Three activities, one allocation problem
Rafael Pass models compute allocated to mining, inference or combined “duplex” work. The analysis characterizes equilibrium allocations and prices using duplex overheads and a parameter relating token adoption to the inference market.
The paper names three regimes: Bitconia, which resembles conventional proof of work; Fortessia, where duplex replaces mining while useful output remains unchanged; and Duplexia, where rewards subsidize inference and expand its supply. It also finds that, after equilibrium prices adjust, majority-attack cost remains connected to the block reward. These are conclusions within the model and its assumptions, not measurements of Pearl’s present economy.
A rebate does not stay in one pocket by definition
A hypothetical example helps explain the issue without forecasting a market. Suppose two otherwise identical providers can deliver the same service, but only one receives an additional payment associated with that work. That provider could retain the difference, offer a lower price, spend more on capacity or tolerate a lower utilization rate. Which outcome persists depends on competition and the constraints both providers face.
The extra revenue is therefore not sufficient information to calculate a customer’s future price or a miner’s profit. The calculation also needs costs, demand, overhead and the period being measured. Changing one assumption can change the comparison.
Security economics needs a time horizon
An equilibrium is an analytical state in which the model’s choices and prices fit together. A network in operation also experiences transitions: equipment is ordered, software is upgraded, workloads change and participants arrive or leave. A result about equilibrium does not supply a timetable for those adjustments.
An observer should therefore avoid mixing a long-run theoretical claim with a short-lived dashboard observation. Today’s payout concentration describes today’s attributed blocks. A model of attack incentives describes a different object. Each can inform a serious discussion, provided its scope remains visible.
A practical reading of network growth
The useful questions become more concrete when separated. Are more blocks being attributed to distinct payout addresses? Are providers reporting comparable end-to-end costs? Are customer workloads actually being served? Are those reports repeatable and independent? These questions require different datasets.
Pearl Pulse can help with the chain side: blocks, rewards, source-labeled producers and historical observations. It cannot derive inference revenue, customer retention or provider margins from a transaction count. An absence of that information is a measurement boundary, not a finding that those activities exist or do not exist. The model provides a vocabulary for asking better questions; operating evidence is still needed to answer them.
The Pearl Pulse view
We think the strongest growth story would be a service valued by people who do not care about mining. If rewards improve the compute offer, participation could spread through ordinary demand. That remains a conditional thesis, best judged through customer outcomes alongside chain statistics, rather than subsidy size alone.
Sources & editorial approach
Original reporting and explanation by Pearl Pulse. The editorial perspective is our interpretation; it is separate from the researchers’ findings. Links below lead to the primary sources.