Great points — and I think this tension will define stablecoin evolution for the next cycle. Smart-contract-enforced overcollateralization offers transparency and decentralization, but it struggles with capital inefficiency and volatility in collateral value (as we've seen during sharp market drawdowns). Fiat-backed models like USDC and USDT, for all their centralization risks, still command trust because they're simple to understand and redeem.
I’d argue the future likely isn’t binary. Hybrid models — combining on-chain transparency with off-chain fiat anchoring, or dynamically adjusting between overcollateralization and real-world asset backing — might offer a better balance. Frax’s partial algorithmic approach was an early attempt, but there’s still a lot of experimentation to be done around incentive structures and risk management.
It’ll be interesting to see if decentralized protocols can build strong-enough confidence loops without needing a fiat-backed “anchor” to lean on. Until then, centralized stables probably aren’t going anywhere.