Considering open sourcing our entire database of EVM blocks. Includes traces, transaction receipts etc. Have data for eth mainnet optimism polygon and goereli. Cost about 90k for all of the requests to trace everything. Figure it could lower some barriers 

Anyone interested?

Built to Build. Angel investor. Previously at Eng and Ventures Coinbase

Any interesting insights? Guessing these are geth traces?

Two things I've been curious about -- (1) distribution of obscure opcodes that actually get executed and (2) distribution of precompile calls

Things I thought about a lot writing an EVM impl and optimizing

Before: @StanfordGSB, Kubernetes @Google, @Blackstone

- % of calls that end up REVERTs
- % INVALID opcodes (somehow bad state)
- CREATE2 v. CREATE

Could probably figure out #1 and #3 from this dataset