I trace the shadow before it casts. Last week, that shadow fell across a Bitcoin mailing list thread, and it wasn't a bug report or a vulnerability disclosure. It was an old argument, dressed in new clothes, asking a question that cuts to the bone of what Bitcoin is: Should the block reward ever be permanent?
Adam Back and Peter Todd โ two of the ecosystem's most recognizable technical voices โ are now on opposite sides of that question. Todd argues for a small, never-ending issuance to keep miners paid after the last subsidy vanishes around 2140. Back calls it a trap dressed up as engineering. The Bitcoin++ conference account resurfaced Todd's talk on the topic this week, and the debate ignited all over again.
The timing matters less than the mechanism. But the mechanism, as usual, is where the truth hides.
The Mechanics of a Vanishing Subsidy
Bitcoin pays miners in two ways. Block subsidies mint new coins, and transaction fees ride along with each block. The subsidy is roughly halved every four years, and it hits zero around 2140. After that, fees alone must carry security.
Todd argues that fee revenue swings too violently to hold the chain together. In a world where fees are the only reward, miners would face a perverse incentive: reorganize the chain and re-mine fat-fee blocks rather than build forward. A fixed reward, he says, kills that pull. It stabilizes the game theory at the cost of a tiny, asymptotic inflation.
His case leans on lost coins. Todd models supply against a loss rate and finds it settles at a ceiling โ coins vanish as fast as fresh ones appear. Therefore, he frames tail emission as a stabilizer, not inflation. He points to Monero, which already runs a small permanent reward. Its apparent inflation rate keeps sliding toward zero.
There's a quiet elegance to the math. I spent six weeks in 2017 auditing the Ethlance Crowdsale contract line by line, and that experience taught me that the most dangerous flaws are never the loud ones. They're the assumptions that feel so reasonable nobody bothers to interrogate them. Todd's model has that feel. It's clean. It's internally consistent. And it's seductive precisely because it solves a real problem.
The problem is real. Miners currently earn 3.125 BTC per block, with roughly 30 halvings still ahead. Each one thins the subsidy further while fees remain lumpy and unpredictable. On a quiet Sunday, a block can carry almost nothing in fees. On a busy Tuesday during a liquidation cascade, fees spike. That volatility is a security cost that someone, eventually, will have to pay.
The Trap Narrative
Back rejects the framing outright. His counterargument is not about the math โ it's about the narrative machinery. He points to BIP-110, the contentious 2026 soft fork that tried to filter non-payment data out of blocks, as the model for how these campaigns get sold.
"The trick is finding ways to trigger and rally people to your dangerously inadvisable cause with simple though false narratives," Back wrote. "110 used 1) JPEG spam and illegal content could be stopped but devs are captured so they won't, 2) anti layer2 anchors devs want to etheriumize bitcoin."
That pattern has a recent scoreboard. BIP-110 failed after two blocks this month, with miner support near 2.53% against a 55% bar. Back had predicted the stall weeks earlier. The backers now chase a breakaway coin instead.
The parallel is explicit. Bitcoin commentator Trey Sellers wrote that a supply-schedule fork would fail as hard as BIP-110, if not harder. Michael Saylor raised a related concern, warning about protocol neutrality whenever consensus rules bend to one camp. Bitcoin's 21 million cap is not just a parameter โ it's a social contract. It's the one invariant that even the loudest factions rarely question.
I've been in enough audit war rooms to recognize a certain pattern. When someone proposes a change that sounds purely technical, but the timing aligns with their own position in the system, my skepticism sharpens. Todd is a smart engineer. He's also a vocal proponent of sidechains and alternative scaling paths. A permanent block reward doesn't just change miner incentives โ it changes the competitive landscape for anyone building on top of Bitcoin.
That doesn't make his argument wrong. It makes it worth examining from more than one angle.
The security question survives the politics. Bitcoin Knots developers spent August claiming the network faces attack, while miner incentive disputes drew in former Ripple CTO David Schwartz. In contrast to those fights, this one carries no deadline. There's no emergency. There's no exploit in the wild. There's just a long, slow countdown that nobody alive today will see reach zero.
The Asymmetric Costs of a Hard Fork
One difference cuts against Todd. BIP-110 asked for a soft fork, which needs only miner cooperation. Raising the cap demands a hard fork, and every holder would have to accept it. That's a fundamental asymmetry.
A soft fork is a tightening. A hard fork is a renegotiation. Soft forks fail when miners don't cooperate. Hard forks fail when the community doesn't. Bitcoin has survived contentious hard fork attempts โ the blocksize wars proved that โ but each one costs more than the last. The scar tissue accumulates.
Here's where the contrarian angle emerges. In my 2020 formal verification work on Curve's stableswap invariant, I ran 10,000 simulated arbitrage attacks against the AMM model. The invariant held. But the process taught me something about security that applies to Bitcoin's cap debate: resilience is not the absence of pressure. It's the ability to absorb pressure without changing shape.
Todd's proposal would change Bitcoin's shape. It would turn a fixed supply into an asymptotically stable one. The difference is small โ fractions of a percent per year โ but the principle is enormous. Once the cap becomes a parameter instead of an axiom, every other parameter becomes negotiable.
And yet, Back's dismissal feels too fast. The fee-only future has its own failure modes. If fees are too low, the chain is underpinned by nothing. If fees are too high, only whales can transact. The current design assumes that demand for block space will grow smoothly and predictably over a century. That's an assumption I've seen fail in nearly every market I've analyzed.
I think about the NFT generator logic I reviewed back in 2021 for Art Blocks Curated. The random seed entropy had a subtle predictability flaw โ nothing that would trigger an alarm, just a slight bias in the block hash dependency. The artist and team fixed it quietly. The integrity of the work survived. That's how good security operates: it catches the subtle bias before it becomes a systemic failure.
Bitcoin's supply schedule has a subtle bias too. It biases toward scarcity. That bias has a cost โ it assumes fees will scale โ but it also has a value that no model can quantify. It's the bias that made Bitcoin worth trusting in the first place.
The Century-Long Test
The argument will not be settled this week, or this year. Fees may yet fund the chain on their own. Nobody alive today will see that test finished. But the debate itself is a form of stress testing. It forces the community to articulate why 21 million matters, not just that it matters.
From my angle in the DeFi security space, the deeper issue is less about inflation and more about governance. I've watched stablecoin yield products like sUSDe build maturity mismatches and stacked risks that work beautifully in bull markets and detonate in bear markets. The same pattern appears in protocol debates: elegant arguments that assume favorable conditions hold.
Todd's model assumes a stable loss rate. Back's position assumes a stable fee market. Both assumptions are elegant. Both will be tested.
The most honest answer I've found comes from the code itself. Bitcoin's consensus rules are beautiful because they're rigid. That rigidity is a feature, not a bug. It's what makes the system predictable. It's what makes audits possible. In the void, the bytes whisper truth โ and the truth is that changing a constant in the consensus layer changes everything downstream.
Vulnerability is just a question unasked. The question of whether Bitcoin's security can survive on fees alone is a legitimate one. But the question of whether Bitcoin should change its fundamental supply schedule to answer it is a different matter entirely.
The Shape of the Next Decade
The 21 million cap is not a technical parameter. It's the line between Bitcoin as money and Bitcoin as a mutable experiment. Breaking that cap would not be a bug fix. It would be a redefinition. Security is the shape of freedom โ and the shape of Bitcoin is, and should remain, 21 million.
I listen to what the compiler ignores. The compiler ignores the social contract. It doesn't care about narratives or false frames. It only executes logic. And the logic of a hard fork to change the supply is the logic of a system that no longer trusts its own axioms. That's not a security improvement. That's a surrender to the chaos that Bitcoin was designed to escape.
The debate will return. The argument is too elegant to die. But when it does return, I hope the community remembers what BIP-110 taught us: simple narratives are the most dangerous attack surface of all. Logic blooms where silence meets code, and some silences are worth protecting.
