1/ Why Ethereum Needs a Dynamically Available Protocol. Ethereum's next consensus protocol should have a provably secure, fast “heartbeat” layer that never stops producing blocks, regardless of how many validators are asleep.
2/ Ethereum has never stopped producing blocks — not during the Merge, not during client bugs, not during outages. Dynamic availability is the formalization of this property. The next consensus design should strengthen it, not compromise it.
3/ We make the case for dynamic availability as a strict requirement — grounded in resilience, censorship resistance, and application-layer continuity.
4/ Moreover, we cover why an impossibility result forces the two-layer split, why LMD-GHOST doesn't meet the bar, what Goldfish brings, and how ~256 validators per slot unlock both faster slots and a near-term path to post-quantum signatures.