Back To The Future of Software: How to Survive The AI Apocalypse with Tests, Prompts and Specs at Vaadin Cruise Meetup 2026
Abstract
Translating ideas into code has always been messy. Specs, user stories, and endless discussions often leave developers guessing: “Is this what they meant?” Misunderstandings pile up, and the real work—writing great code—gets bogged down in back-and-forths. What if there was a way to smooth that process? Imagine starting with a prompt, written in plain language, that describes what a feature should do. From there, the AI generates tests, turning the prompt into something concrete—something the whole team can align on. And when those tests look right, the AI writes the code to make them pass. It’s not about skipping steps but cutting through the noise, so developers can focus on solving problems and shipping great features. Could this be how we close the gap between product intent and implementation? Could test-first workflows like this make quality implicit and productivity effortless? In this talk, we’ll explore these possibilities, from the way prompts reshape collaboration to the role of tests in bridging understanding—and how AI fits into it all. Maybe building software doesn’t need to feel like a grind. What if it could feel… easier?
Resources
- Programmers bore the brunt of Microsoft’s layoffs in its home state as AI writes up to 30% of its code
- Tessl - Make agents work in real codebases
- Assessing the Quality and Security of AI-Generated Code: A Quantitative Analysis
- Faster code, greater risks: The security trade-off of AI-driven development
- StackOverflow Developer Survey 2025 – AI Tools (Dis)trust
- Curse of knowledge
- Stochastic
- Test Driven Development
- Behavior Driven Development
- Gherkin Specs
- GitHub Spec Kit
- The Intent Integrity Chain Explaned + Demo Code
- Andrej Karpathy