Wednesday
Room 2
14:40 - 15:40
(UTC+02)
Talk (60 min)
Where thinking stops
I decided to rebuild a retrospective tool from scratch. This was an internal tool we'd been using for years. AI would write the code, and I'd make the architectural decisions. Or at least that's what I thought.
Although things worked, something felt off. Implementations were technically correct but oddly inconsistent.
I realized that AI didn't make any mistakes, but that I hadn't made the decisions. AI couldn't cross any boundaries, because I never drew them. Speed made it impossible to ignore implicit thinking.
This is not a talk about prompting, frameworks, tooling or models. It's about recognizing where decisions were missing all along. AI makes it uncomfortably fast to see where boundaries are missing.
Yngve Bakken Nilsen
Yngve Bakken-Nilsen is a senior software developer and consultant with over 20 years of experience, currently working with Norsk Rikstoto. He works mainly with Angular and .NET in large, long-lived systems, with a focus on architecture, developer experience, and keeping complexity under control.
Lately, he’s been building a SaaS product where most of the code is written with AI. That experience has shifted his attention away from writing code, and toward something else: figuring out where things start to go wrong when it becomes too easy to build.