There are no downsides for the customer if the software is much more capable than what the user needs. So there's not much of a business case for offering software with less features.
Because fully fledged software is dirt cheap already. Even Photoshop is only $20 per month. That is nothing to a professional who needs the tool. People who care about that kind of money will never pay any money at all for software, they will look for something for free.
1. People buy and use fully fledged applications like Photoshop or get Affinity.
2. People ask AI to do their image editing / creation tasks.
The number of people who will have AI make programs for image editing is so small as to not even count. And they will only do it for tinkering reasons to make a vector image of Tux the penguin and then never use it again.
So in your mind, there is no scenario where people ask AI to create a more focused (even single-purpose) image editing app, and either make it free or charge a small amount?
Because I'm seeing that happening more and more. Just as one random example, I've lost count of the number of vibe coded captioning apps I've seen for FCP, because it's a known pain point in the software.
Sure, but there isn't much of a business in it I would say. People who are too cheap to pay for Photoshop aren't going to pay for a small tool either. And Affinity is already free and very powerful.