I am a generalist quantitative researcher. I am open to volunteering and paid work. I welcome suggestions for posts. You can give me feedback here (anonymously or not).
Vasco Grilođ¸
Karma: 11,764
Hi Austin and Carol. Great point and case study.
Hello. I liked you questions. Keep up the curiosity.
Hi Toby. I like this experiment. The arguments you gave for the featured page made sense to me.
Hello. Thanks for sharing. I had already shared the thoughts below with Jasmine a few weeks ago. I am publishing them here in case others find them useful.
I appreciate the intentions behind the survey, and I would like to take part in principle. However, I do not know how to answer the questions in practice. I feel like they would have to be operationalised in much more detail for me to give a probabilistic forecast.
I can see âpost-AGIâ having already been achieved, or never being achieved depending on how it is defined.
I do not know what âAnimal sufferingâ means. Does moving away from noxious stimuli count as suffering? Does it require metacognition? One could say it refers to suffering in the phenomenal sense (negative qualia), but I do not think phenomenal consciousness exists (I endorse illusionism), or that sentience as traditionally conceptualised is relevant for ethics (relatedly).
I understand âAnimal sufferingâ is that of animals with non-trivial moral weight, but I do not know what this means in practice given my large uncertainty. I can see the moral weight of shrimps ranging from 10^-12 to 1. I could speculate about a distribution, get its mean, and then compare it with a guess for what is trivial moral weight. However, I feel like my view is better described by âI have basically no clue about whether shrimps have trivial moral weight or notâ. Likewise for many of the questions in the survey when I try to think about potential operations.
With respect to âBenchmarks will become uselessâ, I think it matters e.g. whether we are talking about 90 %, 99 %, or 99.9 % of current benchmarks, whether we are talking about current benchmarks, or benchmarks at some point in the future (what point?), and the degree to which they will become useless (e.g. 90 %, 99 %, or 99.9 % useless).
These are not minor details for me in the sense my answers could range from strongly disagree to strongly agree depending on definitions. The results of the survey above could still provide some vibes about what panelists are thinking, especially if they are repeated across time (as I understand you are doing). However, I think they will still be very difficult to interpret. As a rule of thumb, I would clarify the questions up to a point where they could be published on Metaculus. I understand this would decrease engagement. On the other hand, with the current methodology, I will put very little weight on the results. I think the results will be influenced a lot by how people are interpreting the questions, and I also do not trust forecasts produced in a few minutes about such hard questions.
All this said, your methodology may well make sense given your target audience and goals.
Hi Matthew. Great post.
This uncertainty [about consciousness] gives us some reason to be suspicious of very low estimates of animal consciousness.
I think consciousness is usually defined in a way that is either unfalsifiable or too vague for estimating a probability to be useful (like what-is-it-likeness). I believe it would be better to focus just on what animals can do (for instance, probability of a given animal having a score of at least x % on a given test), which can be be empirically assessed.
If a creature behaves dramatically, as would be expected if it were in intense pain, we should think itâs reasonably likely that the creature does, in fact, feel intense pain [as intense as intense pain in humans].
I agree that is a possibility, but I struggle to quantify how likely it is. Their pain during dramatic behaviours could be much more intense than their typical experiences, but still much less intense than human pain.
A third consideration is evolutionary: there is every evolutionary reason for simple creatures to feel intense pain. Pain in humans causes them to avoid various behaviors that are likely to be harmful. Pain in fish, insects, and decapods serves similar functions. We should expect intense pain in these creatures to be adaptive for the same reason that pain in humans is adaptive. If a creature just felt mild discomfort, that would be unlikely to sustain the kind of behavior needed to safeguard survival. Intense pain could be adaptive for eliciting rapid and dramatic behavior.
Evolutionarily speaking, nothing would change if the rewards and punishment had a value of âalternative valueâ = k*âvalueâ, where k is positive? The expected value of each action would have been k times as large, but the ratio between the expected value of any 2 actions would have been exactly the same. k can in principle be arbitrarily close to 0. So experiences could have an intensity arbitrarily close to 0 while leading to exactly the same behaviour?
A similar argument can be employed against many of the objections to significant pain in animals. Animals have relatively simple brains. But if we came across a human with a comparatively simple brain, who seemed to respond aversively to external damage the way animals do, we would be hesitant to be confident that its conscious experience is only mild. We should have similar skepticism directed towards arguments for such conclusions about animals.
This argument only applies to mammals at most?
The eighth and final argument for taking seriously the suffering of simple creatures: the most detailed report to date which attempted to estimate intensity of valenced experience in animals guessed that even simple creatures were, in expectation, intensely conscious. This report came from the team of Fischer et al. (2024). Using a mixed function taking into account various different behavioral proxies, the final estimate (p.233) was that shrimp suffer 8% as intensely as humans and carp 24% as intensely. The median estimate was that shrimp suffered 5% as intensely as humans.
The estimates for sentience-adjusted welfare ranges in Bob Fischerâs book are calculated for 3 models. The value for shrimps as a fraction of that of humans is 40 % for the equality model, 0.035 % for the neurophysiological model, and 6.6 % for the simple additive model. However, it is very unclear to me how to combine these in a principled way. I would say that the key takeaway is that a best guess equal to any of those 3 values is reasonable. Moreover, the 3 models only cover a very tiny fraction of possibilities. Models leading to estimates of the type âestimate for the neurophysiological modelâ^âexponentâ with high exponent could easily lead to very small sentience adjusted welfare ranges as a fraction of that of humans. For example, 1.23*10^-7 (= (3.5*10^-4)^2) for shrimps for an exponent of 2. How do you rule out such models? I read the whole book, and I did not find any justification.
Using a more integrated model that took into account even more inputs (see Duffy, 2024, table 2) their mean estimate was that shrimp suffered a whopping 19% as intensely as humans, bees 14.8%, and carp 17.5%.
Bobâs book superseded that work. âFor updates, please see this [Bobâs] bookâ.
And when humans lose many neurons, so long as their brains perform similar functions, we donât normally think they are vastly less conscious.
This could be true, and animals still have experiences of astronomically low intensity? Say the intensity of experiences (or the sentience-adjusted welfare range) is proportional to âindividual number of neuronsâ^2. Decreasing the number of neurons by 1 % would decrease intensity by 2 %. Nothing dramatic? Estimates for the number of neurons in humans range from 61 billion to 99 billion. So the intensity of experiences would vary by a factor of 2.63 (= (99/â61)^2) specifically because of differences in the number of neurons, which seems like a reasonable upper bound to me. If so, the experiences of shrimps could have an intensity as low as 10^-12 (= (10^-6)^2) times that of the experiences of humans, as shrimps have 10^-6 times as many neurons as humans.
On the other hand, I am also open to the experiences of shrimps being roughly as intense as those of humans. For intensity proportional to âindividual number of neuronsâ^âexponentâ, values from 0 to 2 cover the best guesses that I consider reasonable.
Shriver (2024) provides a number of reasons to be skeptical of neuron count proxies.
âNumber of neuronsâ^0.188 explains very well the estimates for sentience-adjusted welfare ranges in Bobâs book.
More importantly, the highest estimate for the number of neurons is humans is 1.62 (= 99â61) times the lowest estimate, whereas humans have 1 M times as many neurons as shrimps. So I think the number of neurons explaining very little of the variation in the intensity of human experiences is practically no evidence for human experiences being roughly as intense as those of shrimps.
I suggest emailing the applicants a copy of the the submitted answers, although they can also save them in a doc (I usually do this when I apply to jobs).
I suggest adding an option to delete draft posts.
I see.
AcÂcelÂerÂatÂing value growth is as valuable per perÂcentage point as deÂcreasÂing value exÂtincÂtion risk, whatÂever the fuÂture looks like
Hi Aidan.
Among our next priorities will be to re-evaluate Longviewâs Emerging Challenges Fund and EA Fundsâ Long-Term Future Fund in our reducing catastrophic risks cause area and EA Fundsâ Animal Welfare Fund in our animal welfare cause area â as we last evaluated these evaluators in 2023.
Have you considered evaluating all evaluators of a given area in the same year, alternating between catastrophic risks, animal welfare, and human welfare besides catastrophic risks? This would ensure all the recommended funders of a given area were assessed based on the same methodology at roughly the same time.
Hi Toby.
Not a direct answer to your post, but this is a good example of why Iâm not currently planning on downranking AI-written posts.
Good to know. I agree the degree to which a post is AI-written should not directly influence its ranking. Karma influences the ranking, and users can vote taking into account how much AI was used (although I think it is better to value just based on the value of the post).
Thank you, Sarah. Your thoughts made sense to me. Best wishes for your new role.
I had already acknowledged that the analogies were no longer helping, and then I fell into offering another one, so I concede that mistake. I do not think my underlying point is coming across, but I have laid it out as clearly as I can in our exchange.
I personally found your analogy quite helpful. It gave me a chance to explain why I think many properties considered categorical are actually graded. In addition, I think trying to present properties which are categorical is important to illustrate sentience can be so too. The absence of any categorical properties would arguably be strong evidence against sentience being categorical.
Your final question captures the crux: if I regarded sentience as merely a matter of human convention, I would not retain it as a scientific gate.
Makes sense. I think sentience is a matter of human convention, which is why I would not use it as a scientific gate for the affective capacity analysis.
I think you should just expect this, and the answer is not to deny the possibility of St Petersberg lotteries in objective quantities, but to figure out a good way to deal with them (e.g. ignore small enough probabilities, use a bounded utility function, use commitments, use bracketing of some form), or accept that they raise difficult normative problems.
Why do you think I should expect all distributions to have infinite or undefined expected value instead of rejecting such distributions?
Are you saying you believe in the existence with 100% credence in anything (such as ghosts) that is not ruled out by current evidence, as long as its effects couid be arbitrarily small and are so far indistinguishable from its nonexistence? Or must it also have no important normative implications (under classical utilitarianism?)?
I corrected my answers in my past comment. You can see what I crossed out, and wrote after âEdit after Michaelâs comment just belowâ.
Some could be vengeful and want harm to fall upon those that have caused them harm in their lives (whether or not they enact it themselves).
For the âghosts [I mentioned in my last comment] which have exactly 0 measurable effects on the worldâ, the benefit and harm they could cause would be sufficiently small to be practically negligible.
Hi Alistair. Thanks for asking about the effects on soil animals. I think they may be much larger than the effects on farmed animals as I illustrate in the post you linked to. At the same time, I do not know whether decreasing/âexpanding agricultural land increases or decreases the welfare/âsuffering/âhappiness of soil invertebrates.
What do you think about the St Petersburg problem now?
Thanks for pushing me to think about this more. I had only looked into your post a few days after you published it around 3 years ago, but just had a look again. I agree the money pump you described there does not require prospect which could have an infinite value. It only requires prospects with infinite expected value as you have been saying.
I think there is exactly 0 empirical evidence for distributions with infinite expected value for the same reasons I believe there is exactly 0 empirical evidence for infinities. As far as I can tell, exactly 100 % of the empirical evidence that could ever be gathered in principle could be exactly 100 % explained by distributions with finite expected value. Do you agree? I agree distributions should not have a maximum because one cannot be exactly 100 % confident there are not higher values. However, a lack of maximum does not imply infinite expected value.
If thereâs no finite upper bound on what the value the specific probability distribution can take, how can you be 100% confident it is not a probabilistic mixture with a distribution with infinite or undefined EV (but finite for every actual value), like 0.0001% probability to it being drawn from something like a St. Petersburg lottery?
This proves too much? One could argue there is a probability above exactly 0 of any given quantity being a probabilistic mixture involving a distribution with infinite or undefined expected value. In this case, all distributions would have an infinite or undefined expected value? For me this is a bitter bullet to bite than fully rejecting distributions with infinite or undefined expected value.
I think your arguments re infinities conflict with Occamâs razor, and the principle of indifference should be applied across models within the same complexity (or submodels), otherwise you will assign 0 or too little credence to simpler models that are special cases, e.g. parameter value=0 or effectively eliminating some type of feature.
I would apply the principle of indifference to models which explain the same empirical evidence. If a curvature of 0 had a probability above 0, and the values of the curvature just above 0 followed a continuous distribution, the curvature of 0 would be infinitely more likely than a positive curvature arbitrarily close to 0. This is very counterintuitive to me because the curvatures would have an arbitrarily close explanatory power. I would rather concede all universe models are wrong with probability 1 while acknowledging simpler ones are more useful for further scientific progress all else equal.
Would you assign exactly 0 probability to photons having exactly 0 mass?
Yes. I think there will always be infinitely many values arbitrarily close to 0 which explain exactly the same empirical evidence as a value of 0.
Exactly 0 probability to there being no additional fundamental force, because it could just be vanishingly weak? And shouldnât this get you to infinitely many fundamental forces? For any finite set of fundamental forces, you could posit another one and just say itâs very weak.
Yes.Edit after Michaelâs comment just below. I would assign a probability of exactly 0 to any physical law because there are arbitrarily many physical laws arbitrarily close to any physical law. So I would also assign a probability of exactly 0 to any set of physical laws, including the set of laws involving any given number of fundamental forces.Exactly 0 probability to there not being ghosts [ghosts existing with probability of exactly 1], because their effects could just be very small or rare?
Yes, but the effects of the ghosts would have to be sufficiently small or rare to be unfalsifiable.I assume the existence of ghosts is falsifiable under some typical definitions. Likewise for some defitions of God. Edit after Michaelâs comment just below. I would not assign a probability of exactly 1 to something falsifiable.I think consistently applying your arguments suggest you should assign 0 probability in these cases, and so your models blow up in complexity and you become too credulous, contrary to Occamâs razor.
Do you see any undesirable implications of believing in ghosts which have exactly 0 measurable effects on the world? I think this is effectively the same as not believing in such ghosts.
On the scientific question, yes: under a categorical account of sentience, arbitrarily small changes near the relevant biological threshold could in principle place otherwise extremely similar organisms on opposite sides of it.
Would it make sense to run WFFâs affective capacity analysis for only one of 2 organisms which only differ infinitesimally?
I do not regard this [the sentence quoted above] as incoherent. Continuous underlying variables can produce categorical system-level events; a neuron, for example, may or may not fire an action potential as its membrane potential crosses a threshold. This analogy does not establish that sentience is categorical, but it shows that gradual biological change does not rule out a categorical onset.
I found this example quite thought-provoking. Yet, I do not think neuronal firings are categorical. Membrane potential and neuronal firings are represented in the mid and bottom plots of Figure 9A below from section 9.1 of the book Neuronal Dynamics. Membrane potential varies continuously. It is not binary. So humans have to come up with a convention for which types of variations in it count as a firing. In principle, 2 different time series of the membrane potential could differ infinitesimally, and one have 0 firings while other has many firings which fall just short of counting as such. If one was interested in a property like the maximum membrane potential, it would make sense to analyse both time series
There is also no fact of the matter about whether a given membrane potential corresponds to a neuronal firing or not. The number of neuronal firings is a property of the time series of the membrane potential. It is not a property of a neuronal state. In contrast, I understand sentience (existence of affective experiences) is supposed to be a property of a state (of an entity at a given time). Membrane potential is a property of a state, and it is continuous/âgraded (not binary/âcategorical/âsharp).
The physical world varies continuously. So I think all properties tracking it have to be continuous too. People point out phase transitions as evidence of categorical changes, but they involve continuous changes.
Would you keep the sentience gate in WFF if you thought it was a matter of human convention instead of scientific fact?
I think falsifiability for a Bayesian could just mean we can imagine evidence that would warrant assigning very very low credences to the hypothesis.
Makes sense. I cannot imagine any evidence that would update me. However, if I did, there would be falsifiability.
Actual infinities are often falsifiable in principle in this way, in specific cases. I think the probability you should assign to you being infinite in spatial size is extremely small. Similarly for the Earth being infinite in size.
The way I see it, the probability of Earth having a radius larger than X tends to 0 as X goes to infinity. So I would say the probability of Earth having an infinite radius is exactly 0.
I think our credences that the universe is infinite in spatial extent should increase with our credences that the universe is globally flat (0 global curvature), which has been measurable.
Do we really have any evidence that the universe is globally flat? From Wikipediaâs page on the shape of the universe:
Final results of the Planck mission, released in 2018, show the cosmological curvature parameter, 1 â Ί = ΊK = âKc2/âa2H2, to be 0.0007Âą0.0019, consistent with a flat universe.[47] (i.e. positive curvature: K = +1, ΊK < 0, Ί > 1, negative curvature: K = â1, ΊK > 0, Ί < 1, zero curvature: K = 0, ΊK = 0, Ί = 1).
We have evidence that the universe is close to flat. However, there are infinitely many values arbitrary close to exactly 0. So applying some sort of principle of indifference results in a probability of exactly 0 of curvature being exactly 0 (or any other sharp value)?
As an aside, there have been many cases where quantities in physics were assumed to be 0, but then turned out to be just small values, like the mass of neutrinos. It often makes sense to round a quantity to 0 for simplicity, but a sufficiently small value would explain exactly the same empirical evidence.
Hi Matthew. Thanks for crossposting this. I am sharing below the comment I left on the original post.
Hi Matthew. Nice post.
I estimate going to the protest is worth donating a few hundreds to dollars to the Shrimp Welfare Project (SWP). For 30 billion insects raised per year (you say âThis single farm, Protix, raises tens of billions of farmed insectsâ), your assumption that âProtixâs continued operation will lead to an extra 300 billion farmed insectsâ implies each person increases by 0.01 pp the probability of 10 years of counterfactual impact. Maybe 3 years is more accurate considering the rest of the industry could increase their production to offset the loss of Protix. Even then, each person would save 3 M insects (= 30*10^6*3/â10). I estimate SWPâs Humane Slaughter Initiative (HSI) helps 15 k shrimps per $ (https://ââforum.effectivealtruism.org/ââposts/ââEbQysXxofbSqkbAiT/ââcost-effectiveness-of-shrimp-welfare-project-s-humane). Say saving 1 insect from Protix is as beneficial as HSI helping 1 shrimp, and that donations to SWP are roughly as cost-effective as HSI has been. In this case, going to the protest would be as beneficial as donating 200 $ (= 3*10^6/ââ(15*10^3)) to SWP.
Given the above, going to the protest may still be worth it for people close by, but it is not obvious. For no financial cost, and a time cost of 20 $/âh, it would have to take less than 10 h (= 200â20). On the other hand, there may be significant indirect benefits from going to the protest. In particular, connecting with people who care about neglected animals.
Hello Hazem. You can try to turn it into motivation to publish a post (including a linkpost/âcrosspost) for each untitled draft.