Spring Office Hours S4E26 - Groovy, Google and Gemini with Guillaume Laforge
Sep 9, 2025
Guillaume Laforge, co-founder of the Apache Groovy programming language and Developer Advocate at Google Cloud, shares his journey from Groovy's inception to the forefront of serverless tech and generative AI. He discusses Groovy's seamless Java integration and its practical applications in tooling like Gradle and Jenkins. Guillaume highlights how Java developers are embracing AI and the essential criteria for evaluating large language models. With insights on the Agent Development Kit and the latest Java features, he shows how coding is evolving in the AI landscape.
AI Snips
Chapters
Books
Transcript
Episode notes
Prefer Embedded AI Over Ubiquitous Chatbots
- Avoid adding chatbots everywhere; prefer targeted AI features like summaries and classification.
- Use LLMs to enhance search and proactive UX rather than defaulting to chat-first designs.
Run Evaluations After Model Upgrades
- Treat LLM upgrades like code changes by running your evaluation tests after each upgrade.
- If tests fail, adjust prompts or configurations before rolling new models to production.
Speed Is A UX Factor For Chat
- Latency and first-token speed matter for chat experiences as much as raw accuracy.
- Fast models like Gemini Flash improve perceived responsiveness and user experience at scale.
