Whatever you think about generative AI—and you know what I think—you have to concede that it isn’t very good at making maps out of the box: there are so many examples of it going horribly wrong. Interestingly, Google agrees. “General-purpose LLMs and standard coding tools struggle with real-world spatial problems because planetary datasets are massive, with specialized map and satellite formats that require careful multi-step planning to combine accurately.” Google touts its suite of Google Earth AI models, which, it just so happens, do all those things; they’re expanding agent access to them. This is pro-level stuff, mind—and from what I gather comparing machine-learning tools (“AI”) to consumer-grade large-language models (also “AI”) isn’t exactly apples-to-apples. [David G. Smith]
A German court has ruled in favour of a man suing Google for invasion of privacy over aerial images of his property in Google Maps and Earth without his consent. [PetaPixel]
Trump suggested renaming the Atlantic Ocean to Google CEO Sundar Pichai, but Pichai balked: “Please don’t ask us to do that.” [Engadget]