From Frustration to Flow: When AI becomes your partner instead of your opponent

Time: Thursday, 20th Nov, 16:30 – 18:00
Place: Alsion, M304-M307 (go to the main entrance and take lift to the 3rd floor, head East 😉
Sign-up: Sunday, 16th Nov, 23:59

Sandwiches will be served, if you have special dietary needs then you can order a chicken or vegetarian sandwich by sending an email to Krzysztof (krzys@mmmi.sdu.dk)

Topic: From Frustration to Flow: When AI becomes your partner instead of your opponent

Description
Are you tired of the AI hype too? Of chatbots that guess wrong, and the talk about robots taking our jobs? Have you tried GitHub Copilot and thought: “It’s just glorified autocomplete – and it makes mistakes!” And in a busy workday filled with deadlines and bugs to fix, who has time to “play around” with a tool that feels slower than just coding it yourself?
But what if the problem isn’t the tool – but how we use it? What if the skills we already have as experienced developers are precisely the key to turning frustration into progress?
Thomas Kure Thorngreen, Senior Software Engineer at Danfoss Drives, has spent the past year on exactly that journey. From the first frustrating attempts with ChatGPT in Python side projects (August 2024) to today, where GitHub Copilot is an integrated part of his daily work – from protocol debugging on embedded systems to building complete project management tools. With a background stretching from C development in the 90s to modern C++ in large embedded systems, he has experienced firsthand how AI changes what’s possible.
The evening isn’t about selling a product or painting an unrealistic picture. It’s about sharing a discovery: That AI tools, when you learn to guide them correctly, don’t just suggest code – they extend your reach as a developer. Suddenly lab engineers with domain knowledge can build their own tools. Complex problems can be approached in new ways. And the curiosity that once drew us into this field can be reignited.
But just like learning to drive with a manual transmission: It’s frustrating at first. You don’t drive faster than on a bicycle. But when it clicks – and it does – possibilities open up that you didn’t have before.
You’ll see concrete examples of features built with AI, live demonstrations of GitHub Copilot’s capabilities, and deep dives into advanced topics like MCP servers and agentic development. We’ll talk about how to get started without it consuming project time, how to avoid the worst pitfalls, and what this actually means for our future as software developers.
Leave with an honest, nuanced picture of what AI can and cannot do – and perhaps a spark of enthusiasm to try it again, this time with fresh eyes.

Speaker
Thomas Kure Thorngreen, Senior Software Engineer, Danfoss Drives