PulseAugur / Brief
EN
LIVE 02:15:07

Brief

last 24h
[9/9] 221 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. TOOL · Email — AI Tool Report English(EN) ·

    ⚡️ Anthropic finds 10K flaws

    Anthropic's Project Glasswing, utilizing its Claude Mythos Preview model, has identified over 10,000 software vulnerabilities, with 1,094 confirmed as high or critical severity. A notable discovery was a critical flaw in WolfSSL, which has since been patched. This initiative highlights a significant advancement in defensive AI capabilities, prompting Anthropic to advocate for faster patch cycles and tighter deployment timelines. AI

    ⚡️ Anthropic finds 10K flaws

    IMPACT Accelerates the discovery of critical software vulnerabilities, potentially improving overall cybersecurity defenses.

  2. TOOL · Medium — Anthropic tag English(EN) ·

    Project Glasswing: An initial update reads like a warning shot, not a progress report

    Anthropic has released an initial update on Project Glasswing, a research initiative focused on AI safety. The update is framed as a cautionary report rather than a straightforward progress announcement. It suggests potential risks and challenges associated with the project's development. AI

    IMPACT Highlights potential risks in AI safety research, prompting caution for developers.

  3. SIGNIFICANT · Mastodon — fosstodon.org Polski(PL) · · [5 sources]

    Perplexity released Bumblebee – an open-source scanner for macOS and Linux that allows instant detection of malicious AI libraries and extensions on computers

    UC Berkeley Law is implementing strict AI usage policies starting in 2026, prohibiting students from using language models for academic work. Meanwhile, Google has launched its AI Mode in Poland, which uses Gemini 2.5 to provide direct answers instead of links, signaling a shift in search engine functionality. In cybersecurity, Project Glasswing's Claude Mythos has demonstrated an ability to find critical vulnerabilities at a pace exceeding human development teams, and Perplexity has released Bumblebee, an open-source tool for detecting malicious AI extensions on macOS and Linux. AI

    IMPACT AI's integration into education, search, and cybersecurity continues to evolve, prompting policy changes and new detection tools.

  4. TOOL · Mastodon — fosstodon.org English(EN) ·

    The attribution for Mastodon's CVE-2026-46349 (CVSS 5.3, retracted boost reissuance) is interestingly reported as: "This security issue has been reported by Doy

    A security vulnerability in Mastodon, CVE-2026-46349, with a CVSS score of 5.3, has been disclosed. The vulnerability, related to retracted boost reissuance, was reported by Doyensec in collaboration with Anthropic Research. This collaboration has led to speculation about undisclosed project affiliations, specifically Project Glasswing. AI

    IMPACT Speculation around AI collaboration in a software vulnerability disclosure.

  5. COMMENTARY · Mastodon — sigmoid.social 한국어(KO) · · [2 sources]

    AI Leaks and News (@AILeaksAndNews) announced plans to expand Project Glasswing to more partners and later release Mythos-class models to the public after implementing stronger safety measures. This concerns strengthening safety before deploying large-scale models and partner expansion strategies.

    Anthropic's Jack Clare shared optimistic predictions about AI's future capabilities. He anticipates AI systems will be able to design their successors by the end of 2028. Furthermore, Clare suggested that within the next 12 months, humans and robots could collaborate on Nobel Prize-worthy discoveries, highlighting rapid advancements and potential for autonomous research. AI

    IMPACT Highlights potential for accelerated AI development and human-robot collaboration in scientific discovery.

  6. SIGNIFICANT · 36氪 (36Kr) 中文(ZH) · · [18 sources]

    Google says it has discovered hackers using AI to develop zero-day exploit tools for the first time

    Google's Threat Intelligence Group has identified the first instance of cybercriminals using artificial intelligence to develop a zero-day exploit. This AI-generated tool was designed to bypass security measures in an open-source system administration tool, potentially for a large-scale attack. While Google successfully thwarted this specific attempt and notified the affected company, researchers believe this marks a significant escalation in AI-assisted cybercrime, with more sophisticated attacks anticipated. AI

    IMPACT Signals a new era of AI-powered cybercrime, potentially accelerating the discovery and deployment of sophisticated exploits.

  7. RESEARCH · HN — anthropic stories English(EN) · · [5 sources]

    We reproduced Anthropic's Mythos findings with public models

    Researchers have successfully replicated Anthropic's Mythos findings using publicly available AI models like GPT-5.4 and Claude Opus 4.6. This suggests that advanced AI capabilities for discovering software vulnerabilities are no longer exclusive to frontier labs and are becoming accessible through public models. The focus for defenders should now shift from the exclusivity of these tools to validating and operationalizing AI-generated security insights. AI

    We reproduced Anthropic's Mythos findings with public models

    IMPACT Confirms that advanced AI vulnerability discovery capabilities are becoming accessible via public models, shifting the focus to defense and operationalization.

  8. RESEARCH · Ben's Bites English(EN) · · [4 sources]

    Anthropic built a model too risky to release

    Anthropic has developed a new AI model named Claude Mythos, which demonstrates significant advancements in benchmark performance, particularly in identifying software vulnerabilities. Due to its advanced capabilities in finding and exploiting security flaws, Anthropic has opted not to release Mythos publicly. Instead, the company is providing limited access to select organizations through "Project Glasswing" to aid in cybersecurity research and vulnerability discovery, alongside a substantial commitment to open-source security initiatives. AI

    Anthropic built a model too risky to release

    IMPACT Restricted release of advanced AI model highlights growing safety concerns and the potential for AI in cybersecurity, influencing future development and deployment strategies.

  9. TOOL · OpenAI News English(EN) · · [4387 sources]

    Introducing OpenAI

    OpenAI has launched a preview of its Codex coding assistant within the ChatGPT mobile app, allowing users to manage coding tasks remotely across devices. The company is also highlighting how various organizations, including Ramp, NVIDIA, and AutoScout24, are leveraging Codex and GPT-5.5 for accelerated code review, faster development cycles, and AI-assisted research. Meanwhile, Anthropic's Project Glasswing initiative has identified over ten thousand high-severity vulnerabilities in essential software, emphasizing the need for the industry to adapt to AI-driven security analysis. AI

    Introducing OpenAI

    IMPACT Expands accessibility of AI coding assistants and highlights AI's role in identifying software vulnerabilities, potentially accelerating development and improving security.