Research. Without the noise.
Independent reporting on Pearl research, with clear explanations and a considered editorial perspective.
What if the work
was useful?
The research behind proof of useful work from arbitrary matrix multiplication.
Read the analysisResearch, in perspective.
Pearl’s wager: one computation, two purposes
A closer look at the attempt to put matrix multiplication to work for both AI and a monetary network.
The hard part of useful work is keeping it fair
Why arbitrary matrix multiplication matters — and why a useful answer alone is not enough for permissionless consensus.
Who benefits when computation earns twice?
A model of mining and inference shows why the destination of the subsidy matters as much as its existence.
Hawkeye asks a harder question than “close enough”
Reproducing GPU arithmetic on a CPU could make independent inspection more concrete — within a carefully defined scope.
Behind “400×”: the cost of moving data
Pearl’s hashing note puts the spotlight on an often overlooked part of compute performance: where the bytes live.
Making smaller numbers do more work
Hadamard transforms, quantization and the practical challenge of fitting new computation into an inference service.
These are real publications from Pearl Labs’ research directory. Our articles distinguish published findings, our editorial perspective, protocol activation and what the explorer can verify.