WhatA platform that benchmarks AI systems by comparing their problem-solving efficiency directly against calibrated human performance on the same tasks, producing a meaningful 'human-equivalent' score rather than abstract percentages.
SignalMultiple commenters struggled with the puzzles themselves and joked about not being AGI, while others noted the scoring now measures efficiency relative to human solvers — there's clear appetite for AI evaluation that's grounded in human-comparable metrics rather than opaque leaderboard numbers.
Why NowARC-AGI-3's new scoring function (square of efficiency ratio vs. second-best human) pioneers human-relative scoring, and enterprises need this kind of interpretable metric as they make buy/build decisions on AI capabilities.
MarketCTOs and AI leads at enterprises evaluating AI vendors; adjacent to the $3B+ AI governance/evaluation market; Patronus AI and Arthur AI don't offer human-calibrated reasoning benchmarks.
MoatProprietary dataset of calibrated human performance baselines across reasoning tasks creates a unique reference standard that's expensive to replicate.
Automated EU Legislative Threat Monitoring for Tech CompaniesP6/10A SaaS platform that continuously monitors EU legislative proposals, amendments, and council votes that impact tech companies' products, and generates compliance impact assessments with actionable timelines.
Privacy-First Self-Hosted Communication Suite for EuropeansC5/10A turnkey, self-hostable communication platform (chat, file sharing, video) designed for non-technical users and small businesses who want to keep data entirely off third-party clouds.
Civic Engagement Platform for EU Digital Rights AdvocacyC5/10A mobile app that makes it dead-simple for EU citizens to identify their MEPs, auto-generate personalized messages on active digital rights issues, and track legislative outcomes — a 'one-tap lobby' for privacy.
Local-First AI Video Generation Desktop AppP6/10A desktop application that packages and optimizes open-source video generation models for local execution on consumer GPUs, removing content restrictions and API costs.
Killed by AI — Product Shutdown TrackerC5/10A community-maintained tracker documenting every AI product and feature that gets shut down, with timelines, dependency warnings, and migration guides for affected users.
AI Platform Risk Scoring for EnterpriseC6/10A B2B SaaS that continuously monitors AI vendor stability — financials, product churn, API deprecations, leadership changes — and generates risk scores to help enterprises decide which AI platforms to build on.