In the past year, the world has been introduced to increasingly innovative smart assistants, powered by generative AI. More and more complex questions are being answered, and with the introduction of agents, tasks can even be performed automatically and partially autonomously.
But what does this mean for the Mendix Community? Mendix Assist simplifies Mendix application development for both new and experienced developers. But there is also a lot going on at the edge of the platform. There is a global discussion about the potential for generative AI to replace the high-code developer partially. How can we use this high-code assistance during our widget development workflow? We explored it for you!
AI Showdown - The Comparison
This Meetup is the result of weeks of research in which Erick Souza, Tobias Kuhn, Lukas Tost and myself used different AI code assistants in developing a widget for a fictional use case. Which assistant provides valuable answers? How up-to-date is the assistants' knowledge? Which code examples are qualitative? And does it actually make the development of a widget faster, better, or of higher quality? Last but not least, what is our role in interacting with the assistants, and what did we learn about asking the right questions?
The AI assistants we cover are the publicly available versions of ChatGPT 4, Google Gemini, Microsoft Copilot, and Amazon CodeWhisperer.