Was discussing with Federico, and bringing it to ethresearch per his request
Great talk at ethcc. Couple of thoughts I’d love to share with you. I work on cryptoeconomics at EF, and found your presentation fascinating (especially the historical framing). Added Schelling and the Federalist Papers to my reading list.
- "What people would think vs what I think” – While I do think that the coordination problem as an incentive mechanism is quite clever, it may not work in all cases, especially if there are times when a juror thinks “I can see how most people in my juror set would support person A, but if they could hear my argument for person B, they may change their mind,” then the protocol would encourage the juror to agree with the consensus rather than vote her mind.
This would work in most cases but may inadvertently reward “groupthink” in the subtle cases.
- Cryptoeconomics – I was beginning to think about how this model would scale in each of low value, medium value and high value disputes. In high value disputes, subtlety may require deliberation and the early version of the protocol may not be suitable. In low value disputes (even the $500 web dev case), to pay for 5 jurors say $20 of their time, it would eat up 20% of the value in “legal costs.” This has a few problems. Good actors are used to centralized parties being able to recoup all costs in case of successful dispute resolution. In order to do that, the counter party will have to be penalized, which will require a financial stake to do work (i.e. web dev, which could be an interesting solution). Also, depending on how much kleros tokens are required to be a juror, the “yield” of being a juror may not be enough ($1000 stake for $20 yield per case for example is starting to become unattractive to the “lazy” segment of people)… especially net of the risk of getting “slashed.” In which case, it attracts jurors that are ideologically driven rather than are purely financially driven. This may leave medium value disputes around ICO funding or other high value crypto (and/or all digital) transactions the most compelling initial use case.
Would love to hear what you think about these topics.
Super excited for your team’s progress. Do agree that this is a layer that should be abstracted away from most networks as an outsourced model.