Exploration and Discovery on Blockchains
Blockchain explorers are now a critical part of the blockhead toolbox. Blockchain.com, Etherscan.io, SolanaBeach.io. AlgoExplorer.io — every major blockchain has a corresponding explorer. People use these websites to verify that payments were sent and received, that smart contract calls were completed, and to look up activity on addresses of interest.
Most blockchain users do not have direct access to blockchain nodes. Without explorers, they would be flying blind.
Critical though blockchain explorers are, they leave a huge gap in the user experience. The gap of discovery.
Not a small world after all
The number of wallets on Blockchain.com (which is a rough estimate of the number of blockchain users) has grown from roughly 10,992,000 on December 31, 2016 to roughly 76,315,000 on August 31, 2021. That’s more than a 7-fold increase in just 5 years. What’s more, this growth rate is accelerating.
This increased adoption is forcing blockchain technology to evolve so that it can support the increased transaction volume. Advances like Bitcoin Lightning, side-chain based protocols like Polygon, Ethereum’s upcoming migration to proof of stake consensus, and the rise of proof of stake blockchains like Solana and Algorand mean that there are no technological barriers to increased adoption. The world of blockchain users was small in 2017. It has become a big world in 2021 and is well on its way to becoming a global movement.
Small worlds are like villages. Everybody knows everybody else’s business. All they have to do is query the grapevine to find out what’s happening in their neighbors’ lives. When blockchain users lived in a small world, the blockchain explorers were the grapevine. When the world was small, they were the perfect tools to understand what was happening on the blockchains.
The world is now big, and the amount of raw transaction information on the blockchain explorers is overwhelming for even sophisticated users of blockchains.
The Ethereum blockchain processes around 50,000 transactions per hour. A single Ethereum node can add hundreds of thousands of pending transactions to its transaction pool in the same amount of time. The Solana blockchain can process that many transactions per second.
The first step to discovery
This is far more information than a human being can manually sift through in a blockchain explorer. Processing this volume of information requires the help of algorithms which summarize blockchain activity and single out particularly interesting events for human consideration. The big world requires tools that aid discovery, not those built for mere exploration.
The first step to discovery is to build on top of the technology of the explorers. The most primitive discovery tools would present aggregate statistics about the transactions crawled by the explorers. They would do things like call users’ attention to outliers such as accounts in the upper percentiles in terms of number of recent transactions executed, or recent value transferred. They would inform users about accounts that were front running smart contract calls, and tokens that were being pumped, or dumped.
Filters, analyzers, and AI
Where the most primitive discovery algorithms mentioned above are essentially streaming algorithms on blockchain data, their successors would be more stateful. They would identify different markets — e.g. NFT markets, token markets, decentralized exchanges, etc. For each of these markets, they would model the baseline behavior of the markets and then alert the user about activity that deviated significantly from the baseline as events of interest.
The next step would be to move from simple statistical analyses of these markets, and of their participants, to machine learning based evaluations of whether or not an event is noteworthy for a given user. This stage of discovery represents a very rich set of techniques which push at the outer limits of our technological capability. It is also inevitable. We will be using these tools by 2025. The only question is, who will build them?
Enter Moonstream
Moonstream is an open source project focused on bringing discovery to blockchains. We are building the future described above, in which machine learning algorithms choose events of note to surface to blockchain users based on their preferences and risk exposure.
We have started off by focusing on Ethereum, because this is where defi activity is concentrated, with the convergence of ERC20 tokens, NFTs, decentralized exchanges, and even more sophisticated protocols. Users can currently subscribe to events occurring on the Ethereum blockchain and in the Ethereum transaction pool.
Bitcoin and Solana are next. We plan to support every major blockchain.
Last week, we entered the first phase of discovery by launching an Ethereum whale watch:
Each whale watch card presents a summary of the past hour’s activities on the Ethereum blockchain. It shows the top accounts by number of transactions and value (currently in wei) sent and received. This has already been extremely useful in discovering massive token pumps and popular NFT mints that we would not otherwise have known about.
This month, we plan to transition to the second stage of discovery. We will do this by building an index of the Ethereum NFT market and calling attention to NFTs that deviate statistically from the baseline expectations of the index. This will evolve into ongoing analysis and classification of NFTs and NFT holders using machine learning models.
And then we will take these techniques beyond the NFT market, to tokens, decentralized exchanges, and even across chains.
We believe in inclusion, both financial and social. The only way for software to be inclusive is for it to be free (as in freedom). You can find all of our code on GitHub. We hide nothing. You can use Moonstream at https://moonstream.to.
If you share this vision of discovery, join our community on Discord. Let us build this future together.
