Let's use Netflix as an example: my assumption is right now the majority non-personnel costs for netflix are licensing, production, and network-related costs. In the future, I could see them using the same infrastructure to generate bespoke series and movies for user's prompts. In the ToS it will specify the product of those prompts belongs to Netflix, and the users will trade and show off their prompts with friends on Netflix's social interface. This eliminates a lot of the cost in focus grouping, casting, actors, editors, production, licensing, royalties if any.
In the new AI netflix, new tastemakers emerge, people who are known for having the best prompts, and with 0 cost of licensing or production Netflix has a never ending stream of -exactly- what the users want. Knowledge of what the user wants to see is a big part of what Netflix is concerned about in this current day.
So, with this example in mind, I go between two extremes: 1) this will be a job for talented few, like being an actor now, making prompts will be a niche or a cause-célebre, 2) this will be something everyone does, like taking beautiful photos/videos on your consumer camera.
What do you think, will this be a job some day or will this just be something that everyone 'does' to get a better output from their AI 'helpers'?
At first, only a few talented people can do this. Then as the tools they develop get better, the job will be more accessible.