I think AIs will quickly progress to the point where the boss can just tell them to do X, instead of telling the prompter to tell the AI to do X.
I think AIs will quickly progress to the point where the boss can just tell them to do X, instead of telling the prompter to tell the AI to do X.
No magic, all dumb for loops and string formatting and a bunch of calls to LLM. Just a single prompt from the manager. Writers can now go do a roofing job to put food on the table while the retirement age always grows proving them how AI is finally freeing humanity from having to work.
There's no boss in the world who wants to be left alone with the machine. They are bosses, they are not technical people; they don't like machines... machines cannot go to jail or be accountable for mistakes.
If there is a boss2 above boss1, then boss1 is accountable to boss2 for subordinates and there is almost no difference between blaming and firing an employee or blaming and firing an LLM or LLM developer. Also important to remember that an LLM doesn't snitch if you ask it to do shady stuff.
> There's no boss in the world who wants to be left alone with the machine
Most bosses want to be richer and want obedient underlings.
Remember it won't be a scary mainframe. It's basically a benign looking contact in your chat. Same thing they use to message their misstreases. They will be fine.