#753: Amazon Bedrock Mantle and Developing at the Speed of AI
Mon Jan 26 2026
In this episode, Simon speaks with Joe Magerramov (VP & Distinguished Engineer) to explore the transformative impact of AI-assisted coding on software development workflows. Joe shares his team's real-world experience achieving a 10x increase in code throughput using agentic development, but warns that simply bolting AI agents onto existing practices is like "adding a turbocharger to a car with narrow tires and old brakes." We dive deep into the critical infrastructure changes needed to sustain high-velocity development, including the mathematics of bug probability at scale, innovative testing approaches inspired by aviation industry practices, and the evolution of CI/CD pipelines that can handle dozens of commits per hour rather than per day. The conversation reveals why the biggest opportunity isn't just writing more code faster, but using AI to make previously impractical engineering practices economically viable—from comprehensive end-to-end testing with fake dependencies to rapid feedback loops that prevent the entire development pipeline from grinding to a halt when issues arise.
https://blog.joemag.dev/2025/10/the-new-calculus-of-ai-based-coding.html
More
In this episode, Simon speaks with Joe Magerramov (VP & Distinguished Engineer) to explore the transformative impact of AI-assisted coding on software development workflows. Joe shares his team's real-world experience achieving a 10x increase in code throughput using agentic development, but warns that simply bolting AI agents onto existing practices is like "adding a turbocharger to a car with narrow tires and old brakes." We dive deep into the critical infrastructure changes needed to sustain high-velocity development, including the mathematics of bug probability at scale, innovative testing approaches inspired by aviation industry practices, and the evolution of CI/CD pipelines that can handle dozens of commits per hour rather than per day. The conversation reveals why the biggest opportunity isn't just writing more code faster, but using AI to make previously impractical engineering practices economically viable—from comprehensive end-to-end testing with fake dependencies to rapid feedback loops that prevent the entire development pipeline from grinding to a halt when issues arise. https://blog.joemag.dev/2025/10/the-new-calculus-of-ai-based-coding.html