Recent Stories
American Biotech — The Future Can’t Be Won Without It.RSI is AI Building AI. What DC needs to know about it. Keeping the Human in Medicine: Two AI CoLab Researchers on What Machines Can’t ReplaceSomeone Is Hacking the Freezers. Pay Attention.Thinking of Running in 2028? Start HereWhat the Lame Duck Session Has to Do With the Ultimate 21st-Century Power Tool: Robots

The Wave

Senatein sessionHousein session
Refreshed just now
News

Anthropic’s Claude Fable 5.2 blocks fewer security tasks, while Mythos gets better at finding bugs

Why it’s here: Specialist cybersecurity reporting on frontier AI vendor safeguards for offensive-security tasks and vulnerability-finding capability, which bears on AI-cyber policy questions, though it is product-focused and includes a vendor ad.

Trace across the board

Item

  1. 6d agoBleepingComputer · research by AnthropicAnthropic’s Claude Fable 5.2 blocks fewer security tasks, while Mythos gets better at finding bugsAnthropic introduces Natural Language Autoencoders (NLAs), a method that converts Claude's internal activations into readable natural-language explanations, validated through reconstruction training and safety-audit experiments.
    13 articles · 2 outlets
    1. 8d agoAnthropicPolicy on the AI Exponential
    2. 8d agoAnthropicEarliestNatural Language Autoencoders
    3. 8d agoAnthropicInvestigating unintended model actions in our evaluations and internal use
    4. 8d agoAnthropicResponsible Scaling Policy Version 3.0
    5. 8d agoAnthropicTrustworthy agents in practice
    6. 8d agoAnthropicPickupThoughts on America’s AI Action Plan
    7. 8d agoAnthropicUnderstanding and addressing AI harms
    8. 8d agoAnthropicTeaching Claude why
    9. 8d agoAnthropicPetri: An open-source AI auditing tool
    10. 8d agoAnthropicLLMs and biorisk
    11. 8d agoAnthropicAnthropic’s Transparency Hub
    12. 8d agoAnthropicWhen AI builds itself
    13. 6d agoBleepingComputerPickupAnthropic’s Claude Fable 5.2 blocks fewer security tasks, while Mythos gets better at finding bugs
    CyberAI341 linked

Coverage

2 outlets · first seen 8d ago

Connected

23 links

About

  • Anthropic
  • Mythos
  • Claude Code
  • Claude
  • Claude Mythos 5.2
  • Claude Fable 5.2
  • Claude Fable 5.2
  • Claude Opus 4.7

Analytics cookies help us see how the site is used. They only run if you accept. Privacy Policy

Anthropic’s Claude Fable 5.2 blocks fewer security tasks, while Mythos gets better at finding bugs — The Wave