Joe Magerramov's blog: The New Calculus of AI-based Coding

AI-assisted coding can increase development velocity by 10x, but this requires fundamental changes to engineering practices. Higher commit rates mean bugs occur more frequently, demanding better testing strategies like comprehensive fake dependencies and local end-to-end tests. CICD pipelines must become significantly faster to handle dozens of commits per hour, and teams need tighter communication loops to coordinate rapid decision-making. The key insight is that AI agents can both accelerate code production and make previously expensive quality practices economically viable.