Thoughtworks Technology Podcast · Thoughtworks

How it feels to be a software engineer when AI is changing our relationship with code

·41 min

Nate Shuda, cloud solution architect and co-author of 'The Fundamentals of Software Engineering,' joins to discuss how AI is reshaping software development. He argues that despite AI's ability to generate code quickly, engineering fundamentals become more critical because developers now read and manage more code than ever before. Shuda introduces the concept that 'code is cheap, software is expensive,' emphasizing the gap between generating working code and building robust, scalable systems. He shares examples of how AI enables economically viable custom tools, like his co-author's content moderation system, that would have required months and millions of dollars previously. The discussion covers the persistent importance of soft skills, noting that universities and bootcamps teach coding but not engineering practices like stakeholder communication and system thinking. Shuda compares working with AI to managing an intern, requiring clear direction and oversight while maintaining responsibility for the output. He raises security concerns about AI-generated code, noting that while tests can verify what a system does, they cannot prove what it doesn't do, creating potential attack surfaces. The conversation explores whether AI agents represent a new hype cycle and how developers might transition from writing code to managing AI systems. Shuda predicts that reading code will remain a fundamental skill during the transition period, especially as 'shadow IT' applications built with AI tools eventually require professional maintenance. He emphasizes that communication skills and architectural thinking will become even more valuable as AI handles routine coding tasks. The discussion touches on the evolution from tactical coding skills to strategic engineering capabilities, with Shuda noting that every five years, half the software industry turns over, requiring constant re-education about hard-learned lessons. He concludes by suggesting that software engineers will likely do more people-focused work in the next three to five years as AI takes over more technical tasks.

As heard by us

A calm look at AI as the next shift in software work, sharpened by an intern analogy.

The discussion treats AI as the next shift in software work, not as magic or a harmless extra. Nate Shuda, who calls himself a cloud solution architect, keeps the focus on day to day practice: prompting, guardrails, and the habits engineers build over time.

Read the full review in PlayNext →

Why you'd press play

If AI makes coding feel familiar and strange at the same time.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify