Artel, I think you hit the nail on the head. It’s obvious the 4:1 witness discount is fundamentally arbitrary. I don’t see a physical justification for assigning permanently committed bits different economic weights with respect to the actual physical cost Bitcoin must bear. The cost of Bitcoin is ultimately paid in conserved energy and permanent memory, not in our software abstractions. This is the problem when our understanding of Bitcoin is not defined in the physics of work (joules) and memory (bits), but thru fiat interpretations. Bitcoin physically defines money with respect to energy, time and information (memory). Once you pull the logic to limit you realize that the second law of thermodynamics expressed in Bitcoin states that the 2.1 quadrillion satoshis will tend to disperse across an increasing number of UTXOs, unless there is sufficient economic pressure and energy to re-consolidate them. That pressure should emerge naturally from an unmanipulated fee market rather than from arbitrary discounts embedded into consensus accounting. The illusion comes from separating the active UTXO set from the permanent ledger. Yes, a spent output can be removed from the active UTXO set, but the witness and every other committed bit remain permanently embedded in the chain. Every archival node must store them forever. Every new node must download them forever. Every copy of Bitcoin must preserve them forever. The distinction is one of active validation state, not physical memory. Viewing Bitcoin through Landauer’s principle changes the discussion. The dominant cost is not validating the current state; it is the irreversible commitment and preservation of information. Once a bit crosses the proof-of-work boundary, it becomes part of Bitcoin’s permanent physical state. Consensus commits serialized memory, not CPU cycles. Expanding effective blockspace through witness discounts is a high-time-preference solution. Bitcoin conserves demand for finite blockspace. That cost is paid either immediately in satoshis through the fee market or deferred as a perpetual tax on every node that must transmit, download, verify, and preserve the additional physical bits for as long as Bitcoin exists. Looking at the system through energy density makes the distortion clearer. Whether expressed as joules per satoshi per mb^2, joules per mb^2, the total proof of work securing the block remains unchanged while the amount of permanently committed memory increases. The same work is now spread across more conserved bits. The energy density of Bitcoin’s permanent memory decreases. UTXOs are less energetically dense because the same conserved work is distributed across more conserved memory. This is the deeper engineering issue. We have altered the economics of blockspace by assigning different prices to identical permanently conserved bits. That distinction is based on how software uses those bits after commitment, not on the physical reality that every committed bit occupies immutable consensus memory. If Bitcoin is fundamentally a conserved ledger of energy and information, then every permanently committed bit must ultimately be measured by the same invariant accounting. This also highlights something deeper. To conserve the coin, Bitcoin must first conserve the bit. The distinction between “active state” and “historical state” has physical meaning. While an output remains unspent, the satoshis exist as physical bits within the active UTXO set. Once spent, the value simply changes location in memory to the new block of time. The spent TXO disappears from the active state, but the historical bit remains permanently embedded in the ledger. The state has transformed, but nothing has disappeared. The coin has moved through both memory and time, while the information required to preserve that history remains conserved forever. The cost remains as bits. This is precisely the reinterpretation of Landauer’s principle that Bitcoin makes possible. The original formulation measures the minimum energetic cost of erasing a bit. Bitcoin demonstrates that the more fundamental quantity is not erasure, but irreversible state transition through permanent memory. The thermodynamic cost is the cost of committing a new distinction (bit) to history. Once committed, that distinction is never destroyed; it becomes part of the conserved informational substrate of the ledger. What appears to be “deletion” is simply a transformation between active and historical state from a perspective of incomplete information. Looking at Bitcoin this way changes the engineering tradeoff entirely. The cost that would ordinarily be expressed through a competitive fee market for a finite bitspace has instead been partially shifted onto every archival node through the perpetual obligation to preserve additional physical bits. The network still bears the cost; it has simply been socialized across every node that chooses to maintain Bitcoin’s complete history. From a low-time-preference perspective, the honest engineering solution is to pay that cost upfront through fees in sats because it acknowledges the true physical cost of permanently committing information. 1 bit must = 1 bit. Once proof of work commits a bit to consensus, its future role inside the protocol becomes irrelevant to its physical cost. Whether the protocol later classifies it as witness data, transaction data, or historical validation data, every archival node must transmit it, verify it, download it, back it up, and preserve it indefinitely. Consensus does not preserve “validation effort”; it preserves immutable physical information. Removing the witness discount is simply an acknowledgement that permanently conserved bits should be accounted for uniformly; the fee market is the free exchange pricing of sats<>bits. The protocol may distinguish between active and historical state for purposes of validation, but the laws of thermodynamics do not; 1 bit = 1 bit regardless if satoshis are present in said memory. If Bitcoin is fundamentally a conserved ledger of energy, information, and time, then every permanently committed bit contributes equally to the irreversible memory of the system. Without conserving the bit, there is no mechanism by which the coin itself can remain conserved. Witness or not, if my node must store it forever, it is consensus memory. I simply want every permanently committed bit to my node to pay the same price because the physical cost of preserving it is the same. We have to be logically consistent. Why is market manipulation thru arbitrary discounting of bits and inflation of Bitcoin’s finite bitspace good in the long term for the protocol and economics, yet the same behavior in fiat land would be mocked by our community? Make it make sense to me. If Segwit/LN is truely good for scaling finite blockspace, then why do we have to tax the nodes? Let the protocol change speak for itself without taxation in bits. Only then we’ll know its true value. Devalue the bit and you devalue the sat; few truly understand this.