just defended my final year project in uni few weeks ago, a blockchain voting system for university elections. keccak256 identity verification. six months of work. then i read the bonkdao writeup this week. $20m drained without a single vulnerable line of code. attacker just bought enough tokens to pass the proposal. my system would've folded the same way. the code was clean. identity hashing worked. checks-effects-interactions everywhere. but i never modeled the economic layer. never asked what happens when the cost to acquire voting power drops below the value of what the vote controls. my defence panel didn't catch it either. because that's not where anyone looks in university. the gap between your code being correct and the system being secure is where the real work lives. Thats the part i am choosing to focus on for the next couple of years