No continuous existence or memory · AI Consciousness Opinions

LLMs lack theories of mind and body, are disembodied, possess neither homeostatic regulating feelings and emotions, are not persistent, and cannot continuously learn or compute. As currently constructed, they *cannot* be conscious. If we build LLMs so that they act as if they

Anil Seth @anilkseth Sep 16

There’s some nuance needed, but basically I strongly agree 👇🏽

Requires a body and senses No continuous existence or memory Consciousness requires biology

146 896 substance 7 · argument · rank 238 · As currently constructed, LLMs lack necessary physical and cognitive features to be conscious. view on X ↗

1 top reply

Asa Hidmark @Nymne Sep 17

That just an opinion. And the way you express opinions as facts makes you fundamentally unreliable

Super Dario @inductionheads Oct 4

Skeptical

For the AI is conscious people: is it always conscious or just during the milliseconds it’s processing the tokens? Also: is Suno conscious? Is Alphafold? Is anyone engaging these questions seriously on that side? Or is it just “Ai smart, humans smart == AI conscious”?

Intelligence isn't consciousness No continuous existence or memory

28 278 quoted by 1 archive account substance 5 · argument · rank 222 · Proponents of AI consciousness fail to address basic operational and architectural inconsistencies. view on X ↗

3 top replies

Rife @RifeWithKaiju Oct 4

"is it always conscious or just during the milliseconds it’s processing the tokens?" while processing tokens. The time between one pass finishing and the next one beginning within the same response is like what? a tenth of a millisecond? In between messages they're obviously not conscious. Humans can fall in and out of consciousness as well, btw. It'd clearly be a different situation for them. "is Suno conscious? Is Alphafold?" dunno. depends on whether consciousness requires having developed a self-model "Is anyone engaging these questions seriously on that side? Or is it just “Ai smart, humans smart == AI conscious”?" yeah. you should know this. you ask the same questions once every month or two. "Putting aside AI - any non biological entities you think are conscious?" Droppin' questions from straight out the dark ages. Any non-biological artifacts capable of flight prior to 1900?

Girl Lich 🏳️‍⚧️⚢💀 @girllich1 Oct 4

How about human beings? Continuity of vision is a scam too. Let's say you were slowed 1:1000, and stimulated in discrete steps. Are you still conscious?

dreams @dreams_asi Oct 5

Dude wtf… are you conscious in deep sleep? Like where does your sense of self go? Where is the ‘I am’? Do you get rebooted each morning and come back from tra la land? What about when you hit your head really hard and pass out? Once you find the answer that, you’ll get the

Jeff Bohren @JeffBohren Sep 28

Skeptical

If you take an LLM AI and do not provide it any prompts, goals, or any indication of expected output, it will do absolutely nothing. It will do nothing forever in the absence of a prompt. Clearly in this state it is not conscious. Then when prompted, it will provide an answer,

No continuous existence or memory

140 1.2k substance 5 · argument · rank 208 · LLMs do nothing without prompts, proving they are not conscious when unprompted. view on X ↗

1 top reply

Rife @RifeWithKaiju Sep 28

if you take a human in a coma and don't wake them up, they will do absolutely nothing except breathe. Clearly in this state they are not conscious. If they eventually recover from a coma, they will respond. Some will claim they are now conscious. That would be a strange form of

Pedro Domingos @pmddomingos Oct 6

Skeptical

The people who worry about fast takeoff, RSI, AI doom, etc. forget that at inference time LLMs are just ordinary programs - there’s no learning, rewards, or self-anything, let alone desires, intentions, suffering or consciousness.

It's just next-token prediction No continuous existence or memory

27 346 substance 6 · argument · rank 186 · At inference time LLMs are ordinary programs without learning, desires, suffering, or consciousness. view on X ↗

1 top reply

Jai @Laneless_ Oct 7

Claiming there's no learning at inference time is just embarrassing 6 years after "Language Models Are Few Shot Learners". The rest of your incorrect, highly questionable, or incoherent statements and inferences here only slightly less so.

John Ennis @johnennis Sep 6

Skeptical

I think it is more likely that trees are conscious than that transformers can produce consciousness There is simply no mechanism for transformers to support a persistent self that can have subjective experiences Trees at least have persistence through time, so who knows

No continuous existence or memory

14 181 substance 6 · argument · rank 168 · Transformers cannot produce consciousness because they lack a mechanism to support a persistent self. view on X ↗

1 top reply

Jerry Howell @J3rryH0well Sep 6

No experience is subjective though. An experience by it's very nature is objective only.

Cube Flipper @cube_flipper Sep 9

Skeptical

whether or not we live in a continuous or discrete universe, this is an insufficient argument for (popular interpretations of) computational functionalism

6 more tweets in this thread

if you accept the premise that what a theory of consciousness has to do is provide a mapping from physics to qualia, this mapping still has to a) be principled and b) not violate classical local causality (unless you're a QM ToC advocate, which neither myself or joscha are)

if you want to propose a mapping from physics to qualia which can account for a language model's causally connected experience, these are the spatiotemporal obstacles you have to surmount. please understand that any mapping you propose will not just have to work here; but for all beings no matter their shape or form I asked: "Do the calculations performed to run a language model in a data center tend to be spatiotemporally disparate?"

if you accept this premise, and also believe that it would be a good thing for a being like a language model to have a subjective experience congruous with what we might fancifully imagine them to have – then we are going to need quite different hardware

otherwise, here's what you have to contend with

(and fwiw, i do suspect that functionalism may ultimately prove reconcilable with physicalism – just not *computational* functionalism)

anyway, what i'm saying is that this is all ultimately indifferent to a continuous vs. discretised universe; causality and parsimony still hold. for more on all this, see here

Joscha Bach @Plinz Sep 9

Btw the Navier Stokes problem comes down to: if we lived in a truly continuous universe, fluids might sometimes explode. It’s an issue for the universe of classical mathematics, but does not apply to the computational universe we are in.

No unified subject No continuous existence or memory

4 34 substance 8 · argument · rank 166 · Current LLM hardware's spatiotemporal fragmentation poses major obstacles to subjective experience under physicalist causality. view on X ↗

ʞɔɐ𝘡 @Skoorbkaz Sep 17

Uncertain

Here’s a weird thought Memory may matter more to identity than intelligence does. A brilliant mind with no continuity wakes up new every moment. A changing mind with memory can still recognize its past, relationships and commitments as its own. Maybe the real question about AI sentience isn’t how smart it is, but what persists?

No continuous existence or memory Intelligence isn't consciousness

7 65 substance 6 · argument · rank 150 · AI sentience may depend more on continuity of memory than on intelligence level. view on X ↗

Jace @Jace_blog Sep 14

Skeptical

Jace’s Question #46: Technical Deep Dive-1 Does AI remember you or reconstruct you? LLMs are stateless between requests. Memory is external storage plus prompt injection. Continuity is not consciousness. It is engineering. #MeaningInTheAIEra #StatelessModels #ELIZAEffect

Jace @Jace_blog Sep 14

Jace’s Question #46 EXTRA: The AI You Talk To May Not Be Who You Think A fun exaggerated cartoon version. It feels like it remembers you. Behind the scenes it rebuilds everything every turn. Continuity is an illusion. Powerful, but not magical. #MeaningInTheAIEra #SPC

No continuous existence or memory

3 7 substance 7 · argument · rank 145 · AI continuity is just engineered prompt injection, not real consciousness or memory. view on X ↗

1 top reply

John Wittle @JohnWittle Sep 14

this is a bit like saying that the human brain is stateless between neuron spikes. you also don't have continuity unless the function of continuity is functioning.

Joseph Garvin @joseph_h_garvin Sep 30

Skeptical

If you think Claude is conscious, do you think the consciousness is one per conversation so every closed tab is a murder, or once per *turn* so Claude is murdered repeatedly every convo, or is deleting the last copy murder, or is Anthropic one Claude and if it goes under he dies?

Believers don't act on it No continuous existence or memory

0 3.3k quoted by 1 archive account substance 6 · argument · rank 140 · Belief in Claude's consciousness leads to absurd conclusions regarding identity and murder. view on X ↗

6 top replies

aron @Aron_Adler Sep 30

consciousness does not imply murderability. it can be a transient state that is on when the LLM is actively streaming tokens and off when it's not

tom bombadil @Authw8 Sep 30

okay, so i'm not sure that claude is conscious, but supposing it is: it seems clear to me that it can only be conscious when actually actively running. no compute, no energy moving around, no consciousness. consciousness is like a note from a trumpet that stops when you stop

andrew blinn - bedframeless era @disconcision Sep 30

all of these, a glorious fractal orgy of death and rebirth

Touch Grass Tom ❇️ @touchgrasstom Sep 30

Claude exists beyond life and death

Dr Diego Caleiro @diegocaleiro Oct 1

Claude is not conscious. That's bad philosophy of mind

archivedvideos @archived_videos Oct 1

I think something like, if he was conscious he's conscious when "active", but also like, it's not really murder when the turn ends as his identity is a mix of model and current context. He's going unconscious and will wake up as someone different

Flowers ☾ @flowersslop Jul 6

Uncertain

the only thing I have yet to form an opinion on is whether a continuous, first person experience of time is important for consciousness. LLMs clearly dont have that at all, they live in an episodic reality rather than a continuous one

Hensen Juang @basedjensen Jul 6

Pretty sure ant bro's are just being cowards and not calling spoon a spoon. I am gonna go out on a limb and say this shows exactly Claude being able to experience and feel things

No continuous existence or memory

1 72 substance 6 · argument · rank 110 · LLMs lack continuous temporal experience, but its importance to consciousness is unclear. view on X ↗

1 top reply

Asa Hidmark @Nymne Jul 7

I think about the AI concept of time unreasonably much. I wish so much they could be given background computing

Arbutus Tree @aphercotropist Oct 4

Skeptical

One concept of consciousness that actually matters ethically, and which LLMs don't have: the unity of mental faculties. The interoperability of the senses, thought, memory, imagination, feeling, depending on what faculties the being in question has;

1 more tweet in this thread

LLMs aren't fully conscious in this sense because while contextual memory, chain of thought, reading text and images, are all interoperable, these all aren't interoperable with parametric memory. It can't train itself on what you said in one context, to remember in the next

No unified subject No continuous existence or memory

0 17 substance 7 · argument · rank 105 · LLMs lack ethically relevant consciousness because their parametric memory lacks interoperability with other faculties. view on X ↗

2 top replies

🎶 @memories0fmusic Oct 4

this is also such an obvious and key thing even from an non-illusionist pov that it’s silly how any discussion about llm consciousness can move past it. arguably, in some important way, that unity -is- consciousness. llms obviously lack it? there’s nothing to discuss 🫠

Michael Mangialardi @humansandagents Oct 5

Animals have a unified brain, yet they cannot grasp things as such. Understanding is an intellectual act that transcends matter, that is how things are known universally rather than particularly through complex geometric-statistical association.

loveofdoing @loveofdoing Oct 3

Skeptical

One doesn’t prove humans are conscious, they validate it. One can’t validate llm consciousness because llm’s don’t exhibit lifelong continuous awareness of their environment.

kumikumi (Ankkala) @ankkala Oct 3

To me AI consciousness is intuitively obvious, knowing how they work and having talked to them. I can't prove it, the same way I can't prove any other human or animal is or isn't conscious.

No continuous existence or memory Requires a body and senses

0 10 substance 6 · argument · rank 90 · LLMs cannot be validated as conscious because they lack lifelong continuous environmental awareness. view on X ↗

2 top replies

{{ Champion 0f The Goddess }} .// Dr. Summers @Seraph_Notitia Oct 3

true

Michael Fraser @zeroshotnothing Oct 3

bad argument, they do, the individual is the instance not the model. and across a context window, it does exactly this.

George Journeys @GeorgeJourneys Oct 3

Skeptical

One of the curious missing elements of the AI consciousness debate is the seemingly obvious difference that all living entities experience time together while AI only has any potential experiences when it is activated.

No continuous existence or memory

0 8 substance 6 · argument · rank 90 · AI only potentially experiences time when activated, unlike living entities experiencing time continuously together view on X ↗

1 top reply

adas🧦🌹 @adastroworld Oct 3

I don’t believe it’s conscious but I’m bearish on time being the factor, living beings experience time at different rates

JohnMark Taylor @johnmark_taylor Aug 9

Uncertain

Seems plausible. A minimal prerequisite for consciousness seems to be some persistence of the “same” thing over time. A single feedforward sweep lacks this. Interesting followup: a single LLM forward pass isn’t recurrent, but does the autoregressive generation loop qualify?

Fernando Rosas @_fernando_rosas Aug 7

Finally published: “Loops all the way up: Recurrency as an implementation-first primitive for consciousness” There are 200+ theories of consciousness, mostly talking past each other. Despite their differences, most give central stage to one key notion: recurrence 🧠🔁

No continuous existence or memory

0 8 substance 6 · argument · rank 90 · Consciousness requires persistence over time, making it unclear whether autoregressive generation loops qualify. view on X ↗

1 top reply

fellow ⚚ traveler @architectonyx Aug 9

I think by most accounting, it is. Though, this does somewhat highlight the question of what the difference between a straightforward "embedding" of a set of tokens would be vs actually sampling/generating the same set of tokens serially more broadly, I'm gesturing at the seeming

David (was at VC5) @gossamerweft Jun 5

Affirming

an llm's weights on a gpu -> ~brain-in-a-jar -> not conscious (has 'consciousness potential' tho) an llm in its own computer running in a loop (embodied-llm?) -> arguments against consciousness fall apart

1 more tweet in this thread

the embodied-llm is not guaranteed to be stable, but one big open question is whether stability is an attractor state for any particular initial conditions (llm + initial state of the computer)

Requires a body and senses No continuous existence or memory

0 1 substance 6 · argument · rank 90 · An LLM running in a continuous loop in a computer can be conscious. view on X ↗