What’s new in image understanding

A Podkey summary of What’s new in image understanding, from WWDC 2026.A lot of this week’s Apple developer news comes down to one idea: images are getting a lot more usable, a lot more interactive, and a lot less locked into one narrow workflow. Vision can now cut out objects with a tap and refine the result as you go, Foundation Models can work directly from photos in prompts, and the two frameworks are starting to work together in a pretty practical way. There’s also a nice little watchOS angle here, where even something as simple as smarter image cropping can make a tiny screen feel much more helpful.Tap to segment any objectMore than one way to define the objectSegmentation you can refine as you goFoundation Models can take images nowWhy Vision and Foundation Models complement each otherTool calling brings the two togetherBuilt-in OCR and barcode helpCustom tools for specialized image tasksA very watchOS kind of improvementThis podcast was created with Podkey. Make your own at https://podkey.fm