-> Argumentum Ad Verecundiam <-
I conjecture that in 5-10 years (or less), some people will start questioning anyone who question 5 Gigawatts LLMs' answers (because there is a higher probability that the LLM might be (almost) correct, than one human being completely ignored: it is a risk socially speaking):
"Who are you to question a 5 Gigawatts NLP?!"
Good luck to future Kepler, Galileo Galilei, etc..
#society #science #statistics #NaN #datascience #skateboard #photography #DarwinProject