For those not in the know:
https://en.wikipedia.org/wiki/Chinese_room
Very first words:
The Chinese room argument holds that a computer executing a program cannot have a mind, understanding, or consciousness,[a] regardless of how intelligently or human-like the program may make the computer behave.
😆
But Caine will.
nuh uh he can’t
I suddenly understand a bit more about that episode.
You just have to type “what is a” before that and it pops up. I don’t think the llm is choosing not to show us what it is. If it actually did, it wouldn’t be as much of a Chinese room as we think, although I guess it could emulate deception just as well as anything else (as in the blackmail thing from a while back).
The LLM cannot do things that are not in its model.
Which is why its actually bullshit to call any of it AI because a fancy pattern recognition algorithm with a search index behind it doesnt mean Artificial intelligence.
It’s a fancy chatbot.
The Eliza Effect was noted in 1966.
Yup, I either call them LLMs or “AI” Marketing departments have utterly embraced it because everyone, myself included, thinks of science fiction AI and they are so not.
A fun thing I’ve noticed is that for image gen ones, they threw in so much gundam and warhammer 40k art that getting things outside of those when you ask for a “mech” is surprisingly difficult.
Its really surprising that Games Workshop hasnt done anything on that.
As long as you accept that the understanding is embedded in the software, not in the computer.
What a useless conclusion to take from a really insightful argument. Also, there’s nothing forcing AI to resemble the Chinese room, it’s a computer running it, not a look-up machine.
