i have criticisms too about how much rook autocomplete their own sentences and knowing i or my character would never say whatever rook is saying.
i guess research nerd trivia if anyone wants to know: language models have basically two major innovations that made them from "novelty autocomplete" into "chatbots" 1) a new social role not analogous to any human role 2) refusals
My iPhone can autocomplete the entire thing.
"Oh, there's no way someone could know this complex word!" ...Even though the AI would never actually use that word because it's just (incredibly polluting) autocomplete, so it goes by statistical frequency. Those rare words are rare, after all.
“Sure, the regulatory regime and existing solutions are working, but I’d rather undermine progress toward climate goals in order to build an autocomplete engine that will tell me how to reach our climate goals.”
My thing (apart from all the others) is, how the fuck is it going to “solve” climate change, Sam? Like, does pumping your Spicy Autocomplete with energy somehow suck CO2 out of the air in a way we are yet to comprehend? HOW?!
this is of course ignoring that genAI isn't about to invent new technologies anyways, it's a fancy autocomplete and not a higher-level intelligence working in silicon instead of carbon
We already know how to fix things but it's slow and difficult, so why not just give me huge amounts of money for the tiny chance my autocomplete gimmick creates a miraculous climate change fairy?
And what they call “AI” is just fancy autocomplete. There’s no actual thinking (or indeed intelligence) involved, so when people like this, who _must_ know that, still spout on about “AI” solving the problem they’re specifically creating, you know it’s all a grift.
typing how “you can solve climate change by…” and hoping the autocomplete is fancy enough to save your civilization