2. Purpose of Law
โ Prevent involuntary interactions
โ eg. theft, murder, etc
3. Purpose of Law
โ Mediate ongoing interactions
โ eg. I offer to sell you shirt for $10, you send
me $10, I do not send shirt
4. Types of fraud
โ Total
โ not sending a product
โ Partial
โ sending a product with very bad quality
โ Visible
โ defect is immediately obvious
โ Invisible
โ defects take months to manifest
โ defects have only a small chance of manifesting
โ Types of underlying commerce
โ Product, one-time service, ongoing service
โ Financial service
5. Idea: incentives
โ No legal system
โ Profit from acting honestly: $3
โ Profit from cheating: $10
โ Legal system
โ Profit from acting honestly: $3
โ Profit from cheating: -$40
โ Problems
โ How to prove fraud (information asymmetry)
โ Cost of filing a claim
โ Cost of enforcement
6. Cryptoeconomies
โ Involuntary interaction
โ Violent attack impractical
โ Hacking, perhaps practical
โ Ongoing interaction
โ Fraud more difficult (information is more public)
โ Fraud easier (people are anonymous)
โ General problems
โ Commerce international (enforcement harder)
โ Rapidly changing society (harder to set
standards)
8. Solutions
โ Know whom you are trusting
โ Add disincentive to cheat
โ Require collusion to cheat
โ Make cheating more visible
โ Make cheating impossible
9. Know whom you are trusting
โ Idea: if A trusts B, and B trusts C, then C is
more likely to be trustworthy
โ Secondary benefit: if C cheats B, C loses
reputation with A too
โ Naive implementations vulnerable to sybil
attack, be careful
10. Add disincentive to cheat
โ Deposits
โ A submits N coins as deposit, if A cheats
then deposit goes to B
โ Sacrifices
โ A sacrifices N coins to be able to
participate in a system
โ Idea: allow reputation to go negative
โ Hostages
โ A submits N coins as hostage, if A cheats
then coins are destroyed or sent to charity
11. Require collusion to cheat
โ Multisig
โ Oracles (for data)
โ Wallets (for personal security)
โ Multisig 2.0
โ Withdrawal limits
โ Rights of refusal
13. Make cheating impossible
โ Factum law
โ Idea: the definition of how much money
you have is based on the results of past
contracts
โ If you cheat, the system automatically
deducts $X
โ Problem: not everything is cryptographically
verifiable
โ Computational resources: yes
โ Math problem solutions: yes
โ Subjective effort: no
โ Offline effort: no
16. Benefits
โ Cost of enforcement tends to zero
โ Scalable (works for both $1
microtransactions and very large trades)
โ No barriers to entry, anyone can participate
17. Problems
โ How to determine precedents?
โ What level of bad quality is actually
fraudulent?
โ What are the defaults?
โ Network effects
โ Only effective if many people using it
โ Solution: start in a community