@tess I.remember Cory Doctorow written a pretty good article on it so let me parapharse a quote I really like. Advancing LLMs will not create machines which you are capable of speaking to for the same reason that breeding horses does not result in locomotives. LLMs are not able to process concepts - they are bs-ing the most likely word to come. By far, AI has been damaging the public by draining us of the energy by making us consider outrageous far-away at best scenarios.
As Photoshop came into existence, we did not suddenly start distrusting photographs and as AE came in we keep watching videos. AI might make faking material more accesible, and it might even become plausible one day, but verification systems nowadays are in place for a reason.
We should be focusing our efforts on improving labour conditions so that dumb people who believe the hype don't put us out of job, instead of considering the scenario that (gen)AI's might create material comparable to that of human origin.