Furthermore, someone who's genuinely good at prompt engineering, to the point that they can make the AI do magic tricks, has a skillset they can market to other employers. Employers have no use for an ape who asks ChatGPT how to do their job, but they can absolutely use an employee who uses AI to completely eliminate those jerbs.

This is also why AI investors are unable to describe any AI success stories. It's not that they don't know of any, it's that the public doesn't want to hear how utterly replaceable most workers really are

(http://www.autoadmit.com/thread.php?thread_id=5825753&forum_id=2\u0026mark_id=5310911#49613661)

Also correct that other LLMs can engineer prompts for gpt better than gpt can. OpenAI publishes detailed prompt guidance with every new release that any LLM can learn. And they have a prompt optimizer tool for premium account users that nobody knows about.

(http://www.autoadmit.com/thread.php?thread_id=5825753&forum_id=2\u0026mark_id=5310911#49613706)

https://x.com/i/status/2015492341570703804

(http://www.autoadmit.com/thread.php?thread_id=5825753&forum_id=2\u0026mark_id=5310911#49618520)

But the photog could alternatively have put his phone down and looked behind him to the number on the outside of the nearest building, and known exactly where he was (and in around 1/100th of the time it took him to upload the video to GayGPT for confirmation of his location).

(http://www.autoadmit.com/thread.php?thread_id=5825753&forum_id=2\u0026mark_id=5310911#49619305)