Best breakdown of this I've seen. You ended up exactly where I did: the math is the easy part, and clearing the leverage leg is the whole problem.
I actually shipped it on Sepolia. Gimbal is options-not-debt running live on Sepolia as an auto-rolling wrapper. Backed only by ETH, the two legs always sum to 1 ETH, the core invariants are checked in Halmos, and it's all open source.
Where I'd push back a little: you frame it as needing an options exchange, but the roll is just one-sided market making. That can be a permissionless on-chain auction baked into the asset itself. Maybe the venue is the asset.
Still, depth for the leverage leg is the real frontier, and that's where all my time has gone.
https://t.co/RPyEaA85zx
https://t.co/5pf14TWNoE
Would really appreciate anyone that can send a good amount of Sepolia eth to this address for dapp building and testing!
0x3684abb8e19F1b00A91d8f1Bf7189B3bCA2af78A
11/
And the big one: UNAUDITED. Testnet only. Bring curiosity, not money β break it and tell us.
Design: "index-tracking assets on top of options instead of debt" βhttps://t.co/ijDLZzhSEN by Vitalik Buterin, with refinements from the thread built in.
Gimbal is live on Sepolia testnet β https://t.co/YIw4Eh6JuR
Soft peg any asset, backed by nothing but ETH in a system with no liquidations, because there is no debt to liquidate.
Built from a Vitalik research post. Every line written by AI. Honest details π§΅
12/
β³ Still proving (solvers running as I post): the remaining monotonicity region, the peg-value cap, and the headline no-bad-debt redemption theorem. Zero counterexamples in 13+ hours of symbolic search; 128,000-call randomized invariant campaigns all pass.
@avsa There are a lot of these types of brutal warehouse jobs where itβs not a bottleneck, and making special purpose tech is more expensive than just paying some dude minimum wage to go through the same motions for 8+ hours a day. A general purpose robot fits great here