An illusion is perception not accurately representing external reality. So the perception by itself cannot be an illusion, since an illusion is a relation (mismatch) between perception and reality. The Müller-Lyer illusion is a mismatch between the perception “line A looks longer than line B” (which is true) and the state of affairs “line A is longer than line B” (which is false). The physical line on the paper is not longer, but it looks longer. The reason is that sense information is already preprocessed before it arrives in the part of the brain which creates a conscious perception. We don’t perceive the raw pixels, so to speak, but something that is enhanced in various ways, which leads to various optical illusions in edge scenarios.
cubefox
Karma: 1,898
I assume the minimum number was put into place in order to prevent another method of gaming the system.
I think it’s trainable mainly in the indirect sense: If someone has a lot of social grace, they can pull off things (e.g. a risque joke in front of a woman) that would be perceived as cringe or creepy in people with significantly less social grace. These latter people can become less cringe/creepy by learning not to attempt things that are beyond their social grace capabilities. (Which reduces their extraversion, in contrast to treating anxiety, which boosts extraversion.)
I think already children show significant differences in social grace. I remember a kid in elementary school who got bullied a lot because of his often obnoxious behavior. He had both very low social grace and very low social anxiety. I assume with time he learned to become less outgoing, because that wasn’t working in his favor. Outgoing behavior can be scaled down at will, but social grace can’t be easily scaled up. Even people who are good at social grace don’t explicitly know how to do it. It’s a form of nonverbal body/intonation language that seems to be largely innate and unconscious, perhaps controlled by an (phylogenetically) older part of the brain, e.g. the cerebellum rather than the cortex.
Of course that’s all anecdotal and speculation, but I would hypothesize that statistically the level of social grace of a person tends to stay largely the same over the course of their life. The main reason is that I think that lack of social grace is strongly related to ASD, which is relatively immutable. It seems people can’t change their natural “EQ” (which includes social grace) much beyond learning explicit rules about how to behave according to social norms, similar to how people can’t change their natural IQ much beyond acquiring more knowledge.
You might be surprised to learn that the most prototypical LessWrong user (Eliezer Yudkowsky) is a moral realist. The issue is that most people have only read what he wrote in the sequences, but didn’t read Arbital.
Human moral judgement seem easily explained as an evolutionary adaptation for cooperation and conflict resolution,
That’s not true: You can believe that what you do or did was unethical, which doesn’t need to have anything to do with conflict resolution.
and very poorly explained by perception of objective facts. If such facts did exist, this doesn’t give humans any reason to perceive
Beliefs are not perceptions. Perceptions are infallible, beliefs are not, so this seems like a straw man.
or be motivated by them.
Moral realism only means that moral beliefs, like all other contingent beliefs, can be true or false. It doesn’t mean that we are necessarily or fully motivated to be ethical. In fact, some people don’t have any altruistic motivation at all (people with psychopathy), but that only means they don’t care to behave ethically, and they can be perfectly aware that they are behaving unethically.
Unfortunately I think social grace can only be trained to a small degree, for reasons similar to ASD not being curable. Some people just have a natural social grace, others are much more socially awkward, and removing their social inhibitions too much may make them “cringe” or even “creepy”.
Current LLMs can already do this, e.g. when implementing software with agentic coding environments like Cursor.
This is also discussed here. LLMs seem to do a form of imitation learning (evidence: the plateau in base LLM capabilities roughly at human level), while humans, and animals generally, predict reality more directly. The latter is not a form of imitation and therefore not bounded by the abilities of some imitated system.
Thanks, that was an interesting post, it seems like an overall plausible theory. In fact more plausible than the recent one by Chipmonk you linked to, as your theory is much wider and somewhat includes the one by Chipmonk (per point two in your list).
I think one common reason for social anxiety is still missing in this list though: Fear of being humiliated. A rejection, or a cringe comment, can feel excessively humiliating to someone with social anxiety, even if they don’t believe the other person will feel awkward or will dislike them.
I think that’s indeed something exposure therapy can help with. Just thinking something like “this fear of humiliation is clearly exaggerated, let’s not worry about it” is like thinking “this fear of spiders is clearly exaggerated, let’s not worry about it”. It won’t help much because the fear by itself isn’t really what’s exaggerated, it’s the consequence of something that is exaggerated. The fear comes from correctly predicting that a spider touching you would freak you out excessively, just as you’re correctly predicting that you would feel excessively humiliated if a social faux pas or a rejection were to occur. It’s more a phobia than a proper anxiety. I don’t think you can reason yourself out of a phobia without some form of “exposure therapy”.
Though again, that’s only one additional potential cause for social anxiety which doesn’t apply to every case.
This seems interesting! One question: Why the focus on the true positive rate (TPR, sensitivity)? This is the probability of the lie detector being positive (indicating a case of deception), given that deception is actually taking place. Which is simply P(positive | deception). Unfortunately this metric doesn’t accurately measure how powerful a lie detector is. For example, the TPR would be maximal for a lie detector that always returns “positive”, irrespective of whether deception is or isn’t taking place. It’s even possible that both P(positive | deception) (sensitivity) and P(negative | no deception) (specificity) are high while P(deception | negative) is also high.
To adequately describe the accuracy of a lie detector, we would arguably need to measure the binary Pearson correlation (phi, MCC), which measures the correlation between a positive/negative lie detector result and the actual presence/absence of deception. A perfect correlation of +1 would indicate that the lie detector returns “positive” if and only if deception is taking place (and therefore “negative” if and only if no deception is taking place). A perfect anti-correlation of −1 would indicate that the lie detector always returns “positive” for no deception and “negative” for deception. And a value of 0 would indicate that the probability of the lie detector returning “positive” / “negative” is statistically independent of the model being / not being deceptive. In other words, the higher the correlation, the better the lie detector.
So to me it seems we should assume that actually a lie detector with high correlation (high accuracy) pushes the model toward honesty, not necessarily a lie detector with high TPR.
A population of AI agents could marginalize humans significantly before they are intelligent enough to easily (and quickly!) create more Earths.
Exactly. Bringing a new kind of moral patient into existence is a moral hazard, because once they exist, we will have obligations toward them, e.g. providing them with limited resources (like land), and giving them part of our political power via voting rights. That’s analogous to Parfit’s Mere Addition Paradox that leads to the repugnant conclusion, in this case human marginalization.
whether you’re asking a clarifying question that other audience members found useful
This is a frequent problem in math heavy research presentations. Someone presents their research, but they commit a form of the typical mind fallacy, where they understand their own research so well that they fatally misjudge how hard it is to understand for others. If the audience consists of professionals, often nobody dares to stop the presenter with clarificatory questions, because nobody wants to look stupid in front of all the other people who don’t ask questions and therefore clearly (right!?) understand the presented material. In the end, probably 90% have mentally lost the thread somewhere before the finish line. Of course nobody admits it, lest your colleagues notice your embarrassing lack of IQ!
I never did the rubber hand test, but recoiling from something doesn’t mean you believe you are in pain. (It doesn’t even necessarily mean you believe you have been injured, as you may recoil just from seeing something gross.) And the confusion from the rubber hand illusion presumably comes from something like a) believing that your hand has been physically injured and b) not feeling any pain. Which is inconsistent information, but it doesn’t show that you can be wrong about being in pain.
About (not) noticing pain: I think attention is a degree of consciousness, and that in this case you did pay some attention, just not complete attention. So you experienced the pain to a degree, and you knew the pain to the same degree. So this isn’t an example of being in pain without knowing it. Nor of falsely believing you’re not in pain.
There is a related case in which you may not be immediately able to verbalize something you know. But that doesn’t mean you didn’t know it, only that not all knowledge is propositional, and that knowing a propositional form isn’t necessary for knowing the what or how.
I disagree. Knowing that I’m in pain doesn’t require an additional and separate mental state about this pain that could be wrong. My being in pain is already sufficient for my knowledge of pain, so I can’t be mistaken about being in pain, or about currently having some other qualia.
If a doctor asks a patient whether he is in pain, and the patient says yes, the doctor may question whether the patient is honest. But he doesn’t entertain the hypothesis that the patient is honest but mistaken. We don’t try to convince people who complain about phantom pains that they are actually not in pain after all. More importantly, the patient himself doesn’t try to convince himself that he isn’t in pain, because that would be pointless, even though he strongly wishes it to be true.
You can define “being in the state of seeing 15648917” as “knowing you are seeing 15648917″, but there is no reason to do it, you will get unnecessary complications
I think it’s the opposite: there is no reason to hypothesize that you need a second, additional mental state in order to know that you are in the first mental state.
because it’s not knowing about state, it’s being in a state.
All knowing involves being in a state anyway, even in other cases where you have knowledge about external facts. Knowing that a supermarket is around the corner requires you to believe that a supermarket is around the corner. This belief is a kind of mental state; though since it is about an external fact, it is itself not sufficient for knowledge. Having such a belief about something (like a supermarket) is not sufficient for its truth, but having an experience of something is.
Memories of qualia are uncertain, but current qualia are not.
I am not an AI successionist because I don’t want myself and my friends to die.
An AI successionist usually argues that successionism isn’t bad even if dying is bad. For example, when humanity is prevented from having further children, e.g. by sterilization. I say that even in this case successionism is bad. Because I (and I presume: many people) want humanity, including our descendants, to continue into the future. I don’t care about AI agents coming into existence and increasingly marginalizing humanity.
To make this valid we need two things.
Phrase this only in terms of events/propositions, not in terms of single words like “context”, “situation” or “problem”. Probability only applies to events which occur or don’t occur, or to propositions that are true or false. Otherwise it is unclear what e.g. P(x | Situation) means.
Bayes theorem involves exactly two events (or propositions), so we must make sure we express merely similar sounding events (Situation looks like this / Situation like this arise) with one and the same event, in order to not exceed the total amount of two events overall.
Before you read on, you may want to first try the above approach yourself. The following is my attempt at a solution.
Formalization attempt
Here is a possible formalization which uses only two propositions (albeit with two “free variables” and the indexical “current”, which arguably is another variable):
Idea A applies to the current context.
The current context is of type B.
P(Idea A applies to the current context | The current context is of type B) = P(The current context is of type B | Idea A applies to the current context) * P(The idea A applies to the current context) / P(The current context is of type B)
Or perhaps more abstractly:
Not sure whether this accurately captures what you had in mind. It probably needs to be refined.
The first paragraph suggests both cases can be phrased as “idea is applicable”.
No, it is not at all misleading. He is quite explicit about that in the linked Arbital article. You might want to read it.
They definitely would not. They would immediately qualify as moral realist. Helpfully, he makes that very clear:
He explicitly classifies his theory as cognitivist theory, which means it ascribes truth values to ethical statements. Since it is a non-trivial cognitivist theory (it doesn’t make all ethical statements false, or all true, and your ethical beliefs can be mistaken, in contrast to subjectivism) it straightforwardly classifies as a “moral realist” theory in metaethics.
He does argue against moral internalism (the statement that having an ethical belief is inherently motivating) but this is not considered a requirement for moral realism. In fact, most moral realist theories are not moral internalist. His theory also implies moral naturalism, which is again common for moral realist theories (though not required). In summary, his theory not only qualifies as a moral realist theory, it does so straightforwardly. So yes, according to metaethical terminology, he is a moral realist, and not even an unusual one.
Additionally, he explicitly likens his theory to Frank Jackson’s Moral Functionalism (that is indeed very similar to his theory!), which is considered an uncontroversial case of a moral realist theory.