Wednesday
Room 2
13:20 - 14:20
(UTC+02)
Talk (60 min)
The playbook we built for 2,000 developers (and the parts any team can steal)
Every team has one: The person who cares about how the team works with AI. The one who sets up the tools, shares the tricks, fixes the mess when an agent pushes to main. Maybe that's you.
This talk is our playbook, or the parts any team can steal. How we think about AI safety after real incidents, from sandboxing and git hooks to agent permissions and supply chain protection. How we evaluate new tools honestly (including when the AI lied in our own analysis). How we built a community of practice from scratch with workshops, coaching, peer-led knowledge sharing. And the one principle that kept us honest: “Does this make the community more capable, or does it make us more central?”
You don't need 2,000 developers to use this. You need one person who cares. Maybe that's you.
Håkon Eriksson
Håkon is a developer and architect focusing on .Net and Python. He spends a lot of time tinkering with the technology, and is drawn to the practical side of AI adoption: what works, what breaks, and how to build a community that grows more capable over time
Øyvind Viken
Øyvind is a .NET oldie, master of artificial intelligence, and likes to be grounded.