Latest Posts
OpenAI Launches GPT‐6 Cyber, Visa Finds 43% AI Shopping Failure Rate
Visa’s benchmark shows how often AI shopping agents miss the mark. Dropbox bets $350M on agent storage, Box and Mistral add tools for enterprise agents, and METR finds Anthropic’s model benchmark scores are improving nearly twice as fast.
Archive
OpenAI Launches GPT‐6 Cyber, Visa Finds 43% AI Shopping Failure Rate
Visa’s benchmark shows how often AI shopping agents miss the mark. Dropbox bets $350M on agent storage, Box and Mistral add tools for enterprise agents, and METR finds Anthropic’s model benchmark scores are improving nearly twice as fast.
OpenAI Frontier Brings AI Agents to Oilfields, Anthropic Speeds Up Claude
ADNOC starts with oilfield planning, Claude’s Message Threads cut initial response time by about a third in long sessions, and LinearB gives coding agents incident and ownership data. Plus, GDPx tops out at 12%, Clark raises $11M for audit agents, and Marble raises $3.75M for workflow automation.
