エピソード

  • Model Hype, Medical AI and Brussels Policy

    Model Hype, Medical AI and Brussels Policy
    2026/08/28
    What matters is not whether this week’s major model release deserves the word milestone, but what it can actually do, for whom, and under what conditions. Jonathan Harris weighs the claims against the familiar machinery of Model Hype, where AI benchmarks often settle less than their authors suggest. A medical diagnostics paper warrants attention beyond its press release, while a Brussels policy move has been widely misread. The useful signal lies in the practical limits and incentives; the noise is the usual rush to declare victory before anyone has checked the wiring.
    続きを読む 一部表示
    3 分
  • AI Agents: Authority Not Just Ability

    AI Agents: Authority Not Just Ability
    2026/07/31
    The practical issue with artificial intelligence agents isn't just how clever they are, but what they're allowed to do. Jonathan Harris explains why governance needs to precede deployment, moving beyond benchmark scores to focus on precise control over access and actions. This week, we examine how the hype around agentic AI often distracts from fundamental questions of authority and security. We also look at the layers of engineering behind AI systems – prompt, loop, and graph – and why confusing them leads to costly misunderstandings. From industrial safety to the structure of the web, the real progress lies in architecture and careful control, not just more confident claims or faster models. Scepticism, as Jonathan Harris notes, is basic maintenance.
    続きを読む 一部表示
    1 時間 3 分
  • Agentic AI, Model Hype, and MCP and Agent Skills

    Agentic AI, Model Hype, and MCP and Agent Skills
    2026/07/24
    Jonathan Harris cuts through Agentic AI, Model Hype, MCP and Agent Skills, Dirty Data, and Workflow Automation in this 60-minute Turing’s Torch: Artificial Intelligence Weekly briefing. What matters is simple: what is useful, what is undercooked, and who carries the risk once the demo glow wears off. Expect plain-English context on power, money, data, labour and control, with the usual vendor fireworks left outside where they belong. Longer episodes also leave room for the awkward plumbing: incentives, security assumptions, governance gaps, and the budget line nobody wanted to read.
    続きを読む 一部表示
    1 時間
  • Model Hype, Dirty Data, and Workflow Automation

    Model Hype, Dirty Data, and Workflow Automation
    2026/07/10
    Jonathan Harris cuts through Model Hype, Dirty Data, Workflow Automation, and AI Governance in this 45-minute Turing’s Torch: Artificial Intelligence Weekly briefing. What matters is simple: what is useful, what is undercooked, and who carries the risk once the demo glow wears off. Expect plain-English context on power, money, data, labour and control, with the usual vendor fireworks left outside where they belong.
    続きを読む 一部表示
    25 分
  • AI Agent Skills and Model Hype

    AI Agent Skills and Model Hype
    2026/07/03
    What matters is discerning the practical utility of AI advancements from the prevailing hype. Jonathan Harris examines how new AI models and agent skills are being implemented, often revealing that the real challenge lies in the surrounding infrastructure and governance, not just the core technology. This week explores the tension between convenience and control, the complexities of decentralised AI, and the critical need for robust data engineering and benchmark hygiene to ensure genuine capability.
    続きを読む 一部表示
    24 分
  • MCP and Agent Skills, Retail Edge AI, and Workflow Automation

    MCP and Agent Skills, Retail Edge AI, and Workflow Automation
    2026/06/26
    Jonathan Harris cuts through MCP and Agent Skills, Retail Edge AI, Workflow Automation, and AI Governance in this 45-minute Turing’s Torch: Artificial Intelligence Weekly briefing. What matters is simple: what is useful, what is undercooked, and who carries the risk once the demo glow wears off. Expect plain-English context on power, money, data, labour and control, with the usual vendor fireworks left outside where they belong.
    続きを読む 一部表示
    23 分
  • Agentic AI and Vendor Lock-in

    Agentic AI and Vendor Lock-in
    2026/06/19
    What matters is understanding the practical implications of AI systems that act autonomously. Jonathan Harris examines how generative AI in planning and finance, alongside vendor partnerships, creates risks of lock-in and accountability issues. We separate genuine progress from the noise, especially concerning agentic AI which moves from advice to action, demanding robust governance and human oversight, not just technical solutions. The episode covers the real costs and consequences of deploying AI in core functions.
    続きを読む 一部表示
    17 分