Fee only security destabilizes mining incentives

When transaction fees dominate the block subsidy, the incentive to extend the chain honestly weakens: miners can profit by re-mining ("sniping") fee-rich blocks instead of building on them, and this dynamic favors large pools.

Statement: In a subsidy-dominated regime, the dominant strategy is to mine the next block. In a fee-dominated regime with variable fee flow, a fee-rich block creates an incentive to fork it and capture its fees rather than extend it — and only miners large enough to plausibly find two blocks in a row can attempt this, so hash power rationally consolidates toward the largest pools. Academic modeling (Carlsten et al. 2016) reached the same conclusion independently: without block rewards, undercutting and fee-sniping deviations become profitable and equilibrium mining behavior becomes unstable.

Nuance: This is a modeled result, not an observed one — Bitcoin has never operated in a fee-dominated regime, and the models assume sharply variable fee flow; a future with deep, steady fee demand (or out-of-protocol arrangements between transactors and miners) could damp the instability. Counterpoints raised in the same conversation: indirect-benefit mining (miners paid via adjacent services, as with internet infrastructure) may fill the gap opaquely, and proposed in-protocol fixes exist (perpetual issuance, security-tax constructions) but each contradicts other strongly held commitments. The claim is well supported as a risk analysis; any confident prediction of how the transition actually plays out exceeds the evidence.

Common misstatements: "Fees will obviously replace the subsidy" and "the subsidy doesn't need replacing" are both asserted on stage as settled; neither is. "The chain will implode when the subsidy ends" (the credible failure mode is degraded incentives and centralization pressure, not sudden collapse — and the subsidy declines gradually until ~2140).