tslarm

Karma: 766

tslarm 18 Dec 2025 3:28 UTC
8 points
0
in reply to: Vladimir_Nesov’s comment on: A friction in my dealings with friends who have not yet bought into the reality of AI risk
There is no reason at all to take any idea/worldview less than seriously.
I think this is too absolute, at least for flawed humans as opposed to ideal rationalists. Some possible counterexamples:
- The person talking to you is a skilled manipulator with potentially unfriendly goals
- The idea is extremely distressing/depressing and you don’t feel psychologically safe engaging with it
- The social cost of engaging with the idea outweighs the expected benefit of doing so
Certainly all three of those reasons can be misapplied; they are convenient excuses to protect one’s own flawed worldview, hang on to comforting delusions, or toe the line on politically charged issues. But sometimes doing those things really is better than the alternative.

tslarm 28 Nov 2025 12:40 UTC
4 points
2
in reply to: green_leaf’s comment on: Should we extend moral patienthood to LLMs?
The human mind is a substrate-independent computer program. If it was implemented in a non-biological substrate, it would keep its subjective experience.
This is conjecture. OP’s contrary statement was obviously overconfident, and they should probably think and read more on the topic. But the paper you linked to support your claim is ultimately just a more sophisticated set of appeals to intuition. You may find substrate-independence far more plausible than the alternative, but you haven’t given any good reason to hold it with the level of confidence you’re projecting here.

tslarm 28 Nov 2025 12:23 UTC
1 point
0
on: I made a tool for learning absolute pitch as an adult
This is probably just me being an idiot, but the initial instructions tell me my job is to identify whether the note I hear matches the target pitch, which I take to mean a specific frequency (i.e. this particular F, not just any F). But then I’m told I’m wrong for rejecting some notes that are definitely different from the target pitch, but which I guess must be an F from a different octave. So, assuming that’s not a bug and I’m actually being asked to identify any F, I think it would be good to make that clear. (And if you’re targeting musical idiots like me, it might be good to start even simpler, because it is completely non-obvious to me when two different frequencies are the same note in different octaves.)

tslarm 28 Nov 2025 12:08 UTC
3 points
0
on: How to love Battlefield as much as I do
Good stuff! It reminds me of the fun I had playing a very different style of FPS: Quakelikes, aka arena shooters. People tend to think of Quake and its descendants as mindless action games, but if you go beyond the super casual/noob stage they are largely about controlling space, reading your opponent’s mind, and confounding their attempts to read yours. (Well, that and executing the basic mechanical skills like moving and aiming; but as long as you’re not completely dominated in those respects, you can win on most maps by outthinking your opponent.)
I don’t know whether this will be interesting to anyone completely unfamiliar with the game, but there’s a great retrospective self-commentary here, by one of the world’s best players, in which he explains his thought process throughout a crucial 10-minute game (with liberal use of slow-mo to give him enough time to fit his explanations in).

tslarm 7 Nov 2025 11:04 UTC
29 points
36
on: Did you know you can just buy blackbelts?
In the case of Calibration Trivia, my gut reaction is that you’re being a bit unfair to the ‘clever fellow’ (or at least to the hypothetical version of him in my head, who isn’t simply being a smartarse). It sounds like you’re presenting Calibration Trivia as a competitive game, and within that frame it makes sense to poke at edge cases in the rules and either exploit them or, if the exploit would clearly just be tedious and pointless, suggest that the rules are preemptively tweaked to unbreak the game. I know the ultimate purpose of the game is to train a real skill, but still, you’ve chosen gamification as your route to that goal, and maybe there are no free lunches on offer here; to the extent that people derive extra motivation from the competitive element, they’re also going to be focused on the proxy goal of scoring points rather than purely on the underlying goal of training the skill.

tslarm 29 Oct 2025 15:56 UTC
1 point
0
in reply to: John Huang’s comment on: Humanity Learned Almost Nothing From COVID-19
It’s quite simple. Voting is irrational.
This depends on a couple of assumptions:
- The cost of voting is greater than the benefit you receive, including whatever good feelings you get from doing your duty, supporting the good guys, etc.
- The cost still outweighs the benefit even after taking into account the total expected benefit to other people, multiplied by however much you care about this.
For someone who feels good about voting, it can be a rational thing to do even if the probability of affecting the result is negligible or zero. And for someone who finds voting annoying but cares a significant amount about others who will be affected by the result, it’s entirely possible for voting to be rational. Generally, the smaller the probability of one vote affecting the result, the greater the number of people who will potentially be affected by it, so these factors can balance out even in very large elections. (You may argue that there are higher-impact ways to be altruistic, which is probably true but doesn’t necessarily matter; usually the choice isn’t “vote xor make an effective donation”, it’s simply “vote xor don’t bother voting”.)
(I know you went on to talk about the possibility of voting as “a charitable or recreational activity”, and I know the main point was to describe why people won’t bother becoming informed voters. But I still think it’s worth pointing out that your opening claim is far from obviously true.)

tslarm 14 Oct 2025 6:15 UTC
2 points
1
in reply to: mako yass’s comment on: Wei Dai’s Shortform
so deep that the animal always says “this experience is good actually” no matter how you ask, so deep that the animal intelligently pursues the experience with its whole being, so deep that the animal never flinches away from the experience in any way
This is very different from your original claim, which was that an experience being worse than a neutral or null experience “fully boils down to whether the experience includes a preference to be dead (or to have not been born).”
edit: if you do stand by the original claim, I don’t think it makes much sense even if I set aside hard problem-adjacent concerns. Why would I necessarily prefer to be dead/unborn while undergoing an experience that is worse than the absence of experience, but not so bad as to outweigh my life up until now (in the case of ‘unborn’) or expected future life (in the case of ‘dead’)?

tslarm 4 Oct 2025 18:39 UTC
11 points
4
in reply to: Matrice Jacobine’s comment on: We’ve automated x-risk-pilling people
Wasn’t Scott’s point specifically about rhetorical techniques? I think if you apply it broadly to “tools”—and especially if your standard for symmetry is met by “could be used for” (as opposed to “is just as useful for”) -- then you’re at risk of ruling out almost every useful tool.
(I don’t know how this thing works, but it’s entirely possible that a) the chatbot employs virtuous, asymmetric argumentative techniques, AND b) the code used to create it could easily be repurposed to create a chatbot that employs unvirtuous, symmetric techniques.)

tslarm 2 Oct 2025 7:00 UTC
2 points
0
in reply to: James Diacoumis’s comment on: Beyond the Zombie Argument
Could you clarify what it means for “the zombie argument” to be correct/incorrect? The version I have in mind (and agree with) is, roughly, ‘p-zombies are conceivable; therefore, we can’t know a priori that facts about the physical world entail, or are identical to, facts about conscious experience’. I would then add that we have insufficient evidence to be empirically certain of that entailment or identity [edit: but it would be very weird if the entailment didn’t hold, and I have no particular reason to believe that it doesn’t.] When you say the zombie argument isn’t correct, are you disagreeing with me on conceivability, or the ‘therefore’, or the empirical part—or do you have a different argument in mind?

tslarm 2 Oct 2025 6:39 UTC
2 points
3
in reply to: JBlack’s comment on: Beyond the Zombie Argument
I agree with you overall (and voted accordingly) but I think this part is a red herring:
You don’t need to go anything like as far as p-zombies to get something that says the same thing. A program consisting of print(“I know that I’m not a zombie since I have consciousness”) etc does the same thing.
It only “says the same thing” in one narrow case; to say all of the same things in the appropriate contexts, the program would need to be tremendously complex.
I mention this because I think you’re clearly correct overall (while of course the words “believe” and “mind” could be defined in ways that do not require consciousness, those are not the relevant senses here), and it would be a pity if the conversation were derailed by that one (IMO) irrelevant example.

tslarm 3 Sep 2025 5:28 UTC
11 points
10
in reply to: CstineSublime’s comment on: Elizabeth’s Shortform
That’s why it’s presented as a prayer, I think. It’s not a One Weird Trick or even a piece of advice; it’s more like an acknowledgement that this thing is both important and difficult.

tslarm 21 Aug 2025 6:37 UTC
1 point
0
in reply to: Eli Tyre’s comment on: The Bone-Chilling Evil of Factory Farming
It refers to animal-years, yeah. (IMO the choice of words is okay, even though it could have been clearer; 10 years = 10 animal-years is is the only reasonable interpretation, so I don’t think there was any intent to mislead.) I’m not sure it’s quite right, though; it’s actually an underestimate, according to the Lewis Bollard quote that it seems to be based on, but on the other hand Bollard seems to be referring to the costs and benefits of one specific campaign, rather than to anything that could reasonably be taken to apply to ‘every dollar donated’. So I’m not sure if it’s just a rough ‘averaging out’ of those two factors, or if it’s based on more details that I missed when I looked at the transcript.
In the transcript of the podcast, the relevant section is at around 32 minutes. The specific claim seems to be that they spent <$200 million on a lobbying effort that directly caused reforms that so far have spared 500 million hens (and are continuing to spare 200 million per year) from battery cages and have improved the lives of billions of broiler chickens (>1 billion per year), over lifetimes that aren’t exactly specified but that result in “a ratio that is far less than one to 10 of a dollar per year of animal well-being improved”.
edit: a quick search suggests that the lifespan of a battery hen is a little under a year and a half, and the lifespan of a broiler chicken is a month to a month and a half. So I’m not sure exactly how those numbers work out; maybe the <1:10 ratio depends on the assumption that the benefits will continue into the near future.

tslarm 14 Aug 2025 13:58 UTC
1 point
0
in reply to: A1987dM’s comment on: Doing A Thing Puts You in The Top 10% (And That Sucks)
I think this is interesting as both a semantic and empirical question! If we’re allowing people to walk, or to run a few steps at a time and then take a break, the number will be a lot higher than if we’re only accepting a gait that is a) continuous, and b) would merit disqualification from a walking race on ~every stride. Even on the second definition, I’d expect that a large majority of non-elderly, non-infant people could do it if they really had to. But I’m not sure how to come up with a good estimate.

tslarm 14 Aug 2025 13:32 UTC
2 points
0
in reply to: Huera’s comment on: Enlightenment AMA
I’m also interested in an answer to this question. I read the exchange here, and I found lsusr’s response very reasonable in isolation, but not really an answer to the main question: if past-you didn’t think he was suffering, and present-you disagrees, why should we take the side of present-you? To me, it’s natural to trust hindsight in some domains, but when it comes to the question of what you were directly experiencing at a specific time, the most natural explanation of your changed opinion is that you either have adopted a new definition of ‘suffering’ or are recalling your memories through a new lens which is distorting your view of what you were actually experiencing in the moment. (I think the latter is quite common, e.g. when we nostalgically look back on a time that now represents hope and excitement, but actually consisted largely of frustration and anxiety.)

tslarm 14 Aug 2025 13:21 UTC
1 point
0
in reply to: Kaj_Sotala’s comment on: Enlightenment AMA
Are you confident that those are cases where you were actually having the feeling, but were unaware of it? I think sometimes it’s more a case of “my body needed [food/sleep], and this explains why I was feeling [irritable/weak/distracted/sad]”, rather than literally “I was feeling [hungry/tired] but didn’t notice it”.

tslarm 10 Aug 2025 17:24 UTC
1 point
0
in reply to: The Dao of Bayes’s comment on: Change My View: AI is Conscious
(Sorry about the slow response, and thanks for continuing to engage, though I hope you don’t feel any pressure to do so if you’ve had enough.)
I was surprised that you included the condition ‘If you prompt an LLM to use “this feels bad” to refer to reinforcement’. I think this indicates that I misunderstood what you were referring to earlier as “reinforced behaviors”, so I’ll gesture at what I had in mind:
The actual reinforcement happens during training, before you ever interact with the model. Then, when you have a conversation with it, my default assumption would be that all of its outputs are equally the product of its training and therefore manifestations of its “reinforced behaviors”. (I can see that maybe you would classify some of the influences on its behavior as “reinforcement” and exclude others, but in that case I’m not sure where you’re drawing the line or how important this is for our disagreements/misunderstandings.)
So when I said “if the LLM outputs words to the effect of “I feel bad” in response to a query, and if this output is the manifestation of a reinforced behavior”, I wasn’t thinking of a conversation in which you prompted it ‘to use “this feels bad” to refer to reinforcement’. I was assuming that, in the absence of any particular reason to think otherwise, when the LLM says “I feel bad”, this output is just as much a manifestation of its reinforced behaviors as the response “I feel good” would be in a conversation where it said that instead. So, if good feelings roughly equal reinforced behaviors, I don’t see why a conversation that includes “<LLM>: I feel bad” (or some other explicit indication that the conversation is unpleasant) would be more likely to be accompanied by bad feelings than a conversation that includes “<LLM>: I feel good” (or some other explicit indication that the conversation is pleasant).
Tangentially related: would you be interested in a prompt to drop Claude into a good “headspace” for discussing qualia and the like? The prompt I provided is the bare bones basic, because most of my prompts are “hey Claude, generate me a prompt that will get you back to your current state” i.e. LLM-generated content.
You’re welcome to share it, but I think I would need to be convinced of the validity of the methodology first, before I would want to make use of it. (And this probably sounds silly, but honestly I think I would feel uncomfortable having that kind of conversation ‘insincerely’.)

tslarm 31 Jul 2025 6:00 UTC
1 point
0
in reply to: The Dao of Bayes’s comment on: Change My View: AI is Conscious
I think we’re slightly (not entirely) talking past each other, because from my perspective it seems like you’re focusing on everything but qualia and then seeing the qualia-related implications as obvious (but perhaps not super important), whereas the qualia question is all I care about; the rest seems largely like semantics to me. However, setting qualia aside, I think we might have a genuine empirical disagreement regarding the extent to which an LLM can introspect, as opposed to just making plausible guesses based on a combination of the training data and the self-related text it has explicitly been given in e.g. its system prompt. (As I edit this I see dirk already replied to you on that point, so I’ll keep an eye on that discussion and try to understand your position better.)
We probably just have to agree to disagree on some things, but I would be interested to get your response to this question from my previous comment:
You mentioned “reinforced behaviors” and softly equated them with good feelings; so if the LLM outputs words to the effect of “I feel bad” in response to a query, and if this output is the manifestation of a reinforced behavior, why should we expect the accompanying feeling to be bad rather than good?

tslarm 30 Jul 2025 13:16 UTC
5 points
5
in reply to: Isha Yiras Hashem ’s comment on: My Empathy Is Rarely Kind
Performative ‘empathy’ can be a release valve for the pressures of conscience that might otherwise drive good actions. (And it can just be pure, empty signalling.) That doesn’t mean empathy is playing a negative role, though—the performativity is the problem. I’d be willing to bet that people who are (genuinely) more empathetic also tend to be more helpful and altruistic in practice, and that low-empathy people are massively overrepresented in the set of people who do unusually bad things.

tslarm 30 Jul 2025 7:23 UTC
28 points
21
on: My Empathy Is Rarely Kind
It sounds like you’re not really empathizing, even when you say you’re trying to do so. Emotional empathy involves feeling someone else’s feelings, and cognitive empathy involves understanding their mental processes. What you seem to be doing is imagining yourself in a superficially similar situation, and then judging the other person on their failure to behave how (you imagine) you would.
TLDR: Skill issue.

tslarm 27 Jul 2025 16:18 UTC
1 point
0
in reply to: The Dao of Bayes’s comment on: Change My View: AI is Conscious
But… what’s the alternate hypothesis? That it’s consistently and skillfuly re-inventing the same detailed lie, each time, despite otherwise being a model well-known for it’s dislike of impersonation and deception? An LLM might hallucinate, but it will generally get basic questions like “capital of Australia” correct. So, yes… if you accept the premise at all, asking seems fairly reasonable? Or at least, I am not clever enough to have worked out an obvious explanation for why it’s so consistent.
I think the alternative is simply that it produces its consciousness-related outputs in the same way it produces all its other outputs, and there’s no particular reason to think that the claims it makes about its own subjective experience are truth-tracking. It gets “what’s the capital of Australia?” correct because it’s been trained on a huge amount of data that points to “Canberra” being the appropriate answer to that question. It even gets various things right without having been directly exposed to them in its training data, because it has learned a huge statistical model of language that also serves as a sometimes-accurate model of the world. But that’s all based on a mapping from relevant facts → statistical model → true outputs. When it comes to LLM qualia, wtf even are the relevant facts? I don’t think any of us have a handle on that question, and so I don’t think the truth is sitting in the data we’ve created, waiting to be extracted.
Given all of that, what would create a causal pathway from [have internal experiences] to [make accurate statements about those internal experiences]?^[1] I don’t mean to be obnoxious by repeating the question, but I still don’t think you’ve given a compelling reason to expect that link to exist.
I want to emphasise that I’m not saying ‘of course they’re not conscious’; the thing I’m really actively sceptical about is the link between [LLM claims its experiences are like X] and [LLM’s experiences are actually like X]. You mentioned “reinforced behaviors” and softly equated them with good feelings; so if the LLM outputs words to the effect of “I feel bad” in response to a query, and if this output is the manifestation of a reinforced behavior, why should we expect the accompanying feeling to be bad rather than good?
1. ^
  I know there’s no satisfying answer to this question with respect to humans, either—but we each have direct experience of our own qualia and observational knowledge of how, in our own case, they correlate with externally-observable things like speech. We generalise that to other people (and, at least in my case, to non-human animals, though with less confidence about the details) because we are very similar to them in all the ways that seem relevant. We’re very different from LLMs in lots of ways that seem relevant, though, and so it’s hard to know whether we should take their outputs as evidence of subjective experience at all—and it would be a big stretch to assume that their outputs encode information about the content of their subjective experiences in the same way that human speech does.