News Summary

The Laude Institute has announced its inaugural batch of Slingshots AI grants aimed at enhancing the science and practice of artificial intelligence. This program serves as an accelerator for researchers by providing resources often unavailable in standard academic settings, such as funding, compute power, and product engineering support. In return, participants are expected to output a tangible work product, which could be a startup or an open-source codebase. The first cohort comprises 15 projects focusing on AI evaluation, including familiar initiatives like Terminal Bench and ARC-AGI. New projects such as Formula Code and BizBench offer innovative approaches to AI agent evaluation, while CodeClash, led by SWE-Bench co-founder John Boda Yang, introduces a competition-based framework for code assessment. This initiative is poised to significantly impact AI evaluation practices.