1/ Today @UMAprotocol is publishing our design for a Data Verification Mechanism (DVM), a blockchain oracle with **economic guarantees** around the cost of corrupting the system.
We're ALSO releasing our v1 code + deploying it to Kovan!
3/ Here’s our bold claim: **every** blockchain oracle is corruptible, which means any smart contract relying on off-chain data is corruptible and manipulatable **for some price**. This means that your DeFi smart contract could be at risk.
1/ How @UMAprotocol enables "Trustless Tokenization", aka how to create an ERC20 token for synthetic crypto or *real-world* asset exposure in ETF-like format. Mechanics below for creation, re-margining, and redemption (or default):
2a/ Context: Investor wants gold exposure with notional N and is willing to pay a fee F for this service. Provider is responsible for passing the total return of gold to Investor through the smart contract and earns fee F for doing so.
2b/ Context cont'd: Investor and Provider back their promise with margin. Provider uses digital asset M as collateral, which must be above margin requirement MR at all times. A default penalty DP is assessed if Provider fails to maintain M > MR at all times (aka they default).
1/ Important paper published by my friend Prof Tonetti (@ChrisTonetti) and his colleague Prof Jones @StanfordGSB. They have rigorously studied the "economics of data" to show that there are large social gains from letting users own and sell their own data. Some key takeaways:
2/ Jones and Tonetti show that data can be considered a "non-rival" good. Just like ideas, data is not depleted through use. This is in stark contrast to most goods in the economy which are "rival" and can only be used by one person or firm at a time.
3/ Currently, most data is treated like it is a rival good—it is owned by the firm that produces it and is *not* shared. This is functionally equivalent to a firm being given a lifetime patent on an idea and disallowing other firms to use that idea. This is far from optimal.