Laude Institute Announces First Batch of ‘Slingshots’ AI Grants

Why It Matters

By concentrating resources on evaluation research, the Slingshots grants could standardize how AI performance is measured, reducing dependence on proprietary benchmarks and accelerating trustworthy AI development across the industry.

Summary

The Laude Institute unveiled its inaugural Slingshots AI grant cohort, funding 15 research projects that tackle the notoriously hard problem of AI evaluation. The program supplies resources—cash, compute, and engineering support—in exchange for deliverables such as startups, open‑source codebases, or other tangible artifacts. Projects range from familiar benchmarks like Terminal Bench and ARC‑AGI to novel efforts such as Formula Code’s code‑optimization test, Columbia’s BizBench for white‑collar agents, and John Boda Yang’s competition‑driven CodeClash. Recipients are expected to produce a concrete work product by the program’s end, aiming to push both scientific understanding and practical tools for assessing AI systems.