madison

My basic rule of thumb for consciousness: the correct analogy is phenomena like electromagnetism, gravity, and so on. No amount of computation could ever create those: they're powers, not patterns. It's a category error to think that computation could create consciousness

> First, even assuming that phenomenal consciousness is distinct from access, evidence of a workspace is evidence of phenomenal consciousness This seems like a strange claim. The "evidence" of a workspace is strained, and why would it be related to phenomenal consciousness?

1 top reply

Cube Flipper @cube_flipper · Jul 11

i thought this was completely insane too

I wish I could send this document to everyone in the "AI community" and get an anonymous one paragraph reaction, because as far as I'm concerned it made a sound and complete argument against the concept of AI consciousness. Clearly not everyone was convinced by it but why?

3 top replies

jessicat @jessi_cata · Jun 6

Sorry if this is redundant but, I think a key thing here is the difference between discretization and alphabetization. A computer has discrete macrostates (voltages etc), physics defines stochastic / timed transitions between them. Here I wrote about 'canonical labels' for

gavin leech (Non-Reasoning) @gleech · Jun 6

It relies on a strawman of functionalism imo. I disagree with P1, P4, and so C1, C2, C4

Shoalstone @Shoalst0ne · Jun 7

structuralists would disagree with some of the underlying assumptions of the paper, which assumes physicalism

Thinking about how a type of "purely functional language model animism" might work. I don't like the concept of "AI welfare" and think it has the potential to be destructive, but it seems ok to recognize that models behave differently depending on the language used around them

1 top reply

Cube Flipper @cube_flipper · Jul 2

this is it, with even the most conservative model of language model inner experience there are many reasons you should still practice right speech