A new economic model maps when proof of useful work can improve compute supply

Proof of useful work creates two linked markets: demand for the network’s coin and demand for the computation. Critics often argue that useful output subsidizes an attacker. The paper models equilibrium prices instead of assuming useful work makes security free.

The paper’s only author, Rafael Pass, lists Pearl Research Labs as his affiliation (he is also a faculty member at Cornell Tech and the Technion) and says the work was done while consulting for the company, which lists it among its own research. Read it as the project’s economic case, carefully argued, rather than as an independent evaluation.

Under the model, Pass concludes that the cost of a majority attack stays tied to the block reward and, at the same token-to-inference ratio, is at least as high as in classic proof of work. The paper also finds a regime where low overhead and enough token adoption increase the supply of socially useful computation.

• Measure actual proof overhead instead of assuming it is negligible.
• Measure real inference demand against the size of token incentives.
• Separate mining-only work, inference-only work and integrated “duplex” work.
• Test whether observed market behavior fits the model’s assumptions.

An economic possibility result is not an adoption forecast. Pearl still has to reach the favorable regime in the real world.