Revisit your AI disappointments.

Folks, a few days ago I wrote a post about how “pocket agents” change everything, illustrated with a photo of Kattis's watch, where Marvin popped up and pointed something out to her. Completely automagically, based on her real life, and in a genuinely smart and insightful way. Exactly what she needed, right then.

On Saturday I used that image and that story in a Swedish session, and now I need to do the same for an English keynote I'm giving later today. Which meant I needed an English version of the image. I've attached the result here. If you indulge in extreme pixel-peeping, you could point out that the text is a little too sharp (and I could probably have prompted that away, but for this particular use it's an advantage as it is). Otherwise, it's an extremely faithful representation of the original image.

My thinking right now is that we have (among other things) two problems with how we work with AI today. One is that we're stuck with an outdated idea of what AI can and can't do. This would have looked like cat shit a year ago. The other is that we're dazzled by the latest cool thing, which right now might be extremely advanced videos and 3D renderings…

But there's also that practical side, which is easy to miss. Where something that didn't work yesterday can be a substantial boost to efficiency or quality today. And where things that otherwise wouldn't have been done get done.

So. When did you last revisit your digital parking lot 🅿️ You know, the place where you've collected your past AI disappointments? Go there and have a look now. It's almost certain that AI's capabilities have caught up with your needs. Or perhaps it's YOUR ability to actually use the services that has caught up with your needs...