Has anyone used the Apple Intelligence Feature image playground ?
Its a good idea but has a long way to go. Its actually really pretty bad most of the time. I suppose it is still a very early beta…so hopefully will improve.
I even asked it "Display the text “THIS IS TEXT” and it came up with “THIS THILS 1 ETS”
I haven’t been able to get it to do anything remotely useful to me.
This isn’t unique to Apple Intelligence. I run into this same issue with other AIs. For some reason, AIs struggle with mixing images with text. I even tried telling the AI, “Do not include text in the image” only to have the AI change the text, or even increase it. 
Thanks, I’ll check it out.
Yep, and wrote a blog post about it.
I think it’s going to be far more useful than most of my fellow geeks who seem to be coming down hard on it. In the latest App Stories Federico said it’s so bad he’s not even going to cover it on MacStories.
I think it highlights the differences of perception between normal users who do not listen to tech podcasts or spend time in tech forums and the geeks who tend to be, IMHO, overly critical. I made a handful of images (shared in blog post above) and shared with my extended family and the overwhelming response was VERY positive.
This isn’t meant to be professional illustration, it’s meant to be upgraded clip art. That’s it, that’s the use case. And yes, no doubt, it will get better.
Thanks for your post and the blog article. I think, for the use cases you describe, the images are fine—some are very good. I’d add a use case, one I’d use often: illustrations for blog articles. I’ve already used ChatGPT for this purpose.
I don’t have Image Playground. I assume this is still in beta?
Yes, you need to be in the beta for 18.2 to get it. We should have it available to everyone this month. 9to5Mac predicts next week.
Good article @Denny, and key point.
I’m quite a techie (taught computing and had many years running large networks and IT support in education) but I often find myself puzzled by much of the tech commentary from podcasts and blogs and especially by the expectations and norms in tech forums. There’s often a disconnect between my experience of what software and hardware is good, what experiences work and especially what design is pleasing and what is often a common set of assumptions and even viewpoint from a whole range of commentators.
That might just be me being a grumpy old man, but I don’t think so. To give a few examples relevant here: I don’t want MacOS on my iPad Pro, thank you. I don’t need to do absolutely everything I ever want to do via a single keyboard shortcut. I particularly don’t want to be sneered at because I like tactile computing (trackpads ftw). I don’t need to clutter my machine with a dozen esoteric menu bar utilities, and I think Apple is one of the most successful companies of all time, by any reasonable measure, and generally has got it right with the design choices it has made (some better than others)
I was talking recently to a developer whose company has started to interview (by Zoom) as wide a variety as possible of users, seeking out newbies, youngsters, oldsters etc. in a planned way, trying to get feedback in proportion to their user base. They’ve been shocked at how unrepresentative their support forums, reddit, social media etc. feedback actually were and realised they were being pushed into design choices by a vocal minority, that were making their app worse for many users and their use-cases.
MPU is much better than many other podcasts, especially when it reaches out to power users who are outside the inward-looking “podcast and creative” bubble , but I wish the whole techosphere would question its assumptions more often.
I couldn’t agree more! You described my perspective perfectly. That’s probably scary, but I thought I would pass it along anyway. 
Image Playgrounds arrived on my computer today, possibly because I’m in the EU? Yes, the output is just screaming “AI generated”, so this limits the reuse of its generated images by a lot.
Anyway - I had fun playing with it for a moment, and I’m sure kids will love this. Of course, I made “memojis” of the whole family, and turns out it took quite a lot of different reference photos before it created something with a decent likeness. As humans, we’re really good at seeing the difference between two people, which is probably because Playgrounds is struggling.
I tried adding a few keywords, like “not smiling” and “mouth closed” which was totally ignored. Don’t see a big practical use for myself, but will see where it goes.
A bit sad we can’t open up ALL images tagged with a person from the Photos library, as I am missing a few fun ones I’d like to try.
CORRECTION: just found the Photo Picker icon in the lower right 