The fight over what belongs on Bitcoin has shifted from protocol governance to guerrilla software development, with Ordinals advocate Leonidas announcing Friday a new open-source client called DOG Mode that would bypass the stalled BIP 110 anti-spam proposal entirely.
Rather than seeking miner consensus to restrict non-financial data (something BIP 110 has spectacularly failed to achieve, drawing exactly zero signaling support in the current period), DOG Mode takes the opposite approach. It would loosen Bitcoin Core’s relay policies to let near-block-size transactions propagate freely and slash the dust limit to a single satoshi. If even one miner runs compatible software, those transactions confirm. No vote required.
The proposal exposes a fundamental tension in Bitcoin’s architecture: consensus rules govern what makes a block valid, but relay policy determines what most nodes will actually forward. Bitcoin Core’s defaults function as de facto network rules because almost every node runs Core. DOG Mode would create a parallel relay path for transactions Core currently rejects as “non-standard,” even when those transactions are technically valid under consensus.
The Technical Split: Relay Policy Versus Consensus
Understanding what DOG Mode actually proposes requires separating two concepts that Bitcoin users often conflate.
Consensus rules are the hard boundaries. Break them and your block gets rejected by the network. A node that accepts invalid blocks forks itself off Bitcoin entirely. BIP 110 is a consensus change: it would add new rules defining certain data as invalid, requiring 55% of hashpower to signal support during a designated window. That’s why it needs miner buy-in.
Relay policy is softer. It governs what an individual node chooses to pass along to its neighbors. Bitcoin Core ships with defaults that reject transactions it deems “non-standard,” meaning they’re valid under consensus but Core won’t help them propagate. A miner who receives such a transaction directly can still include it in a block, and the resulting block stands because it violates no consensus rule.
DOG Mode targets relay policy exclusively. It would raise the maximum standard transaction weight from 400,000 units to 3,900,000 units. To put that in context: a Bitcoin block holds four million weight units, so Core currently relays nothing larger than roughly one-tenth of a block. DOG Mode would carry transactions filling nearly 98% of one.
The client would also cut the dust limit from its current range of 294 to 546 satoshis down to a single satoshi. That floor exists because tiny outputs cost more in transaction fees to spend than they’re worth, cluttering the UTXO set with economically unspendable coins. But Ordinals and Runes protocols have to pad their outputs with extra bitcoin to clear that threshold.
The $25 Million Locked in Padding
Leonidas claims that eliminating the dust limit would release approximately $25 million in padding back to the Ordinals and Runes ecosystems. That figure deserves scrutiny.
Ordinals embed images, text, and other arbitrary data directly into Bitcoin transactions. Runes, created by Casey Rodarmor, issue tradeable tokens on the network. Both must attach small amounts of bitcoin to their outputs to satisfy Core’s relay requirements. At current prices, millions of satoshis sit locked in outputs that exist only to meet a threshold, not because anyone wanted to store value there.
Whether freeing that capital actually benefits those ecosystems or simply enables more aggressive use of block space depends entirely on your view of what Bitcoin is for. The restrict-data camp sees Ordinals inscriptions as spam that crowds out financial transactions. Ordinals proponents counter that any valid transaction paying market-rate fees has as much right to block space as any other.
The $25 million estimate is unverified in any public analysis, and Leonidas has a direct financial interest in the outcome. He co-founded Runestone, the project that seeded the DOG token. A relay policy that benefits Runes would benefit markets where his own token trades.
BIP 110’s Complete Failure to Gain Traction
To understand why DOG Mode exists, look at BIP 110’s numbers.
The proposal to cap arbitrary data is structured as a user-activated soft fork. It needs 55% of miners to signal support during a defined window. According to the BIP 110 monitor, it has drawn zero signaling in the current period. In no previous period has it ever cleared roughly 1% support.
That’s not a political setback. It’s a categorical rejection. Miners have shown no appetite whatsoever for a consensus change that would restrict what can be included in blocks, presumably because Ordinals and Runes transactions pay fees like any other. Revenue is revenue.
BIP 110’s node support sits in the low single digits, carried almost entirely by Bitcoin Knots, the long-running alternative to Core favored by the restrict-data camp. But node counts don’t matter for a user-activated soft fork. Miner signaling does, and miners aren’t signaling.
This asymmetry matters. BIP 110 requires a supermajority that doesn’t exist and shows no signs of forming. DOG Mode requires exactly one willing miner.
The MARA Slipstream Precedent
Services already exist to bridge the gap between valid-but-nonstandard transactions and miners willing to include them. MARA’s Slipstream is the most prominent example, brokering direct relationships between transaction creators and the mining operation.
This model works but doesn’t scale elegantly. Users wanting to inscribe Ordinals or mint Runes have to know a miner or use a service like Slipstream. DOG Mode would create a relay network that propagates these transactions like any other, reducing friction for anyone running the alternative client.
The question is whether enough nodes would adopt DOG Mode to create meaningful propagation. Leonidas is explicit that the client would deviate from Core far less than Knots does, framing it as a minimal-modification fork rather than a comprehensive alternative. That pitch targets users who want one specific policy change without abandoning Core’s broader architecture.
Vaporware Status: No Code, No Repository
Here’s the catch: DOG Mode does not exist as software.
Leonidas announced an initiative on Friday and asked developers to contribute to an initial release, miners to add support, and users to quote-post a slogan. There is no repository, no version, no benchmark. It’s a community leader announcing a Bitcoin client and asking someone else to build it.
That doesn’t mean it won’t happen. Open-source projects regularly start as announcements seeking contributors. But the gap between “prominent figure proposes idea” and “functional software with meaningful adoption” is substantial. Bitcoin Knots took years to build credibility and still commands only single-digit node share.
The announcement’s timing appears strategic, coming five days after reporting that BIP 110 was approaching its deadline with effectively zero miner backing. If you’re in the expand-data camp and your opponents’ consensus change is dead on arrival, announcing a relay-policy alternative while attention is focused on the debate makes political sense even without shipping code.
The Broader Philosophical Divide
Bitcoin’s block space debate isn’t really about relay policy or dust limits. It’s about identity.
One camp sees Bitcoin as a monetary network where every byte should serve financial transfer. Arbitrary data, images, token protocols, and inscriptions waste scarce resources and price out small transactions. Their solution is consensus rules that define non-financial data as invalid.
The other camp sees Bitcoin as a permissionless system where any valid transaction paying market fees has equal standing. If someone wants to inscribe an image and pays the going rate for block space, that’s as legitimate as sending bitcoin to an exchange. Their solution is relay policies that don’t discriminate.
Neither side is arguing about technical facts. They agree on how consensus and relay work. They disagree about what Bitcoin should be, and that’s not a question code can answer.
Both sides are now fighting through alternative clients rather than through Bitcoin Core itself. BIP 110 proponents run Knots. If DOG Mode ships, Ordinals proponents would run that. The network fragments into ideological relay zones, each propagating transactions the other won’t.
Whether this is healthy competition or corrosive fracturing depends on your tolerance for Bitcoin’s political messiness. The currency has always been a compromise between people who want different things. Alternative clients are just another arena for that negotiation.
For traders tracking the DOG token specifically, the announcement had minimal immediate price impact. Market participants may be waiting to see whether DOG Mode actually ships before pricing in any relay-policy changes. A proposal with no code is a promise, not a product.
What Happens When One Side Doesn’t Need Consensus
The most interesting implication of DOG Mode isn’t technical. It’s governance.
BIP 110 follows Bitcoin’s established path for protocol changes: propose, discuss, signal, activate. It requires broad agreement and imposes coordination costs. That process failed comprehensively here, with miners showing zero interest in restricting data.
DOG Mode routes around that failure by targeting something that doesn’t require agreement. Relay policy is a node operator’s individual choice. If Leonidas can get enough people to run his client and one miner to include the transactions, he wins without ever needing permission.
This creates an asymmetric dynamic. The restrict-data camp has to build supermajority consensus for their consensus change. The expand-data camp only needs to peel off a sliver of hashpower for their relay change. One side plays by the rules that govern Bitcoin’s evolution. The other side found rules that don’t apply.
That’s not cheating. It’s accurate reading of how the system works. But it does raise questions about what happens when ideological factions stop seeking consensus and start building parallel infrastructure. Bitcoin’s unity has always been somewhat illusory, a shared database maintained by people who want incompatible things. Alternative clients make the fractures visible.
The restrict-data camp could respond by hardening Core’s relay policies, making it even harder for DOG Mode transactions to propagate. Or they could accept that relay policy was never the place to fight this battle and refocus on consensus changes with actual miner support (though no such support exists today).
For now, both proposals exist more as political statements than functional software. BIP 110 has code but no miners. DOG Mode has miners’ theoretical interest but no code. The stalemate continues, with block space filling at market rates regardless of what either side prefers.
