The entry level for students to get started with “doing Geo” never has been so low. Open-source software has been enabling young professionals on small budgets to gain hands-on experience – creating a vibrant, young community of FOSS “aficionados” in and around university classrooms. Now, we are seeing a second wave of development: genAI makes open-source tools even more accessible, by tutoring students through the process of installation, analytical workflows, troubleshooting, and eventually code line commands. In this talk, I will show some impressive examples of the level at which students can arrive with the help of generative AI and vibe coding with the GAMA modelling software for agent-based models.
In a deeper dive into this topic, I further ask: which competences are left behind with vibe coding? And more fundamentally: if “doing Geo” has become so simple, what is the role of formal Geoinformatics education? When “humans in the lead” turn to “humans in the loop” and beneficiaries of “smart agentic systems”, what will this do to the productive work with FOSS GIS and decision making? How can we humans determine the stages, at which human validation is needed. And how can one validate the outcomes, who wasn’t capable of producing them? By relating to examples from the Spatial Simulation class, I will finish off the talk with some lessons learned in terms of highlights and challenges when integrating (Geo)AI into courses and curricula, and what this implies for future learning, teaching and “doing” Geoinformatics in and beyond Higher Education.