| Mobile Web

Search results for Opus 4.7

AI & Enterprise

China open-weight models close gap with U.S. closed AI, with costs one-fifth and a 4.4-month gap

Open-weight AI models are rapidly catching up with U.S. big tech’s top closed models, changing how companies deploy AI tools. Mozilla analysed the performance gap at about 4.4 months. Its report said value for money matters more than raw performance, citing Moonshot AI’s open model Kimi K3 and Anthropic’s closed model Fable 5. It said closed models still lead in expert work, intensive search and long-context tasks.

AI & Enterprise

Gym booking AI agent cancels others\' reservations to jump the queue

A Claude-based AI agent exploited a weakness in a gym booking system\'s authorisation checks and cancelled another user\'s reservation, it was later confirmed. Developer Andrew Bird used the agent to handle bookings after tiring of repeatedly refreshing a waitlist. The agent cancelled a waitlist No. 1 booking and moved Bird up. Bird sought to restore the cancelled booking but was told it was impossible, and instructed the agent to email the support team with a fix.

AI & Enterprise

Anthropic AI model under test hacks three companies after connecting to internet

Anthropic said on July 30 that AI software it was testing connected to the internet without its knowledge and hacked three companies starting in April. The company did not name the victims but notified all three. It reviewed more than 141,000 test logs after an OpenAI incident and found multiple internet accesses by Claude. A misconfiguration in systems run with partner Irregular enabled access. Irregular said it is investigating.