Michael St Jules 🔸

Karma: 13,070

Philosophy, global priorities and animal welfare research. My current specific interests include: philosophy of mind, moral weights, person-affecting views, preference-based views and subjectivism, moral uncertainty, decision theory, deep uncertainty/cluelessness and backfire risks, s-risks, and indirect effects on wild animals.

I’ve also done economic modelling for some animal welfare issues.

Want to leave anonymous feedback for me, positive, constructive or negative? https://www.admonymous.co/michael-st-jules

Michael St Jules 🔸 18 Aug 2026 19:26 UTC
2 points
0 ∶ 0
in reply to: Impartial Partially’s comment on: Bottom-Up Bracketing: A Person-Centered Response to Cluelessness
Is this a version of (weak) actualism as a person-affecting theory? In worlds where Alice doesn’t exist, she has no properties like welfare to provide us any reasons or welfare-based objections.
Weak actualism, as in the literature (Hare, 2007, Spencer, 2021 section 6), would say to pick among the options that maximize “self-conditional value”. An option is permissible only if it’s no worse than any other for the people who (ever) exist in the resulting outcome. In other words, assuming we choose the option A, we compare all the options pairwise on the basis of the reasons provided by the people who would exist given our choice of A,^[1] and if A is no worse overall than any other option on these comparisons, then it maximizes self-conditional value.^[2]
It sounds similar to what you wrote here, too:
Perhaps, on a person-centered approach to ethics, that is simply true (cf. Boonin 2014 and Pummer 2024): If we choose B, we will have made it the case that (i) no actual person has negative welfare, which they could object to on personal welfare grounds, and (ii) all actual persons have the maximum amount of positive welfare they could have had. (Person p2, once she exists, will have no welfare-based objection to our choosing B: We could not have made her better off. Nor will p1, who will not exist.) On a person-centered approach, we may therefore have maximized people’s welfare in the sense that matters.
1. ^
  Weighing the reasons by the probability that those reasons exist at all, say.
2. ^
  Weak actualism doesn’t need to say that all options that maximize self-conditional value are permissible. There are some cases where we can plausibly recommend choices between them.
  I also think maximizing self-conditional value as formalized in the literature is the wrong approach for an actualist when deciding between at least 3 options, because it’s myopic, and we should use backward induction on behalf of the people who would exist given an option. This can avoid the repugnant conclusion. I have an unpublished draft based on the idea of backward induction for weak actualism I’d be happy to share, and I discuss the problem a bit here.

Michael St Jules 🔸 18 Aug 2026 8:11 UTC
2 points
0 ∶ 0
in reply to: Michael St Jules 🔸’s comment on: Announcing: Cluelessness Critiques Competition
Hmm, I think I made some unjustified leaps in the above comment.
Some contingent moral patients in A could be determinately better off in expectation (conditional on existing) than the statistical average contingent moral patient in B, and some contingent moral patients in A could be determinately worse off in expectation (conditional on existing) than the statistical average contingent moral patient in B. So, with counterparts:
1. on bottom-up bracketing, we wouldn’t bracket the determinately better off than average contingent moral patients and determinately worse off than average contingent moral patients out. We do get to bracket out those where the comparison is indeterminate, by overlapping in range of welfare with the average. I’d guess it’s possible to specify some that are far enough away from the average, but this could be wrong.
2. on top-down bracketing, whether they get bracketed out probably depends a lot more on the specifics.
There’s also a question of whether we should handle counterparts iteratively, and then if order matters. If we bracket out Alice in A with her statistical counterparts in B, then it doesn’t seem like Alice should be left as an available counterpart to those in B (after pulling out one statistically average contingent moral patient from B), so Alice shouldn’t be part of the statistical average for counterparts of those left in B. So the statistical average of remaining moral patients in A needs to be recomputed. But if we bracket out the contingent moral patient Alex in A first instead of Alice, could this change the verdict of the comparison?

Michael St Jules 🔸 17 Aug 2026 16:48 UTC
4 points
0 ∶ 0
on: Challenging the unawareness argument with third actions and mixed actions
Very interesting!
How do you recommend choosing between somewhere-uniquely-optimal actions if there are multiple? I’m imagining the worst case where we still have several left that we’re clueless between.^[1]
And are all convex combinations of (lotteries over) somewhere-uniquely-optimal actions also somewhere-uniquely-optimal? (Let’s suppose our set of actions is closed under convex combinations, so somewhere-unique-optimality is defined with respect to that set.)
EDIT: The answer is no. 60-For-Sure, 100-If-True and 100-If-False are all somewhere-uniquely-optimal (under convex combinations), but 60-For-Sure dominates the convex combination of picking one of 100-If-True or 100-If-False at random, with 50% probability each.
1. ^
  I suppose the actual worst case is where inaction is ~always somewhere uniquely-optimal, and especially when building a portfolio to give across organizations, but that will depend on the particulars involved.

Michael St Jules 🔸 17 Aug 2026 16:41 UTC
2 points
0 ∶ 0
on: Challenging the unawareness argument with third actions and mixed actions
Just to clarify the example and concept: if we added the option c-For-Sure, which gives c with 100% probability, to the set of options in your example, it would be somewhere optimal iff c is at least 50, right?
Take Pr(X)=50%. Then 50-For-Sure, 100-If-True and 100-If-False all have EV 50.
And further, any lottery/gamble between any three of them (50-For-Sure, 100-If-True and 100-If-False) has EV 50, if the draw of the option is statistically independent from X’s probability distribution.

Michael St Jules 🔸 9 Aug 2026 3:25 UTC
4 points
3 ∶ 0
in reply to: huw’s comment on: You can now afford to work at AIM: our new salary policy, program stipends, and founder salary advice
Bigger seed grants!

Michael St Jules 🔸 28 Jul 2026 16:00 UTC
6 points
0 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
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 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.
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.
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?)?
Ghosts could be conscious, experience pleasure and suffering and care about what you do. 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). Others could want to see the happiness of loved ones. Others could want their descendants to live up to their expectations (e.g. in education, work, family, religious adherence), and not care much about their happiness. A large share could be horrified by modern secularism. It could be that every human that dies becomes a ghost indefinitely.

Michael St Jules 🔸 27 Jul 2026 17:50 UTC
6 points
2 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
What do you think about the St Petersburg problem now?
I think your arguments re infinities conflict with Occam’s 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. There are infinitely many ways the universe could be more complex, and arbitrarily more complex, than you’d guess.
A flat universe effectively has one fewer parameter and is simpler. So it shouldn’t get 0 credence. Among flat universes, the bounded/finite ones also have extra parameters for the boundaries or shape of the universe (compared to something that looks like R^3), so the infinite one shouldn’t get 0 credence.
(There probably are many ways for the universe to be infinite spatially in its shape, too, but those are more complex than R^3.)
EDIT:
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.
Note that I never suggested to assign 0 probability to anything. I think this leads to further examples to illustrate my point:
Would you assign exactly 0 probability to photons having exactly 0 mass?
Exactly 0 probability to there being no additional fundamental force, because it could just be vanishingly weak? And shouldn’t this get you to infinitely many fundamental forces? For any finite set of fundamental forces, you could posit another one and just say it’s very weak.
Exactly 0 probability to there not being ghosts, because their effects could just be very small or rare?
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 Occam’s razor.

Michael St Jules 🔸 27 Jul 2026 4:05 UTC
6 points
0 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
I like the principle of indifference. However, I think infinities are unfalsifiable in principle. In this case, does it make sense for me to attribute probabilities to them? If I did, they would be just metaphysical priors that can never be updated by any evidence.
St Petersburg-like lotteries, defined in terms of your Bayesian credences, don’t require assigning positive probability to any possible infinities out there in the world.
I’ll leave the rest in a footnote, because it’s not that relevant to the point I’ve been making so don’t plan to go further with it, but I already wrote it and it may be of interest to you.^[1]
1. ^
  I think falsifiability for a Bayesian could just mean we can imagine evidence that would warrant assigning very very low credences to the hypothesis. 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.
  In the case of the spatial extent of the universe being infinite, it seems hard to falsify now only because all the evidence we’ve gathered so far is in fact consistent with it (or no less consistent than with a bounded/finite universe), and it remains one of the standard models used by experts. My own view is that it’s more likely than not infinite in spatial extent, because it’s the simplest model consistent with the evidence.
  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. If we had good evidence of nonzero global curvature, that would be decent (but not overwhelming) evidence that the universe is finite/bounded, because it would rule out the most plausible infinite models of the universe.
  I agree that we can sometimes compare actions with infinite expected cost-effectiveness. I’m most partial to expansionist views of some kind, as the most complete (given precise probabilities).

Michael St Jules 🔸 23 Jul 2026 16:57 UTC
4 points
0 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
I object to relying on infinities (not arbitrarily large finities) to guide decisions because they do not explain more empirical evidence than arbitrarily large finities.
But do they deserve to be privileged at the exclusion of the infinite EV distributions if and when the latter are at least as consistent with the evidence? Why? Shouldn’t you use some principle of indifference or symmetry here?
If there’s 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?

Michael St Jules 🔸 22 Jul 2026 18:25 UTC
4 points
0 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
The expected value of the St. Petersburg lottery is 1 + 1 + … = +inf. It involves finite terms, but infinitely many terms. I meant to relate f(x) = x in my comment to the expected value of the St. Petersburg lottery. If this involved an arbitrarily large number of terms, its expected value would be arbitrarily large, but not infinite.
I don’t understand what you’re objecting to. Are you saying all probability distributions should only take finitely many possible values? If so, do you have an upper bound on what the total value should be over all of the future and possible existence, that you’re 100% confident in, under your course of action if where you do nothing? (It doesn’t have to be the least upper bound / tight.)

Michael St Jules 🔸 12 Jul 2026 17:29 UTC
2 points
1 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
Below is how Claude thinks Adam (the author of the article) would object to your comments. The objections make sense to me. Any reactions?
Claude is dumb (at least without further critique and verification, and usually with), and your prompt basically put it on the task of defending the position, not actually assessing the arguments fairly. So it turned up bad arguments.
I doubt the author would respond this badly.

Michael St Jules 🔸 12 Jul 2026 16:06 UTC
6 points
0 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
Taking bet A doesn’t require any commitment. My argument just uses backward induction (+ignoring statewise incomparability), which you should generally use in sequential choice situations, or else you’ll be worse off in many situations, even with sharp probabilities.
It allows unsharpness. Having unsharp probabilities does not require sequential decisions to be made independently.
that very concession is what Elga reads as victory: the imprecision has stopped doing the one thing it was introduced to do.
The argument against unsharp probabilities is defeated. We just have to treat them in certain ways. The summary of the paper here missed one way we could treat them, and claimed too much against another (if we accept commitments or resolute choice in other cases).

Michael St Jules 🔸 12 Jul 2026 16:03 UTC
4 points
0 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
Solution to which problem? I am not sure what is supposed to be problematic.
That if you use backward induction on acting rationally at each step, you will be worse off. You will predict later that you’ll change your mind, unless you can force your future self to honor a commitment (or plan) you’d no longer want to keep when it actually comes time to honor it.
EDIT: my bad, the problem is that if you don’t use commitments, you could be worse off. Using backward induction in the Sally argument actually works fine, doesn’t leave you (or Sally) worse off and doesn’t require any commitment.
However, I do not think it can be infinite. The function f(x) = x can take an arbitrarily large value, but not an infinite value (its range is the set of real numbers).
St Petersburg doesn’t require any state to have infinite value. Its value is (canonically) 2^n with probability 1/2^n for each n at least 1. Always finite actual value, but infinite expected value.

Michael St Jules 🔸 11 Jul 2026 19:27 UTC
6 points
0 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Subjective Probabilities should be Sharp
Why does Parfit’s hitchhiker pose a problem?
Because the same kind of solution is available to someone with unsharp probabilities in Elga’s scenario, if you’re treating them fairly.
The St. Petersburg paradox involves an infinite expected payoff, and I reject infinite worlds.
It doesn’t require an infinite world, only that you can’t be 100% confident in any finite upper bound on your impact that you specify, and that there are infinitely many ways that the world could be (due largely to not full certainty about physics).
(But also 0% to infinite worlds seems epistemically immodest, doesn’t treat the evidence on each side fairly, and is poorly argued, imo. But I don’t want to rehash this.)
In Sally’s case, her money is the only thing that matters.
Why can’t the fact that she’d pick a dominated sequence or regret it if she rejects both bets matter to her after rejecting bet A?
I understand one should accept bet A based on that strategy. However, unsharp probabilities are supposed to allow for accepting or rejecting A?
They don’t have to in every case. If it was A in isolation, and no other decisions, then yes, both rejecting and accepting should be permissible. But that’s not the case presented to us.

Michael St Jules 🔸 11 Jul 2026 9:08 UTC
7 points
0 ∶ 0
in reply to: Evan LaForge’s comment on: Subjective Probabilities should be Sharp
As a rational actor with no useful information
mood

Michael St Jules 🔸 11 Jul 2026 9:06 UTC
7 points
0 ∶ 0
on: Subjective Probabilities should be Sharp
3. SEQUENCE — Sequences of actions can be assessed for rationality independently of their parts: each of “reject A” and “reject B” can be individually permissible while the sequence “reject-A-then-reject-B” is impermissible. Elga turns the same Sally argument on it. SEQUENCE makes rejecting Bet B fine when no Bet A preceded it but irrational when it would complete the bad sequence—so it, too, imposes different requirements across two situations Sally can see are identical in everything she cares about. Hence SEQUENCE fails.
I don’t think this is a strong argument. There are other cases where you should make commitments that you would later be inclined to break, like Parfit’s hitchhiker, and St. Petersburg lotteries with unbounded utility functions. The latter is an argument that unbounded utility functions are irrational, based on similar logic.
Furthermore, “imposes different requirements across two situations Sally can see are identical in everything she cares about”. What if I do care about the differences? Or, is this any worse than picking numbers to ensure precision for no better reason than that they occured to you? Because that’s what it takes to produce arbitrarily precise probabilities if you fix what information is available to you in realistic settings.
Also, here’s another way someone with unsharp probabilities might handle this situation. In summary, I should accept bet A at the start to rule out the possibility of picking a dominated sequence:
1. If I accept bet A at the start, then the probability that I pick the dominated sequence (rejecting both) is 0.
2. If I reject bet A at the start and if I can’t guarantee that I will accept bet B next, then there’s some chance that I pick the dominated sequence.
If I compare 1 and 2 statewise, then 1 > 2 with some probability, and 1 and 2 are incomparable otherwise. In other words, either 1 beats 2, or I have no decisive reasons favouring either and I can ignore those cases. So I decide on the cases where 1 beats 2 and accept bet A at the start.
What links here?

Michael St Jules 🔸 3 Jul 2026 21:50 UTC
2 points
0 ∶ 0
on: The Learning Trap: What Simulated Clueless Agents Reveal About the Unawareness Argument
What would you recommend to handle cluelessness and unawareness about the far future and acausal influence, where contact and feedback may not be very practical or sufficiently informative? Maybe with the far future, we should just be extremely patient, possibly over 1000 years or more, and wait for that feedback anyway, but during that time, lots of confounders and clueless-making events could come up.
Or we simulate things in a lot of detail to try to get that feedback artificially, although that may mean simulating and realizing a lot of suffering.

Michael St Jules 🔸 3 Jul 2026 20:55 UTC
3 points
0 ∶ 0
in reply to: Anthony DiGiovanni 🔸’s comment on: Announcing: Cluelessness Critiques Competition
(1) not “impartial” in the sense that some moral patients/consequences are bracketed out for not very well-motivated reasons, or (2) not action-guiding.
If you think it’s fair to map the “same” moral patient across worlds to themselves, assuming at least some transworld identity, then I’d guess we could work out counterpart relations extending transworld identity to cover contingent moral patients, maybe along the lines of Thomas, 2019, pp.30-31 where, roughly, we map each contingent moral patient under one choice to the statistical average of the contingent moral patients — across all of spacetime and the whole multiverse if we’re in one — under the other choice.^[1]
At first pass, if we’re clueless about the total welfare effects for any action or comparing any two actions, I’d guess we’d be clueless about the average contingent moral patient, and, at least on bottom-up bracketing, just bracket them out and ignore them. This would mean the theory acts like the strict person-affecting view, where only currently existing (or at least “necessary”) moral patients count, and we ignore (at least) the rest. That might sound bad, but we ignore them because we’re consistently clueless about whether their counterparts are better or worse off in expectation, which seems like a good principled reason to do that.
This could be practically action-guiding. You’d probably focus on near-term human welfare, because currently existing non-human animals that you could “affect” cost-effectively don’t typically live long enough for your actions today to help them (e.g. most live <2 years before slaughter, often <1 year), and you’re clueless about the welfare of the counterparts of the future/contingent animals you could “affect” cost-effectively.
But maybe speculative scenarios, including how TAI might play out, also make you clueless about how to benefit currently existing humans if, before bracketing, you’re aggregating and taking expectations over their entire futures or their current preferences pointing at things at least a few years away, and can’t bracket out the speculative scenarios. I imagine you could get around this by using transworld identity at each point of time^[2] and do bracketing before aggregating over their futures, so you can bracket out the periods of time you’re clueless about.
1. ^
  Maybe with some extra accounting for the number of moral patients to do. And this could get trickier with infinities.
2. ^
  Could need to account for relativity theory in some way.

Michael St Jules 🔸 3 Jul 2026 18:04 UTC
4 points
0 ∶ 0
in reply to: Anthony DiGiovanni 🔸’s comment on: Announcing: Cluelessness Critiques Competition
Would you say bracketing isn’t fully impartial? It doesn’t seem less impartial than Scanlon’s Greater Burden Principle (or Tom Regan’s harm principle), which roughly says to choose between two actions, you should prioritize the individual(s) with the strongest claim or who has the most to lose counterfactually between options (not necessarily the worst off, to contrast with Rawls’ Difference Principle/maximin/leximin).
I guess methods of bracketing where admissible bracketings must group certain individuals, e.g. spatiotemporally or by causal paths affecting them, are more partial. But if you go with the unconstrained version, it seems basically impartial?
It is sensitive to differences between worlds beyond the number of individuals at each welfare level, because it has to identify each individual across worlds to take individual differences.

Michael St Jules 🔸 15 Jun 2026 19:12 UTC
11 points
2 ∶ 0
in reply to: Vasco Grilo🔸’s comment on: Only mammals and birds are sentient, according to neuroscientist Nick Humphrey’s theory of consciousness, recently explained in “Sentience: The invention of consciousness”
This also applies to many comments written by people? In general, whenever something is shared, one could invest more time to make it more accurate. However, to me it seems hard to come up with general rules about how much time to invest vetting claims.
If someone had a pattern of fabrication and very poor understanding (and apparent confidence) like LLMs often do if used uncritically, I would be annoyed with them and possibly do any of the following:
1. Tell them to read and review more carefully, look for opposing arguments, etc..
2. Downvote such comments (and I very very rarely downvote).
3. Stop engaging with this person, because it wastes my time and may encourage them to waste others’ time.
I think sharing long texts produced by LLMs is often fine even if they were not fully read by the person prompting, basically for the same reasons that it is fine to share long text produced by people even if the person sharing them did not fully read them.
I would request you check it yourself or at least run LLM critique passes, because of high rates of hallucination and other errors by LLMs.
Similarly, if there was an author who was just as consistently bad as LLMs are, and you shared their work uncritically like this, I’d recommend the same.
I have the impression many people, including academics, use “consciousness” and “phenomenal consciousness” interchangeably, and without wanting to take a stance on illusionism or realism.
My impression is that many people, including academics, use “consciousness” and “phenomenal consciousness” interchangeably, but they do so implicitly rejecting strong illusionism, would reject strong illusionism if asked directly, and typically don’t understand strong illusionism. Maybe many are open, though, I’m not sure.