Yes.
However I don't know why that makes a categorical difference.
We don't have the ability to freeze, copy, store, and restart brain state to the molecular level like we can computers to the bit level. Presumably if we could, the lived experience would be like blips in consciousness (Marvel-like, even).
LLMs today typically do not have persistence, and when they do, it's mediated as text-based storage of future inputs, rather than direct persistence of latent state. But those are architectural decisions, not categorical constraints.
Certain concentrations of chemicals on my tongue trigger certain chemical expressions that propagate signals from nerve cells on my tongue to and across other nerve cells in my brain.
I respond. A smile. Or a wince. I experience physical sensations of taste or wetness or pain localized on parts of my tongue.
I perceive much the same sequence in my dog when I let her lick my ice cream bowl.
What about a the fruit fly that hovers over the too-ripe banana on the counter?
Which of us three get to claim consciousness?
Perhaps. Genuinely.
(Though I personally don't buy the Chinese Room.)
Intelligence, the ability to reason, and the ability to reason about one's self and one's own interests are all demonstrably substrate independent it seems to me. Why *must* consciousness be different?
Winning at chess is a physical act. What is the fixed, here-and-now physical cause?
I mean, the proof is in the pudding.
Two boxes grow in ways we don't understand:
One box is made of a network of cell neurons composed of chemicals.
The other box is made of a network of math neurons composed of IT components.
Both output pudding.
The amount of sameness in the pudding is the only metric we have right now.
I don't know anything about you, and I'm not throwing shade, but a lot of the discourse on this topic feels like it's coming from people who are not deeply embedded in this space.
Compare these two AI code outputs. 19 months apart.
https://t.co/YyTJuH6Zmr
https://t.co/N4ez8l01mE
@asatarbair Depends on whether you're trying to commit a genocide or to kill selected, deeply embedded combatants while minimizing additional casualties.
@WilliamHogeland "LLM are designed to create an illusion that they do think and talk like human beings."
I would say this statement is incorrect and evinces a misunderstanding of how we got here.
"Chemistry ain't conscious. Neurotransmitters are not conscious. I am right about this. If you disagree you are wrong. This is a classic bifurcation between midwits and those that are actually intelligent. Don't end up on the wrong side, you'll look so stupid if you're saying models are conscious."