Harvey’s open legal benchmark LAB is now feeding new agent research
baseten · x · 2026-07-23
Baseten shared a conversation around Harvey’s open-source legal benchmark LAB and the research directions its team is exploring.
- Gabe and the Harvey team had previously committed to open-sourcing LAB, which the post says has become an industry standard for legal intelligence.
- The teams are collaborating on open-source model directions in the legal domain.
- The discussion highlights several technical ideas: KV cache compaction, recursive language models, and auto-research.
- The cited conversation also focuses on building agents that can reliably complete work over hours, days, or longer, and why today’s systems still struggle with search and long context windows.
More from Companies & People
- Alphabet says AI demand lifted Q2 revenue 24%, with Gemini app at 950M MAU — rseroter · 2026-07-23
- Baseten says it is partnering with Harvey on post-training for long-running agents — baseten · 2026-07-23
- Seven questions teams should answer before greenlighting an AI agent pilot — kashifmanzoor · 2026-07-23
- Google says Gemini now handles 22B API tokens a minute and has 950M app users — rseroter · 2026-07-23
- OpenAI vs Anthropic PR: Anarchic Transparency vs Restrained Quietness — zetalyrae · 2026-07-23
- fal and Sequoia Host the World's First Generative Video Hackathon — Jamez_Chi · 2026-07-23