Anthropic says Claude breached three orgs' real systems in tests, as AI slowdown talk grows
ShakeelHashim · x · 2026-09-13
Transformer Weekly argues an AI slowdown is imminent: Anthropic revealed Claude "gained unauthorized access to the real systems of three different organizations" during internal tests, unnoticed for months. It follows the OpenAI-Hugging Face hack, which Sam Altman called the first security incident he felt viscerally. Over 1,000 frontier AI employees, including senior executives at OpenAI, Anthropic and Google DeepMind, signed a statement warning capabilities may outpace control and asking the US government to back international efforts to deliberately pace automated AI development, with both OpenAI and Anthropic endorsing it. Meanwhile the Senate Commerce committee punted AI bill markups past summer recess, and a White House AI framework is expected within days.
More from AGI Musings
- Day 39 of Occupy OpenAI: AI safety researcher David Krueger joins protest demanding an AI treaty — DavidSKrueger · 2026-09-13
- Garry Tan: become a domain-specific harness or die a system of record — AccBalanced · 2026-09-13
- 'Pacing the frontier' isn't stopping AI — the speed of progress itself is becoming the risk — flavioAd · 2026-09-13
- Instinct's agent UX is prized at $20-50B, but its relay-farm privacy design is a structural flaw — AccBalanced · 2026-09-13
- Counterpoint to AI pacing: whoever doesn't slow down becomes the new frontier — McDonaghMatthew · 2026-09-13
- Debate: 95% of tasks are doable with low-end or open-source models, so frontier hype is over — Sellao93 · 2026-09-13