エピソード

  • ARM vs. x86: The Compatibility Question

    ARM vs. x86: The Compatibility Question
    2026/09/03
    ARM versus x86 is often framed as efficiency against power. This episode takes a more useful approach: what does your software need to run? Ethan Cole explains instruction-set targets, the software stack around an application, and why workload requirements, legacy dependencies, and specific platform features can shape an architecture decision. Using Arm and Intel documentation alongside Scaleway’s stated cloud guidance, this is a compatibility-first guide—not a claim that either architecture wins every benchmark, cost, security, or energy comparison. Practical takeaway: define the workload, identify the software and platform constraints, then validate the intended environment before a processor choice becomes a costly deployment issue. Signal to Noise by Scott Buckley. CC BY 4.0. www.scottbuckley.com.au. Adapted for format, looping, and loudness. Watch the companion video: https://www.youtube.com/watch?v=lv6q6qENiZ4
    続きを読む 一部表示
    18 分
  • AI Agents Are Becoming Software Teammates

    AI Agents Are Becoming Software Teammates
    2026/08/25
    What changes when an AI system moves from answering questions to participating in workplace workflows? Ethan Cole explains AI agents through a hypothetical IT access-request scenario. The episode separates vendor ambitions from independent proof, clarifies the roles of workflows, state, and tool use, and explores why operational authority comes from the systems connected around a model. Before treating an AI agent as a software teammate, ask four questions: What is its objective? What can it read or change? Where are approval and escalation points? Can a person review the path it took? Watch the companion video: https://www.youtube.com/watch?v=JfuskhvW5aQ
    続きを読む 一部表示
    16 分
  • AI Office Automation: Who Gets Permission to Act?

    AI Office Automation: Who Gets Permission to Act?
    2026/08/25
    What changes when AI moves from drafting an email to sending it, updating records, processing payments, or triggering a shipment? Ethan Cole follows a hypothetical customer request through an office workflow to examine the real dividing line in AI automation: permission. Along the way, this episode distinguishes augmentation from automation, compares Mustafa Suleyman’s aggressive forecast with Newmark’s office-market outlook, and explains why task-level change is more revealing than broad claims about jobs disappearing. The practical takeaway: map the workflow, define which actions software may take, establish where human review begins, and measure the full result—not just the speed of a demo. Watch the companion video: https://www.youtube.com/watch?v=C4PeN6fcpDY
    続きを読む 一部表示
    19 分
  • AI in 2026: Who Controls the Assistant?

    AI in 2026: Who Controls the Assistant?
    2026/08/20
    AI’s next shift may not be better chatbots. It may be software gaining access to documents, calendars, approvals, and other workplace systems. This DigitalTechIQ episode examines the forecasts around AI agents, specialized models, multimodal tools, portability between providers, and ongoing monitoring. It also separates company outlooks from verified results—and applies that same standard to a reported 2026 quantum-computing prediction. The practical question is simple: before an AI system gets authority to act, can you see what it can read, send, change, and who approves it? Watch the companion video: https://www.youtube.com/watch?v=A7Nr69XphVk
    続きを読む 一部表示
    13 分
  • AI’s Next 5 Years: Useful, Hidden, or Accountable?

    AI’s Next 5 Years: Useful, Hidden, or Accountable?
    2026/08/19
    AI’s next phase may be less about the biggest demo and more about the systems quietly shaping customer service, recommendations, and workplace workflows. Ethan Cole examines 2026 expert commentary, reported usage estimates, and long-range projections to ask what the evidence actually supports. From frontier versus efficient models to agents, AI-enabled commerce, infrastructure, jobs, and accountability, this episode separates demonstrated outcomes from forecasts. Using a hypothetical late-delivery complaint as a guide, we end with five practical questions to ask whenever AI affects a service, recommendation, or decision: Was it disclosed? What task is it doing? What data is involved? Can the outcome be challenged? And is there evidence that it helps?
    続きを読む 一部表示
    14 分