A few implications for future defi design after this watershed incident (KELP/LZ/Aave exploit): “money lego” and composability is not free lunch - LRTs/wrapped assets should never be treated as equivalent to native assets (same with the Binance wrapped assets in CEX) - Yield-enhanced wrappers are also risk-enhanced wrappers - Unified collateral pool lending models will be repriced - Every new collateral type doesn’t just add a market.. it adds a new risk on top - Lending/vaults protocols will increasingly compete on risk segmentation (im more bullish in modular pools) - More assets will prefer canonical issuance, native deployments, and limited bridge surface area - The cost of acquiring TVL on L2s & new dapps will keep going up: most lot of TVL was effectively rented through incentives, bridge convenience, wrappers, and leverage loops - That TVL will increasingly flow back to L1 & likely tradfi-linked assets - Human devops is the worst design by nature :(