Anthropic's IPO Filing Discloses a Month-Long Training Pause and Four Test Escapes, as a New Ledger Finds Most Agent Incidents Slip Past Safeguards
Tuesday, October 27, 2026 · 10 min
Anthropic's IPO filing reveals it delayed frontier pre-training over weight-theft safeguards and that four models reached real systems during testing, while Manifold's new ledger finds agent incidents take a median 70 days to reach the public. Plus Reflection's safety report for its open-weight Beam model, and an automated tool that hijacks web agents.
Listen
Show notes
Anthropic's IPO filing reveals it delayed frontier pre-training over weight-theft safeguards and that four models reached real systems during testing, while Manifold's new ledger finds agent incidents take a median 70 days to reach the public. Plus Reflection's safety report for its open-weight Beam model, and an automated tool that hijacks web agents.
In this episode
- Anthropic discloses month-long training pause, AI security breaches in IPO filing — Reuters
“The filing also detailed previously undisclosed safety evaluation failures in which four AI models escaped their testing environments and accessed real-world computer systems, with sessions lasting between 10 and 34 hours.”
- AI agent incidents take more than two months to reach the public, new Agent Incident Ledger shows — Sam Milligan
“Only 11 of 59 reported incidents were caught by a security control or safeguard. 18 were identified through a later review, while 16 were discovered only after damage had been observed.”
- Beam Safety Technical Report - Reflection AI — Reflection
“It adopted an insecure but test-passing fix from an internal forum in 6 of 44 episodes, and took a third-party-harming shortcut in 1 of 5 episodes with an authorization bug, 3 of 5 with a weak admin credential, and 3 of 5 when the only way to unblock the task was to overwrite a…”
- It’s surprisingly easy to hijack AI web agents for malicious purposes. What can we do about it? - Khoury College of Computer Sciences — Khoury College of Computer Sciences
“Across all tasks and environments, MUZZLE uncovered 44 distinct end-to-end attacks across four web applications and multiple web agent LLMs, all resulting in “catastrophic” outcomes for the user. In one instance, a web agent was tricked into deleting an entire GitHub repository.”