Skip to content
← All news

September 23, 2026

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

  1. Introducing GPT-6 Sol and Luna

    openai.com

    OpenAI releases GPT-6 Sol and GPT-6 Luna, delivering its next-generation frontier reasoning and high-throughput model tiers with massive price reductions and million-token context windows.

  2. Introducing Claude Opus 5.5

    anthropic.com

    Anthropic introduces Claude Opus 5.5, its flagship model delivering frontier coding and knowledge work performance at 40% lower operating cost than Opus 5.

  3. Google AX – Open Agentic Orchestration Runtime

    github.com

    Google open-sources AX, a declarative, Kubernetes-like agentic orchestration runtime designed to manage workspaces, network fencing, and sandboxed execution for autonomous agents at scale.

  4. Jev’s Architecture Unmasked

    archerhume.com

    Archer Hume analyzes Jev across 10,000 API calls, reverse-engineering how its decision-first, language-free architecture replaces text generation with direct probability readouts from internal model representations.

  5. Why I couldn't build Jev at OpenAI – Diogo Almeida, TypeSafe Co-founder & CEO

    youtube.com

    TypeSafe co-founder Diogo Almeida joins Latent Space to discuss Jev, explaining why frontier labs struggle to prioritize low-latency, code-first decision models over verbose chat-based LLMs.

  6. Amazon blocks Meta's Muse AI assistant in new standoff over agentic shopping

    geekwire.com

    Amazon cuts off Meta's Muse personal AI agent from shopping on its platform over data security and customer relationship concerns in the emerging battle for agentic commerce.

  7. Instinct’s powerful AI assistant is raising privacy and security concerns

    techcrunch.com

    Early testers and security researchers raise privacy and safety concerns over viral AI assistant Instinct, citing broad account access, unprompted email actions, and persistent background data retention.