Beyond the Hype: Is AI taking the fun out of software development?

This title could be clearer and more informative.Try out Clickbait Shieldfor free (5 uses left this month).

A podcast transcript from Scott Logic where CTO Colin Eberhardt discusses a four-week experiment measuring AI's real-world impact on software development with developers Dean Kerr and Amy Laws. The team alternated between AI-assisted and completely AI-free ('cold turkey') work on genuine open-source issues in Mock Service Worker. Key findings: AI roughly doubled velocity (1.9x), but the more revealing insights came from the cold turkey phase — developers discovered how deeply dependent they'd become on AI for summarization, search, and contextual answers, not just code generation. Dean developed an Analysis → Implementation → Reflection agentic loop using red/green TDD to improve AI output quality. Amy found that handcrafting code without AI restored a sense of pride and problem-solving joy she hadn't realized she'd lost. The group concludes that AI hasn't eliminated the fun of software development, but has fundamentally shifted where that satisfaction comes from — and raises concerns about how junior developers will build foundational skills in an AI-first environment.

41m read timeFrom blog.scottlogic.com
Post cover image
Table of contents
Useful links for this episodeSubscribe to the podcastTranscript

Sort: