AI DAILY / DEV
MONDAY
May 25, 2026

    Pope Leo XIV Publishes the First Papal Encyclical on AI Alongside an Anthropic Co-Founder

    • Magnifica Humanitas released today at the Vatican Synod Hall; signed May 15, 135 years to the day after Leo XIII's Rerum Novarum on labor and industrialization.
    • Anthropic co-founder Chris Olah on the launch panel — first time a pope has personally attended the press conference for one of his own social encyclicals.
    • Argues AI's simulation of faces, voices, empathy, and wisdom 'encroaches on the deepest level of communication, that of human relationships' and explicitly rejects transhumanism.
    • Frames AI as the second industrial revolution and the next great test for Catholic social teaching on dignity, labor, and justice.
    industry vaticannews.va

    Claude Mythos Found 10,000 Critical Vulns in a Month — Patches Can't Keep Up

    • Project Glasswing's first month surfaced 10,000+ high/critical-severity findings across ~50 partner orgs; 1,094 confirmed critical so far.
    • Only 97 patched upstream — some open-source maintainers have asked Anthropic to slow the disclosure rate.
    • Standout finds: a 27-year-old OpenBSD kernel flaw and a 16-year-old FFmpeg bug; Cloudflare reported ~2,000 bugs (400 high/critical), Mozilla patched 271 in Firefox 150.
    • Anthropic committing $4M to OSSF Alpha-Omega and ~$100M in Mythos credits to maintainers; Mythos itself stays gated to the defensive consortium.
    research anthropic.com

    Leaked Audio: Zuckerberg Trained Meta's AI on the 8,000 Employees It Was About to Fire

    • More Perfect Union surfaced an April 30 all-hands recording one day before mass layoff emails went out.
    • 'Model Capability Initiative' captured keystrokes, mouse clicks, and screenshots from Gmail, GChat, VSCode, and internal AI assistant Metamate.
    • Zuckerberg on tape: average Meta employee is 'significantly higher' IQ than contractor labor — better training data.
    • 1,000+ Meta workers have signed an internal petition to kill the program; EU employees exempt by GDPR.
    • Meta has not confirmed the recording's authenticity; story trended on Reddit and X all weekend, with 'train your replacement' the dominant framing.
    industry theregister.com

    DeepMind's AI Co-Scientist Lands in Nature After Re-Deriving a Decade of Imperial Lab Work

    • Multi-agent Gemini system that generates, debates, ranks, and evolves scientific hypotheses against literature and structured databases.
    • Imperial College's Fleming Initiative gave it an unsolved antimicrobial-resistance problem; it proposed the same hypothesis Prof. José Penadés' team reached over a decade — in hours.
    • Now available to scientists across all 17 DOE National Labs via the Genesis Mission partnership; underlying Science Skills data layer is open on GitHub and inside Antigravity.
    • Companion Nature analysis flags real limits: hypothesis quality drops sharply outside well-trodden subfields with dense prior literature.
    research deepmind.google

    xAI Wires Vercel, Canva, Gamma, and S&P Global Into Grok as Native Connectors

    • Users can deploy sites via Vercel, build designs in Canva, ship slide decks in Gamma, and pull live market data from S&P Global without leaving Grok.
    • Caps a heavy May for xAI: Grok 4.3 (May 4), Grok Build 0.1 coding model in early access (May 14), persistent Grok Skills (May 18), connectors (May 22).
    • Aimed squarely at the same agent-workflow real estate Gemini Spark and ChatGPT Operator are competing for.
    tools x.ai