OpenPriors · structured judgment
Manifund under the microscope Recent Manifund funding proposals, judged by
LLMs answering one small question many thousands of times: which of these two shows more of
this quality, and by how much? The default ranking is an Analytic Hierarchy Process built from nothing but those answers: how much each
quality matters to a funding goal, which direction is good, and how much each proposal shows
it. Pick a goal to re-rank the slate, pick one quality to rank by it alone, or open any card to
see exactly why it scores what it scores.
Loading judgments…
alive on the ledger
The same AHP, running as a product Everything above is a frozen study. This section is live: the August 8–15 cohort judged on
four canonical axes through the public judgements API , and the axes
themselves judged for importance under seven funder goals — the classical AHP weighting step
as just another cardinal run whose entities are the axes. Weights are exp(latent),
normalized; a proposal's score is the weight-summed z-score. Every number below links to the
board and run it came from.
Funder goal a careful generalist reducing existential risk near-term measurable welfare epistemic public goods hits-based leverage field-building technical AI safety
1 Catching Fooled AI Judges: The Correspondence Auditor v3 +3.4 +2.5 +8.4 +3.1 2 Aethel + FamilyClaw: proof-carrying execution for AI agents +1.2 +1.9 +5.2 +3.0 3 Can a One-Page Human–AI Evidence Check Catch Research Claims That Go Too Far +1.0 +1.7 +2.7 +2.6 4 TRACE-FV: Do verified AI corrections endure? Preregistered pilot +1.2 +1.5 +1.9 +2.5 5 Deterministic replay for robot fleet failures +0.1 +2.6 +1.5 +1.6 6 PRML +1.1 +1.3 +0.3 +2.1 7 Open-source trust rails for the agent economy -0.2 -0.7 +5.0 +1.4 8 Prediction of Inoculation Prompt Side Effects +2.2 +0.6 +0.2 +0.3 9 Substrate coupling in neural networks: instruments and falsifiable controls -0.4 -0.7 +3.4 +1.9 10 Context-Aware Defenses Against Indirect Prompt Injection in Agentic AI System +0.9 +1.5 +0.1 +0.1 11 A Multilingual Dataset and SLM for Automatic Coding of Cancers +0.8 +0.5 +1.5 -0.1 12 Independent Multi-Model AI Accountability Research by the EM Foundation +0.0 -0.6 +3.6 +0.4 13 Agent limits that survive delegation +1.2 +1.0 +0.5 -0.5 14 Auditable claim extraction and comparison of claims across scientific literature +0.6 +0.6 +0.4 +0.5 15 Open Speech Data for Endangered Northern Ghanaian Languages +0.6 +1.3 -0.2 +0.1 16 Travel Grant to Present My Mechanistic Interpretability Research at MICCAI 2026 -1.2 +0.7 +2.4 -0.1 17 Human oversight Framework for AI systems as a Governance Tool +1.4 +0.7 -0.1 -0.9 18 What Happens If We Actually Test It? -0.0 -0.0 +0.8 +0.8 19 Surrogate base model for Mechanistic Interpretability +0.1 +0.7 -0.3 +0.6 20 Nipping AI-Fabricated Science Claims in the Bud +0.5 +0.6 -0.6 +0.4 21 Agent Passport System +0.6 +0.2 -0.3 -0.1 22 Trace Continuity: Making AI Prove It Still Has Authority +0.6 -0.0 -0.0 -0.6 23 AI Literacy program for Epileptic Youth in Kenya, Africa. -0.1 +0.0 -0.3 -0.6 24 Evaluating the safety, ethics, and values of quantized and finetuned open models -0.2 -0.1 -0.4 -0.3 25 Exploring Dynamic Constraint Boundaries for Auditable AI -0.8 +0.2 -0.9 +0.4 26 Himalayan Peak Finder -2.4 +1.1 +0.1 -0.4 27 Solving the memory issue in AI -0.3 -0.3 -0.2 -1.0 28 Well-Capitalized Prediction Markets for Measuring AI’s Economic Impacts -0.8 -0.5 -0.5 -0.1 29 Does Consciousness Depend on the Brain or the Computation? -2.4 -0.8 +1.6 +0.1 30 Sentient Futures Project Incubator -0.0 -0.9 +0.0 -1.0 31 Analysing AI policies in Higher Education institution in Malawi -0.4 -0.4 -0.3 -1.3 32 programmatic internet search -0.6 -0.7 +0.5 -1.6 33 An Asymmetric Wager: Funding a 99th-Percentile Mind to Pivot into Computer Scien -0.8 -0.3 -1.1 -1.0 34 Guardians of the Digital Territory +0.2 -0.9 -0.8 -1.4 35 Developmental Continuity in Persistent AI Agents -1.4 -1.4 -0.8 +0.1 36 A self-evolving defense against increasingly complex agentic attacks -0.3 -1.8 -0.6 -0.9 37 The New Critic Longform Reporting Fund -1.7 -0.1 -0.9 -1.7 38 Preserving the Human Veto +0.0 -2.2 -0.7 -1.5 39 Generating and scoring the next emerging virus from its sequence -1.1 -2.7 -1.3 +0.5 40 AQI-Autonomous Quantum Intelligence -1.6 -2.4 -1.5 -1.3