Built a cryptocurrency that races against real time instead of hashes, then spent a week trying to prove myself wrong about it
LapseCoin is a peer-to-peer cryptocurrency using a Verifiable Delay Function instead of proof-of-work: block production is a race to finish a sequential computation tied to real elapsed time, not a race for a lucky hash. The reason that's interesting: the measured hardware-advantage gap is much smaller than PoW's (roughly 3-10x from Chia's own 2019 hardware competition on the same construction, versus the 10,000x+ ASICs opened up over commodity hardware), so it doesn't push toward the same energy-spiral PoW did. No premine, fair launch, FALCON-512 signatures for the quantum-resistance angle.
Spent the last week trying to actually break my own design instead of just shipping it. Found and fixed a real bug that was silently crashing part of the test suite, pulled real chain data off my own live node to calibrate assumptions instead of guessing, and went after a threat I originally thought was a real gap: whether one operator running many builder addresses could gain an unfair edge over the network. Turned out, once actually simulated instead of assumed, it barely holds up as a problem at all: running many addresses at the same hardware tier wins exactly proportionally, and splitting one machine's throughput across addresses is worse than running it as one. Neither needed a fix.