The Laude Institute announces its first batch of AI “Slingshots” grants.

On Thursday, the Laude Institute announced the first batch of Slingshots grants, aimed at “advancing the science and practice of artificial intelligence.”

Designed as an accelerator for researchers, Slingshots aims to provide resources that may not be available in most academic settings, whether that be funding, computing power, or product and engineering support. In return, recipients pledge to produce some final work product, whether it’s a startup, an open source code base, or some other type of artwork.

The initial group consists of 15 projects, with a particular focus on the challenging problem of evaluating artificial intelligence. Some of these projects will be familiar to TechCrunch readers, including the standard Terminal Bench for command-line coding and the latest version of the long-running ARC-AGI project.

Others take a new approach to dealing with the well-established problem of evaluation. Formula Code, created by researchers at Caltech and the University of Texas at Austin, aims to produce an assessment of AI agents’ ability to improve existing code, while Columbia-based BizBench proposes a comprehensive standard for “white-collar AI agents.” Other grants explore new architectures for reinforcement learning or model compression.

SWE-Bench co-founder Jun Buda Yang is also part of the group, as the leader of the new CodeClash project. Inspired by the success of SWE-Bench, CodeClash will evaluate code through a dynamic competition-based framework, which is what Yang hopes to do

“I think people continuing to evaluate based on basic third-party standards leads to progress,” Yang told TechCrunch. “I’m a bit concerned about the future where standards become specific to companies.”