We believe in the power of public blockchains to provide a public good, powered by a collection of self-motivated actors governed by the rules of a shared consensus ledger, in an environment of competition and freedom of choice. We have learned important lessons from the successes and failures of prior cryptocurrencies, and have been creating innovations of our own that will unleash a new era of speed, security, scalability, and usability.
Cosmos is a culmination of these lessons and innovations, designed to be extended to incorporate even future innovations, to create a distributed ledger platform suitable for generations to come.
One of the great tragedies of cryptocurrencies has been their failure to interoperate with existing systems, and even with each other. It is as if each virtual currency, on its own blockchain and with its own community, were too sovereign for its own good. This is made most apparent when you attempt to buy your first Bitcoin: you discover that it is remarkably difficult to buy the new internet money over the internet, at least without divulging considerable amounts of personal information and allowing weeks for it to be verified by third parties (exchanges) you hardly know and have been warned not to trust with too much of your money.
While mechanisms like atomic swaps between chains can help link communities, they suffer from a need for mature liquidity markets and active participants on each side (or suffer long timeouts). The alt-coin ecosystem is an alchemical soup of enthusiasts experimenting in various ways in new cryptocurrency design, but each alt-coin stands alone, siloed by the lack of a general inter-blockchain communication system, many needlessly wasting electricty through convoluted Proof-of-Work (PoW) designs.
Ethereum sought to address the problem, somewhat indirectly, by inventing a new platform for computation, allowing arbitrary financial systems to be written against a common virtual machine using a common development environment. Presuming all new cryptocurrencies launch on Ethereum, they will all have some capacity for interoperability. Ethereum is a wonderful idea, but the design decisions, while motivated by important considerations (like the need for a simple specification and guaranteed determinism), neglect a great many more (like the difficulty of designing a secure general purpose programming environment), and lends Ethereum it's own sovereignty problem, which is that it excludes users who want access to the platform but want alternatives to the current state transition machine.
Of course, Ethereum doesn't address the problem of turning fiat currency into ether, either, short of having banks and major government bodies issue their currency on Ethereum. And while many such institutions are investigating the Ethereum blockchain, there are both scalability and governance concerns that will restrict its full scale adoption, and hence impede integration with existing systems and currencies.
Each existing cryptocurrency supports some cultural ideal. Bitcoin is decidedly libertarian, with a slight anarchist leaning which is ultimately betrayed by the economics of Proof-of-Work. Ethereum is less a political statement and more an avant garde academic excercise in distributed systems, computer science, and contract theory. Dogecoin is a testament to the capacity for Humans to bond over the absurd, a phenomenon more commonly known as "being social". Steem offers a more structured version of being social, with a basis in content production and curation. And so on. But there are many more idealogical currencies to create, many more virtual machine designs to try, each affording a unique experiment in the intersection between governance, culture, and economics. Would it not be a tragedy for all these experiments to be so rigidly siloed from one another by great chasms of incompatibility?
Clearly, what we need is something more general than Ethereum - a sort of meta-Ethereum; something that can integrate existing codebases and the states of existing blockchains; something that can interface in a sane manner with the currencies issued by nation states; something that formalizes the relationship between cryptocurrencies and connects them without compromising their independence; something with extreme flexibility that does not compromise security.
Enter Cosmos. Cosmos is a network and a framework for interoperability between blockchains. It consists of a web of "hubs" and "zones", where each "zone" is effectively an independent blockchain with an arbitrary cryptocurrency design (be it like Bitcoin, Ethereum, ZeroCash, CryptoNote, your local banking institution, etc.), and each "hub" is a multi-asset cryptocurrency that facilitates interoperability between some set of zones.
Hubs and zones are powered by the Tendermint Proof-of-Stake (PoS) consensus algorithm, using the TMSP interface to host applications written in any programming language. TMSP permits enormous flexibility in application design, and enables the application to inherit the security features of a Tendermint-powered blockchain.
Zones communicate with one another through a hub, primarily in the form of assymetric transfer of some set of tokens from one zone to another. Effectively, a hub is a blockchain with many sidechains, but using PoS instead of PoW leads to a dramatic reduction in overhead without sacrificing security. The trick is that each zone acts as a light client for the hub, and the hub acts as a light client for all its zones. Tendermint makes this efficient with compact light client proofs that are secure so long as the validator set is known, which can be achieved by ensuring the light client synchronizes with validator set changes at least as often as they are permitted to occur.
Many people have rallied against the use of Proof-of-Stake, claiming it is impossible to secure. Simultaneously, they acknowledge that it is nearly trivial to secure many PoS chains using a single, secure PoW chain. It is my own position that the correct number of PoW chains is one, and the correct number of PoS chains is in the thousands or millions, on par with the number of currencies, rewards systems, and various other token-based systems in the world.
Cosmos reflects this position in that it makes no distinction between hubs - there is no "top" hub, and the most popular or successful hub is a matter of adoption by zones. Furthermore, there is no fundamental distinction between zones and hubs, such that any zone with sufficient functionality can become a hub. While we will release one particular hub in order to get things started (the aptly named Adam hub), there is nothing stopping this hub from becoming obsolete as other hubs, even the likes of Bitcoin or Ethereum themselves, come to compete with it, and potentially replace it.
Cosmos thus introduces a new kind of market dynamic that formalizes the power struggle between the various cryptocurrency offerings, demolishing the barriers to interoperability and enabling competition between hubs to contribute to the economic security of each hub itself - validators who misbehave will be abandoned for better behaving ones. The result is a heterogeneous network, adapting for each community and each locale according to its needs, and yet retaining a backbone of interoperability that keeps barriers to entry low while holding stewards of the system accountable and quality of service high.
Whether as testing grounds for new cryptocurrency designs, or an upgrade to existing ones, as a means for decentralized exchange, or a platform for scalable smart contracts, Cosmos's potential is that of every existing cryptocurrency and more in synergy. The only question is, what will you build on Cosmos?