosmos, first met. They discussed possible attacks on Jae’s Tendermint algorithm, and shortly thereafter began closely collaborating on the Tendermint software.

Ethan and Jae spent the better part of 2015 working closely on the Tendermint software. They implemented the Ethereum Virtual Machine as part of the Tendermint application state (technically the first iteration of “Ethermint”), implemented various features to make Tendermint more useful, and made various improvements to the consensus algorithm.

Ethan was also completing a Masters Degree at the University of Guelph at the time. His thesis, Tendermint: Byzantine Fault Tolerance in the Age of Blockchains, focused on Tendermint — its consensus algorithm, software architecture, and performance under test. The thesis is widely regarded as an essential introduction to the historical context of blockchains and Proof-of-Stake in the tradition of classical Byzantine Fault Tolerance.

In the second half of 2015, discussions between Jae and Eris Industries over copyright of the Tendermint codebase (which included an implementation of the Ethereum Virtual Machine), led Jae to introduce an abstraction between the Tendermint consensus engine and the application state machine it replicates, a programming interface we now know as the Application Blockchain Interface (ABCI). This interface made it possible for application logic to be completely separated from the Tendermint consensus engine, to be run in a different process, and even written in a different programming language,⁴ making Tendermint the general purpose state-machine replication engine it has become today. This allowed the components related to the Ethereum Virtual Machine, and to any cryptocurrency elements at all, to be refactored out of the Tendermint code base. In particular, at the time, ownership over the Ethereum Virtual Machine components moved to Eris Industries, and that code evolved into what is now the Hyperledger Burrow Project.

By the end of the year, Ethan decided to leave Eris Industries to build a company around the Tendermint codebase with Jae. He became a formal co-founder of All in Bits Inc and served as its CTO.⁵

2016: Cosmos and Adoption

In early 2016, All in Bits (AiB) was trying to raise money from venture capitalists for Tendermint-based enterprise products and Proof-of-Stake application hosting. The VCs were only just starting to understand Bitcoin, and were certainly not ready for this. During this period, Ethan and Jae started working with Peng Zhong, a designer and web developer that Jae had worked with on a previous project. Peng started working on early blockchain explorers and wallet designs to support possible products. They realized after a few months that the timing was not right, and so Jae and Ethan revisited their original motivations: to solve a number of outstanding problems in the security, speed, flexibility, and usability of public cryptocurrency systems, to lay a foundation for a more sustainable approach to money and finance.

After iterating through a number of designs and debating different ways to raise capital, they hit on the concept of Cosmos and wrote the Cosmos whitepaper, which proposed an “Internet of Blockchains’’, based on the Tendermint consensus engine, the ABCI application interface, and a TCP-like protocol for reliable communication between untrusting blockchain networks called the InterBlockchain Communication (IBC) protocol. Jae had written the very first specification of the IBC protocol, which was included in the whitepaper. The first blockchain in the network was to be known as the Cosmos Hub; it would be a Proof-of-Stake cryptocurrency system with a native token called ATOM.

The Cosmos Whitepaper was well received by the cryptocurrency community, and Tendermint was gaining adoption as a flexible and mature platform for building blockchain applications. By the end of summer 2016, Ethan and Jae received some initial funding to build Cosmos.

Meanwhile, All in Bits was engaged by Votem Inc to host the Rock N’ Roll Hall of Fame’s annual vote on a Tendermint blockchain, which it completed in the fall. That means the 2017 inductees into the Rock N’ Roll hall of fame (which included Tupac!) were actually voted for using a Tendermint blockchain, lending more credibility to the project and helping further boost adoption. (See the Votem press release for more details.)

The user base of Tendermint was growing considerably at the time, and the Cosmos project was gaining in recognition. Cosmos won an award for Most Innovative Project at the International Blockchain Week in Shanghai, September 2016, which was co-hosted with Ethereum’s Devcon II.

By the end of the year, Jae and Ethan were hiring their first developers at All in Bits. A list of employees and when they joined can be found in the Appendix. The list is notably incomplete and will be updated over time!

2017: Early Growth

By February 2017, following the model set by Ethereum, a non-profit was established in Switzerland called the Interchain Foundation (ICF). The ICF was governed by a Foundation Council (FC) consisting initially of Jae, Ethan and a local Swiss board member named Guido Schmitz-Krummacher. In the meantime, All in Bits was ramping up development of the Cosmos software — the team was already running testnets of Tendermint based chains with a simple account system sending tokens between each other with a prototype of IBC.

By April 2017, the Cosmos and Tendermint projects had already developed significant grassroots reputation and adoption over the years among early blockchain developers. AiB grew to some ~20 people over 2017. They were working on wallets, an application framework, improving Tendermint, and designing proof of stake algorithms. They were iterating on early versions of the Cosmos-SDK and Cosmos Hub software. They were sketching designs for Ethereum bridges and Decentralized Exchanges and smart contract languages, and so on.

A number of core team members joined AiB in 2017, many of whom are still active in the Cosmos ecosystem today. See the Appendix for more details on who joined in 2017 and what they’re up to now! These were the formative days of what would come to be known as the Tendermint Mafia.

2018: Cosmos-SDK and Testnets

By early 2018 the Cosmos-SDK really started to take shape and the early pieces of the Cosmos Hub started to make use of it. The Cosmos-SDK was aspiring to be the Ruby on Rails for blockchains — a “batteries included” application framework and toolkit for programming blockchains.

Around spring 2018, Zaki Manian officially joined AiB. Zaki had been around since the beginning of Cosmos — he worked through many of the early ideas with Ethan and Jae, he reviewed the whitepaper, was one of the firsts to fund the project, and was always around to provide advice and insight. Zaki would go on to lead the Cosmos testnet program, culminating in the Game of Stakes and the launch of the Cosmos Hub. He would later lead the major Stargate upgrade, and he continues to play a key role in Cosmos strategy and development today.

Over the course of 2018, the testnet program grew considerably and became increasingly decentralized. By Summer 2018, there were decentralized testnet launches with over 100 validators. The team was breaking ground in the deployment of large scale BFT consensus systems. Many of today’s top validators learned the ropes in the early Cosmos testnet program. The technology matured considerably during this period. A new business model — the Proof of Stake validator — was being born.

The team grew and lots of great people were joining (see the Appendix!). Everyone was united by a shared vision to build a sovereign interoperable future for blockchains, a vision that motivates them still, to continue innovating on the frontier.

At the ICF, Michael Niederer was hired to be the CFO, and took over operations in Switzerland. The ICF began an initiative to start funding research more broadly, especially on consensus systems, cryptography, and formal verification. The ICF helped found the Stanford Blockchain Research Center and initiated collaborations with researchers in Switzerland and elsewhere. The ICF started to grow its broader funding program too, expanding the ecosystem by contributing to some early Cosmos-based projects.

By the end of 2018, serious preparations were being made for the imminent launch of the Cosmos Hub. Most notable of these was Game of Stakes, the first Proof-of-Stake, incentivized, decentralized tesnet which served as the culmination of the Cosmos tesnet program and which put the Cosmos software through its most rigorous test to date. Game of Stakes set a new standard for incentivized and decentralized testnets whose legacy is felt across the ecosystem today.

Throughout most of 2018, the Cosmos Hub was famously “two months” away from launch. Chris Goes did offer a prophetic software estimation when in early summer 2018 he predicted that the Cosmos Hub would launch by the end of February 2019. He was off by about 2 weeks.

Stay tuned, for launch and the years that followed, in Part II.

Thanks to the many reviewers who contributed to this history. It has been reviewed and accepted by Jae, Ethan, and Zaki.

Endnotes

1. Fun piece of trivia: Jae Kwon (co-founder of Cosmos), and Jehan Tremback (Interchain Security lead, at Informal Systems) contributed to early debates about Dogecoin monetary policy in this github issue.
2. The security model for Tendermint (and its light client) is defined more formally in this paper by Informal Systems.
3. A cryptoeconomics google group was created. Jae later created a second cryptofinance group to continue discussions.
4. See the original blog post announcing ABCI (which was originally called TMSP, the TenderMint Socket Protocol)
5. A third co-founder, Dustin Byington, was also brought on to All in Bits Inc to lead finance and operations, though he would part ways with the company soon after in 2016.

Appendix

All In Bits Employees
This appendix is notably incomplete, if you are missing from the list and would like to be added or notice incorrect information please reach out to marketing@interchain.io.

2016

• Ethan Frey (now leading Confio which develops CosmWasm and Tgrade)
• Anton Kaliaev (now at Parity)
• Matt Bell (now co-founder of Nomic)

2017

• Brian Crain (now CEO of Chorus One)
• Sunny Aggarwal (now co-founder of Osmosis)
• Judd Keppel (now co-founder of Nomic)
• Greg Szabo (now at Informal)
• Zarko Milosevic (now CTO of Informal)
• Arianne Flemming (now COO of Informal)
• Chjango (now at Osmosis)
• Adrian Brink (now co-founder of Anoma)
• Joon Yun (now at Osmosis)
• Anna Harbaum (now at Interchain Berlin)
• Rigel Rozanski (now contributing to Interchain Security)
• Jim Yang (now at Osmosis)
• Jordan Bibla (now at Blockdaemon)
• Fabian Weber
• Alexis Sellier (now co-founder of Radicle)

2018

• Billy Rennekamp (now at Interchain GmBH)
• Aleksandr Bezobchuk (now at Interchain GmBH)
• Chris Goes (now co-founder of Anoma)
• Zaki Manian (now co-founder of Iqlusion and Sommelier)
• Dev Ojha (now co-founder of Osmosis)
• Jessy Irwin (now at Agoric)
• Shelly Chang (now at Osmosis)
• Jack Zampolin (now co-founder of Stranglove)
• Federico Kunze (now co-founder of EVMOS)
• Aditya Sripal (now at Interchain GmBH)
• Alessio Treglia (now at Ignite)
• Jin Kwon (now co-founder of Saga)
• Ismail Khoffi (now co-founder of Celestia)
• Alex Simmerl (now at Mekatek)
• Mircea Colonescu (now at Informal)