Key Takeaways
- Laude Institute launched the ‘Slingshots’ AI grants to support AI research.
- The program provides resources like funding and engineering support.
- Fifteen projects are part of the initial cohort, focusing on AI evaluation.
- Notable projects include Terminal Bench, ARC-AGI, Formula Code, and BizBench.
Laude Institute’s ‘Slingshots’ Initiative
The Laude Institute has introduced the ‘Slingshots’ AI grants to push forward the boundaries of artificial intelligence research. Announced on Thursday, the grants are designed to act as an accelerator for researchers by providing essential resources that are often out of reach in standard academic settings. These resources include funding, compute power, and product and engineering support.
Focus on AI Evaluation
The initial batch of grants supports fifteen projects, with a strong emphasis on the challenging task of AI evaluation. Among these projects are some well-known initiatives like Terminal Bench and the ARC-AGI project. These projects aim to tackle the complexities of measuring AI performance and effectiveness.
Notable Projects
- Terminal Bench: A command-line coding benchmark.
- ARC-AGI: A long-standing AI evaluation project.
- Formula Code: Developed by CalTech and UT Austin, this project focuses on evaluating AI agents’ ability to optimize code.
- BizBench: A Columbia-based initiative proposing a benchmark for “white-collar AI agents.”
Innovative Approaches
The grants also support innovative approaches to AI research, such as new structures for reinforcement learning and model compression. John Boda Yang, co-founder of SWE-Bench, leads the CodeClash project, which uses a dynamic competition-based framework to assess code. Yang emphasizes the importance of maintaining diverse and comprehensive benchmarks to drive progress in AI.
FAQ
What is the purpose of the ‘Slingshots’ AI grants?
The grants aim to advance AI research by providing resources like funding and engineering support, which are typically unavailable in academic settings.
How many projects are included in the first cohort?
The initial cohort includes fifteen projects, focusing primarily on AI evaluation.
Who is involved in the CodeClash project?
John Boda Yang, co-founder of SWE-Bench, leads the CodeClash project, which uses a competition-based framework to assess code.
Summary
The Laude Institute’s ‘Slingshots’ AI grants represent a significant step in supporting AI research, providing crucial resources to projects that address the complexities of AI evaluation and development. For businesses managing their own software systems, staying informed about advancements in AI evaluation can offer insights into optimizing internal processes and improving overall efficiency.
Source: Laude Institute announces first batch of ‘Slingshots’ AI grants – techcrunch.com