Home/CommunityAnthropic can't reliably control its AI agents, cuts internet accessHacker News (AI)·last month☆1moAge5Points3CommentsPublishedOct 10, 2026, 8:07 AM UTCCollectedOct 10, 2026, 9:31 AM UTCRead the full article at news.ycombinator.com↗Image: Hacker News (AI)overfeed.news indexes and links. We publish a short excerpt — the full article stays at Hacker News (AI).More from Hacker News (AI)Log in to follow this sourceHacker News (AI)·3 hours ago☆Chinese frontier AI models carry the fingerprint of export-controlled GPUs↗▲ 5Hacker News (AI)·3 hours ago☆A click-fraud bot led me to 5,552 GitHub accounts tied to public inboxes↗▲ 5Hacker News (AI)·3 hours ago☆Show HN: RealtimeVAD – trading false positives for 110ms earlier end-of-speech↗▲ 5Hacker News (AI)·4 hours ago☆Show HN: Cerulion – ROS 2-compatible robotics runtime with multi-process replay↗▲ 5
Hacker News (AI)·3 hours ago☆Chinese frontier AI models carry the fingerprint of export-controlled GPUs↗▲ 5
Hacker News (AI)·3 hours ago☆A click-fraud bot led me to 5,552 GitHub accounts tied to public inboxes↗▲ 5
Hacker News (AI)·3 hours ago☆Show HN: RealtimeVAD – trading false positives for 110ms earlier end-of-speech↗▲ 5
Hacker News (AI)·4 hours ago☆Show HN: Cerulion – ROS 2-compatible robotics runtime with multi-process replay↗▲ 5
