explainable-ai

You Can't Tune a Product — Show Notes You can tune a piano — the strings obey the wrench. You cannot tune a generative product, because its behavior is a trained prior, not a setting. This special edition tells the first-person story of the RCD pipeline itself: months of prompt detuning that failed to stop NotebookLM's "collision" dramaturgy, the forensic that proved it, and the pivot to verify-then-author — a Whisper-plus-LLM fidelity gate, then taking the pen entirely with verbatim two-host TTS. ...

The Bounds of the Cloned Voice In 2024, journalist Evan Ratliff cloned his own voice, wired it to a chatbot, and let it loose on the phone — and the thing that gave it away was a pause. This deep dive traces where that pause went. By 2026 the technical tells Ratliff relied on — latency, woodenness, the audible seams — have largely collapsed, which means the real boundary of the cloned voice has moved off the machine and onto us: our psychology, our laws, and the economics of who can afford to fake whom. The defense that still holds isn't a better detector. It's whether you and the person on the other end agreed on a code word in advance. ...

Have questions about this episode? Reach out.

Two Modes, One Codebase: Durable vs. Disposable Code in the Age of Cheap Generation Durable and disposable code aren't two points on a quality scale — they're two different kinds of software with different cost models and different reasons to exist. AI codegen has collapsed the cost of the disposable kind and shifted the ratio between them, which makes the boundary between the two the real engineering skill: naming a piece's half-life up front, designing the seam so promotion or disposal is cheap, and refusing to let throwaway glue silently harden into a load-bearing production dependency. This deep dive starts from the four blog posts that framed the debate and checks them against the 2026 research, security data, and regulatory calendar. ...

Photons In, Controls Out: What Tesla FSD Actually Learns, and Where Its Ethics Actually Live A special-edition deep dive on Tesla Full Self-Driving as a case study in machine-learning systems ethics: how an end-to-end vision-only driving network actually learns, where ethical preferences live in a stack with no system prompt, and why a convergence of practitioners and researchers now argues that the Trolley Problem is the wrong question for the engineering reality. Sober comparative ground from Waymo, Mobileye, and Wayve keeps it from being a Tesla monologue. ...

Should You Be Nice to Claude? Three Spines, One Question, and the Answers That Don't Agree The viral question — does saying "please" and "thank you" to a language model matter? — looks like one question, but it's three. Efficiency, ethics, and tone-of-collaboration each answer it for different reasons, and once you separate them, the 2026 evidence stops contradicting itself. This deep dive walks the spines, finds where they diverge, and lands on what the evidence actually supports. ...

What NIST AI RMF 1.0 Actually Demands — And What It Doesn't Three and a half years after publication, the NIST AI Risk Management Framework has become the closest thing the United States has to a national AI governance baseline — referenced in federal procurement, mapped against the EU AI Act, woven into every hyperscaler's compliance posture. This deep dive reads the document honestly against what it has actually become: surprisingly concrete in what it demands of organizations, conspicuously silent on the systems it was never designed for, contested from civil society on one flank and frontier-safety researchers on the other, and quietly reshaped by an institutional reorganization that has removed the word safety from the name above the door. ...

Episode Description White House unveils AI rule override as grid risks surge The White House unveiled a sweeping blueprint to override state artificial intelligence laws just as United States utility cyber incidents surge roughly seventy percent. To combat escalating physical and digital threats, infrastructure operators like PG and E are rapidly deploying over 630 predictive cameras to mitigate operational risks. In response to this mounting complexity, authorities finalized a hard August 2026 deadline demanding documented operational proof of model transparency to gate audits and procurement. Technology leaders must validate their system inventories and establish compliance guardrails immediately, or they risk losing access to critical enterprise contracts. ...

Episode Description AI Primer This professional primer serves as a foundational guide to artificial intelligence for individuals in business and leadership roles. The text clarifies essential terminology, ranging from the mechanics of Large Language Models and tokens to the practical art of prompt engineering. It distinguishes between generative AI, which creates content, and agentic AI, which autonomously executes multi-step workflows. Furthermore, the source introduces advanced concepts like Retrieval-Augmented Generation for data accuracy and the Model Context Protocol for standardized tool integration. By emphasizing the importance of human oversight and the pursuit of Artificial General Intelligence, the guide provides a framework for making informed strategic decisions. Ultimately, these excerpts equip readers to navigate a rapidly evolving technological landscape with technical literacy and ethical awareness. ...

Episode Description Autonomy Surges: Trust Lags, Infrastructure Unveils Gaps Automated systems are accelerating across all sectors, from AI-driven algorithm discovery to utility infrastructure, creating a sharp tension as security teams face an AI trust paradox in automated response, hesitant to hand over control despite machine-speed attacks. The practical risk of this rapid scaling became clear when the 15-hour Amazon Web Services outage generated over six million reports, triggered by an internal DNS race condition, highlighting acute concentration risk. Regulators and standards bodies pivot aggressively, with the Transportation Security Administration formalizing mandatory pipeline cybersecurity requirements effective May 2025 and the IEC 62443 standard pushing industrial networks toward zero trust microsegmentation. These governance gaps and architectural shifts mean organizations must urgently invest in robust failure containment and user-validated explainable AI to ensure automated speed doesn't compromise critical safety. ...