Skip to content
← All news

September 22, 2026

8 curated links on AI agents, Python, security, Google Cloud, and engineering, selected by Médéric Hurier.

  1. Introducing Grok 4.7

    x.ai

    xAI releases Grok 4.7, its flagship frontier model for coding and knowledge work designed to deliver twice the speed at half the price of comparable models.

  2. MiMo-V2.6

    mimo.xiaomi.com

    Xiaomi introduces the MiMo-V2.6 model series, delivering frontier multimodal intelligence while livestreaming its massive-scale reinforcement learning post-training in public.

  3. An ex-OpenAI researcher just deleted language from the LLM...

    youtube.com

    Fireship breaks down Jev, a decision-only "System 1" AI model by ex-OpenAI researcher Diogo Almeida that eliminates language generation to execute actions 200 times faster without hallucinations.

  4. Harness Engineering Explained – Inside the Stack Behind Antigravity, Claude Code & Cursor

    youtube.com

    Google Cloud Tech breaks down agent harness architecture with Ryan Lopopolo, exploring the execution environments and engineering principles powering tools like Antigravity, Claude Code, and Cursor.

  5. GLM-5.3 delivers Opus 4.8-level cybersecurity results at a fraction of the cost

    semgrep.dev

    Semgrep benchmarks Z.ai's GLM-5.3 against real-world vulnerability datasets, showing it achieves Opus 4.8-level cybersecurity detection performance at a fraction of the cost.

  6. Strengthen your CI/CD pipeline with new Secure Source Manager capabilities

    cloud.google.com

    Google Cloud expands Secure Source Manager to protect software supply chains with unified authentication and authorization across source repositories and CI/CD pipelines.

  7. Arena.ai on X

    x.com

    Arena.ai reports that Xiaomi's MiMo-V2.6-Pro debuted in the top 10 on the Code Arena WebDev leaderboard with 1628 points, ranking third among MIT-licensed open-weights models.

  8. Jensen Huang – The Doomer Hoax, Superintelligence is Here, and The Future of AI

    youtube.com

    Jensen Huang joins the All-In Podcast to discuss AI safety debates, superintelligence timelines, open-source model strategies, and global compute infrastructure.