BayCHI | Jude Yew | Fast AI vs Good AI: Reflections of AI Adoption

Light dinner served at 6:30pm PT. Feel free to bring your own if you prefer. We strive to support both in-person and remote attendance, so we'll also fire up the Zoom at that point to support virtual socializing.

Abstract

Fast AI vs Good AI: Reflections from the trenches of AI adoption on a Software Development team
Jude Yew

AI-powered software development is generating a lot of claims– about productivity gains, about how engineers work, about what's changing and what isn't. But how much of it holds up from the perspective of someone in the trenches of doing this work?
This talk draws on my experience as a UX Researcher embedded in a software development team that is in the midst of employing agentic modes of working across our organizagtion. I'll take some of the most talked-about ideas in AI-enabled development– cognitive surrender, productivity measurement, the pace of model change— and run them against what I've actually observed and experienced. Some of the discourse holds up. Some of it doesn't. And some of the most important things happening aren't being talked about much at all.
The goal isn't to debunk the hype or add to it. It's to offer a more grounded account of what this transition actually looks like from the trenches.
Bio:
Jude is a user experience researcher working on developing efficient and delightful Developer Experiences. His work focuses on using mixed-method approaches towards studying developer workflows and tools in order to identify pain points and challenges. Jude manages a team of UX Researchers and Software Engineers that identifies and addresses software developer experience across a variety of surfaces such as, APIs, SDKs and Documentation.
He was previously a faculty member at the National University of Singapore studying the interaction of humans in online social spaces with a goal to harness collective effort and intelligence for greater social good.