My problem with all this is that the quality of all the generated text and graphical content is utter garbage. Quantity != quality.
The graphics I've produced are absolutely phenomenal IMHO. I find that that textual content depends highly on the prompt, and often does take quite a few iterations to get right.
Often with text, the GPT is more of an assistant/advisor/proofreader for me, rather than a stand-alone creator of quality content.
Sometimes it works really well for emails. Like for producing responses to formal communications. It can save cut the time to respond from 10 minutes to 30 seconds.
We're already witnessing this with the creation of textual and graphical content through ChatGPT. It's now possible to generate various types of text content and a wide range of graphics at the cost of 10 cents for the dozen ChatGPT API calls. And the work is completed in a few minutes, as opposed to several hours. This represents a several orders of magnitude increase in per capita productivity for these specific tasks. As AI technology advances, the scope of applications benefiting from such productivity boosts is expected to widen, which means human civilization will experience a revolutionary increase in productivity, and with it, resource abundance.