Via Stephan Heuel1 I came across this discussion in the FME2 ecosystem: The Elephant in the Room? The elephant, of course, is AI and AI agents and their rising role in conceiving, implementing, and executing geospatial data pipelines. A short excerpt from the thread:
My default tool for random tasks or jobs was FME, now its Claude Code/Cowork. Using FME just feels way too slow now by comparison. Microsoft Copilot is even in Excel now, I never thought I’d go back to using Excel!?!? (…)
Everytime I open up FME now I feel so unproductive because of the time it takes to get something done. It feels painfully slow compared to having the Agent just “Do it”.
Safe Software3 picked the discussion in the Safe Software Community Hub up in a post of their own: Are AI Coding Agents Replacing Tools Like FME? Let’s Talk About It.
Betteridge’s law of headlines applies: “Any headline that ends in a question mark can be answered by the word No.” Still: I agree with Stephan’s kudos towards Safe Software for their openness to discuss the topic.
Both the original discussion thread and the Safe Software post are worth reading. The opinions expressed do not only matter for FME, I would say, but more broadly for similarly positioned tools and solutions.
Footnotes
Side-track: 15 years ago I worked with Stephan, but my take on FME was very different from what he expresses in his post: While I appreciated its robustness compared to other low-code tools for geospatial workflows (especially Model Builder at the time), I did not like FME overall and preferred to write code instead. Granted: That (a small) part of my role included maintaining, repairing, and running fickle FME workspaces that had been built by someone else could have something to do with that 🙂.↩︎
Safe Software’s Feature Manipulation Engine.↩︎
Maker of FME.↩︎