6 ms·

“If you talk about having a subjective experience, then we don't know of any way to prove that even other humans than ourselves have one.“ Wittgenstein kinda b

• threethirtytwo 6mo ago
  The problem with your thinking here is that we are creating artificial beings now that display and output the same subjectivity. The argument you present like many arguments breaks down when the topic becomes self referential. It makes sense for other topics as analyzing subjectivity becomes pedantic when asking questions like why is the sky blue. But now subjectivity itself is in question. The argument you present calls for the subjectivity of others to be taken as true because all reality breaks down if we don’t… but what’s suddenly stopping you from applying the same assumptions to an LLM? That is the heart of the problem. People are questioning whether the burden of subjectivity is applicable to LLMs. Or another way to frame it… what makes humans rise to the level where we can assume their subjectivity is true? What is the mechanism and reasoning behind that? We can no longer simply assume human subjectivity is true because LLMs are now displaying outward behaviors that are indistinguishable from humans. Also stop relying on the wonderings of old school philosophers. We are now in times where you can basically classify their ideas as historically foundational but functionally obsolete and outdated. Think deeper.
  • mrandish 6mo ago
    > what makes humans rise to the level where we can assume their subjectivity is true? To dive into this specific question: to me, there's a better reason than the obvious functional utility of not treating other humans like NPCs. It's in three parts. First, is that I subjectively experience a rich and varied internal mental life (aka qualia). So, I have first-hand evidence that N equals (at least) 1 in terms of qualia existing in humans. Second, there are multiple lines of experimental evidence from fMRI, surgical and brain injury studies which indicate other human brains broadly function in ways similar to my own brain. Third, the consistency of the many self-reports of other humans I know and trust which strongly correlate with consistent reports from humans I've never met and who have little apparent motivation to deceive me (unlike those I know - if I were very paranoid). This all consistently supports a model of reality in which humans experience qualia broadly similar to my own. So when humans show external behaviors similar to my own, I make the reasonable inference that the internal causal mechanism broadly maps to what I internally experience when showing similar external behaviors (in contexts where the human is credible and has no motivation to be deceptive). The alternatives like "I'm a brain-in-a-vat ala The Matrix" or "I'm the sole subject of a constructed reality like the Truman Show" seem far less likely. But that's all general 'Philosophy of Mind', the slam dunk is that the question isn't just about humans but about humans compared to LLMs; in short, "Do LLMs experience human-like consciousness?" To me the answer is quite clear for three reasons: 1. LLMs are dramatically different than humans, mammals or even biological entities. They only vaguely emulate a few traits of neurons but otherwise work by different algorithms, at different scale, different speeds, connected in different ways on an entirely different physical substrate. 2. There's far less supporting evidence, and 3. There exists substantial negative evidence. 2. There are only two lines of evidence supporting LLM consciousness and the first is largely circumstantial, that a) LLMs possess some abilities previously only seen in humans. Specifically high-level verbal fluidity and linguistic manipulation along with instantly accessing a vast and diverse breadth of pre-trained information using a wide variety of non-linear relationships across many dimensions. While that ability is shockingly impressive, completely novel and can be quite useful, it's still only vaguely circumstantial because replicating some previously human-only abilities isn't evidence for the existence of other human traits like consciousness/qualia. However, LLMs are remarkably misleading for humans to reason about because the nature of LLMs essentially hacks our highly-evolved "judging intelligence/consciousness" heuristics. I'd argue we couldn't have designed a system to be more ideal at playing Turing's 'Imitation Game' and convincing humans they are human-like if we'd intentionally tried to. b) The second line of supporting evidence for LLMs is that they generate text which can describe internal subjective experiences much like a human would (as seen in the Dawkins / Claude transcript). Unfortunately, this isn't convincing because we know that LLMs were trained on human sample text to be 'imitation machines'. The algorithms were designed, tuned and tested to generate text output statistically optimized to plausibly simulate how a composite human would respond to the same input (including the invisible system prompt instructing: "You are a Large Language Model, not a human"). We even added a tiny degree of random variability to the processing of the statistical weights because we found that makes the simulation seem a bit more plausibly like what a composite human would say. In short, LLM 'self-reports' cannot be taken at face value any more than the performance of an actor we've hired to pretend something and strongly incentivized to never break character. Note: knowing this should elevate our skepticism to maximum. We're assessing an algorithmic system, designed and iteratively optimized across millions of generations to convincingly simulate the output of something different than what it innately is. 3. But to me the real clincher is the negative evidence against LLM consciousness/qualia. Unlike the philosophical puzzles around trusting human subjectivity, with LLMs we can directly look under the hood at how it works and the entire specialty of Mechanistic Interpretability exists to do exactly that (https://towardsdatascience.com/mechanistic-interpretability- https://towardsdatascience.com/mechanistic-interpretability-...). So we know with a fair degree of confidence that, despite what they may say, LLMs do not experience qualia in the way that humans and even other mammals do (which we have insight on from 'looking under the biological hood' with fMRI, surgical and brain injury studies). And that's why the case for human subjectivity is so much stronger than the frankly flimsy case for LLM subjectivity.
• vidarh 6mo ago
  Well, I am not certain whether it is the case. There is no need to be certain about that to treat other people the same way, because whether or not I believe they have a subjective experience has zero impact on how they react to stimuli. It does, however, have relevance when we consider whether or not other, non-human, entities can have consciousness: If we can't know what consciousness actually mean with respect to humans, that is a strong argument for not insisting that we know whether or not other entities are conscious. If we then choose to treat other humans purely on the assumption that they e.g. do feel distress the same way, we ought to consider that we do not what the pre-requisite to reach a level of awareness to feel distress is.
  • bonoboTP 6mo ago
    If they have no moral weight, then it wouldn't matter how they react to stimuli. They would be instruments to be used for the only conscious being, the solipsistic self. Maybe sometimes they would present powerful obstacles in the way of the ego person, and require some mouth noises and worse in their direction to align them with only moral weight of the universe, the solipsistic ego. A very ugly philosophy.
    • vidarh 6mo ago
      It matters how they react to stimuli because we do not know, and can not know, whether they have consciousness or not, and so the safer assumption is to treat them as if they do. But if we accept that we also need to consider that we do not know, and can not know whether other entities are conscious or not either. We can only tell whether they present as if they are. And we should consider that when deciding how to treat them. Furthermore we should be cautious about how high we set the bar.
• nurettin 6mo ago
  It's all words, man! -- Wittgenstein, probably
  • gizajob 6mo ago
    That about sums it up for the Tractatus yeah.