To be skeptical of AI is to commit yourself to near-constant demands to prove yourself, and endless nags of “but what about?” with each one — no matter how small — presented as a fact that defeats any points you may have.

Vegans will be right at home.

Roose, along with his co-host Casey Newton, would go on to describe this example at length on a podcast that week, describing an entire narrative where “the human actually gets suspicious” and “GPT 4 reasons out loud that it should not reveal that [it is] a robot,” at which point “the TaskRabbit solves the CAPTCHA.” During this conversation, Newton gasps and says “oh my god” twice, and when he asks Roose “how does the model understand that in order to succeed at this task, it has to deceive the human?” Roose responds “we don’t know, that is the unsatisfying answer,” and Newton laughs and states “we need to pull the plug. I mean, again, what?”

Fuck I hate this shit. I can hear their voices and I haven’t even watched it. The fucking techbro speech, “we don’t know, that is the unsatisfying answer” SHUT THE FUCK UP. AND HES FUCKING LYING.

Just two paragraphs earlier for context:

It is transparently, blatantly obvious that GPT-4 did not “hire” a Taskrabbit or, indeed, make any of these actions — it was prompted to, and they do not show the prompts they used, likely because they had to use so many of them.

I take the easier, simpler, faster, direct route, with phrases like “Fuck off” and “Fuck no” and “Fuck that shit”. I owe nobody an explanation on why I won’t use these fake-AI environmental disasters. And the techbro assholes pushing it are obviously not used to people using that kind of terse sweary language, it throws them off-balance.

hey chatgpt make harry potter and superman kiss

implying i would ever waste energy talking to someone stupid enough to like AI