{
  "id": 16168562,
  "title": "Anthropic says it partnered with Theorem to use Claude to formally prove security properties at the assembly level across 100+ real-world open-source libraries (Anthropic)",
  "url": "https://urgent.news/2026/10/23/anthropic-says-it-partnered-with-theorem-to-use-claude-to-formally",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-23T09:15:03.000Z",
  "source": {
    "name": "Techmeme",
    "slug": "techmeme",
    "url": "https://www.anthropic.com/research/formal-verification"
  },
  "original_language": "en",
  "account": "Anthropic has announced a partnership with Theorem, utilizing Claude AI to formally verify security properties at the assembly level across over 100 real-world open-source libraries. This marks the first large-scale effort to apply AI to formal verification at such a granular level. Claude not only proved these security properties but also uncovered bugs that had gone undetected for years, despite extensive testing. As AI models continue to advance, formal verification is expected to become a crucial aspect of secure software development, even for well-tested software like the Linux kernel, which has recently seen over 1,000 vulnerabilities. Formal verification involves translating code into a mathematical model and then proving that it adheres to a defined set of security rules. While this process is complex and requires significant expertise, LLMs like Claude are making it more accessible. The collaboration demonstrates that frontier models can autonomously generate correctness proofs for real-world software, providing a roadmap toward ensuring bug-free code. Anthropic and Theorem are publishing a subset of these proofs and offering a Claude skill to help developers begin their own verification processes.",
  "summary": "Introduction — Even mature, heavily tested software contains bugs, and LLMs are increasingly adept at finding them.",
  "key_points": [
    "Anthropic partners with Theorem to use Claude for formal verification",
    "Claude verifies security properties at assembly level in >100 open-source libraries",
    "AI uncovers undetected bugs in years-old software like Linux kernel"
  ],
  "editors_take": "Applying frontier models to granular assembly-level verification transforms complex mathematical proof generation into an accessible tool for uncovering long-standing vulnerabilities within critical open-source software libraries.",
  "illustration": null,
  "coverage": {
    "outlets": 4,
    "also_reported_by": [
      {
        "outlet": "Live Mint",
        "title": "Claude called itself a ‘high functioning machine’: Monk Swami Sarvapriyan recounts late-night AI chats at Anthropic",
        "url": "https://urgent.news/2026/10/22/claude-called-itself-a-high-functioning-machine-monk-swami",
        "published": "2026-10-22T10:47:35.000Z"
      },
      {
        "outlet": "The Economic Times",
        "title": "Anthropic’s new AI app Claude Duet makes debut",
        "url": "https://urgent.news/2026/10/22/anthropics-new-ai-app-claude-duet-makes-debut",
        "published": "2026-10-22T11:48:02.000Z"
      },
      {
        "outlet": "CBS News",
        "title": "Anthropic previews Claude Duet personal AI agent for subscribers",
        "url": "https://urgent.news/2026/10/23/anthropic-previews-claude-duet-personal-ai-agent-for-subscribers",
        "published": "2026-10-23T14:52:44.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}