{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"Travel Tech Podcast","title":"Why AI Is Slowing Down Experts Before It Speeds Up Work (Brooker, Painter, Deakin, McKenzie)","html":"","width":"100%","height":180,"duration":1349,"description":"AI adoption inside teams is not following the narrative most people expect. In some cases, the most experienced engineers—the ones expected to benefit the most—are actually getting slower.That friction reveals something deeper. The challenge is not just about tools or capability. It’s about trust, accountability, and how work itself is structured. In high-stakes environments, where someone must sign off and take responsibility, AI doesn’t simply slot in—it fundamentally reshapes how teams operate.This conversation with Alex, Ian, Oli, and Adrian explores what happens when AI moves from experimentation into real production environments, and why the bottlenecks are as much human and organizational as they are technical.What You’ll LearnAI can reduce productivity before improving it: Senior engineers may initially slow down due to context switching and deeply ingrained workflows.Trust is not abstract, it is operational: In regulated or high-risk systems, adoption depends on proof, repeatability, and accountability—not just perceived capability.Accountability remains human even in AI-driven systems: Someone must still sign off on outputs, especially in safety-critical environments.Team roles are shifting from building to assuring systems: The future focus moves from writing code to validating system behavior and outcomes.Junior career paths are being disrupted: Traditional entry-level tasks are increasingly automated, forcing a rethink of how engineers are trained.AI adoption varies dramatically by domain: Safety-critical industries like aviation will adopt far more slowly than consumer or enterprise software.Larger code generation introduces new risks: AI can produce more code faster, but also increases bug rates and cognitive load for reviewers.The real constraint is system-level understanding: Teams must still comprehend architecture and system behavior, even if AI generates the code.Productivity gains follow a J-curve: Teams must go slower first to learn how to...","thumbnail_url":"https://img.transistorcdn.com/LxpvuNpWwfSGFL1KA1WhoZf9L55ykAqb5rgjXNFqi3c/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9mY2Yz/ZjA5OGE1ZmEyMTk4/ODJkYmU1YjhlYjRk/YTMzNC5wbmc.webp","thumbnail_width":300,"thumbnail_height":300}