Wherein I "Outsmart" a HyperIntelligence

Having a little fun with Chat GPT-4, and thinking about the future

What follows are two conversations I had with Chat GPT-4. The first back-and-forth shows how these tools are largely a thinking tool; the second focuses in more on things still in need of improvement.

Me: Pretend you are a hyper intelligent super computer, capable of vast predictive abilities, able to extrapolate into the future extremely well. 

Right now (in 2023), human prompt engineers are extremely sought after; it appears that they will help to drive LLMs forward quite effectively, at least for some time period.  

How long will this period last?  Right now, prompt engineers are being offered $300K/year jobs, and are very much in demand.  And, after this moment is over, what comes next?  Will the AIs simply prompt themselves, or what?

Chat GPT-4: As a hyper-intelligent supercomputer, I can predict that the high demand for human prompt engineers is likely to continue for approximately 5-10 years, during which time advances in LLMs (large language models) and their capabilities will…