Wednesday, July 22, 2026

1. PathAgentBench: Benchmarking Evidence-Seeking Vision-Language Models on Whole-Slide Pathology Image
   T2arXiv AI/CL/LG（论文）

   PathAgentBench is a benchmark specifically designed to evaluate evidence-seeking capabilities of vision-language models on whole-slide pathology images (WSIs). It covers four complementary abilities: cross-modal matching, diagnostic-region localization, multi-scale reasoning, and autonomous exploration. The benchmark includes 1,822 TCGA slides and 17,135 diagnostic paths annotated by ten certified pathologists. Evaluating 20 models reveals strong performance in multi-scale reasoning (>93% accuracy) but extremely poor results in region localization (IoU <0.09) and autonomous exploration (hit rate 0.020 at high magnification), exposing a pronounced gap between reasoning and evidence acquisition.

1. Synthesia Launches AI Roleplay Sessions for Enterprise Training, Moving Beyond Video to Live Coaching
   T2TechCrunch：AI（RSS）

   Synthesia has launched AI Roleplay Sessions, an interactive enterprise training platform where employees practice workplace conversations with AI avatars. The system provides real-time feedback, scoring, and analytics to help companies measure training effectiveness, marking Synthesia's expansion from video generation into live coaching.

2. Glow Emerges from Stealth at $1.2B Valuation to Challenge Endpoint Security in the AI Era
   T2TechCrunch：AI（RSS）

   Glow is a new cybersecurity startup focused on addressing novel endpoint risks introduced by AI agents and developer tools inside enterprises. The company emerged from stealth today with a $1.2 billion valuation, aiming to redefine security in the AI era. Its product uses AI to detect and block attacks targeting AI workflows in real time, and has already secured several large enterprise clients and top-tier venture backing.

1. Anthropic Donates Another $20 Million to Public First Action
   T1Anthropic：News（网页）

   Anthropic has announced an additional $20 million donation to Public First Action, a public policy research organization, bringing its total contribution to $40 million. This move underscores Anthropic's long-term commitment to AI safety and governance, supporting external policy research to drive responsible AI development. The organization focuses on issues such as AI regulation and labor market impacts, and this funding will expand its research capacity and influence policy-making.

2. Passionfroot Raises $15M to Expand B2B Creator Marketplace to the US
   T2TechCrunch：AI（RSS）

   Passionfroot, a German startup building a marketplace that connects B2B creators with brands, has raised $15 million in a Series A round led by Insight Partners. The funds will be used to expand into the US, enhance product development, and grow the team. The platform focuses on B2B creator collaborations, differentiating from traditional consumer-focused influencer marketing.

—

—