Vasco Grilo

• Vasco Grilo 28 Jul 2026 15:51 UTC

  1 point

  0

  in reply to: MichaelStJules’s comment on: an­ti­monyan­thony’s Shortform

  Consequentialism should guarantee the best (expected) outcomes for the consequentialist. That’s the whole point. If Hammond’s version doesn’t do that, then it seems like a bad version.

  I think being left in the desert is the best and only outcome if the driver only takes me to town if they are certain I will pay them there.

  Let’s assume the driver will drive you if they predict you’re at least 80% likely to pay. Maybe you can’t fully guarantee a ride, but you can increase the probability by at least temporarily modifying how you make decisions, e.g. by caring more about keeping promises.

  I do not understand why this poses a problem to consequentialist reasoning. If I want to maximise my chances of survival, I should just constrain myself as much as possible to pay the driver in town. I would most likely be wrong if I found myself in town thinking there is only 10 % chance of paying if the driver is calibrated. I could try not to pay, but I would end up paying 80 % of the time. Reliably not paying would not be a live option if the driver is calibrated?

• Vasco Grilo 27 Jul 2026 8:38 UTC

  1 point

  0

  in reply to: MichaelStJules’s comment on: an­ti­monyan­thony’s Shortform

  Yes, you can’t not pay if you got a ride, assuming perfect accuracy. But this doesn’t solve the problem (guarantee a ride to town, which is presumably better).

  Consequentialism (as defined by Hammond) does not guarantee me a ride to town, but I see this as a problem for me, not consequentialism. Why should one expect consequentialism to save the person in the desert? In general, consequentialism recommends actions compatible with bad outcomes for some people, and good outcomes for others.

  For me the upshot of the thought experiment is that I am basically certain to stay in the desert. I will never be able to commit 100 % to paying in town. The simple fact that I would have to think about commiting is a sign that the chance of not paying would be above 0. So a driver who only drove me to town if they were certain I would pay would certainly leave me in the desert. So, regardless of how much I commit, I think I would be certain to end up in the desert.

• Vasco Grilo 23 Jul 2026 19:55 UTC

  1 point

  0

  in reply to: MichaelStJules’s comment on: an­ti­monyan­thony’s Shortform

  This [why I (Vasco) believe in dynamic consistency] seems completely consistent with the backward induction argument I made in the second half of my comment here.

  I agree. Claude says your argument breaks what I referred to as consequentialism. “Consequentialist behaviour, by definition, reveals a consequence choice function independent of the structure of the decision tree [consequentialism]”. Here is what Claude said.

  The structure of the tree must be irrelevant [for consequentialism to hold] — only the set of achievable consequences at the root matters. (And Hammond proves in §5 that this forces the revealed preference relation to be a complete, transitive ordering.) The way to show Michael violates this is to build two trees with the identical feasible consequence set and show his rule behaves differently in them.

  Tree 1 — the one-shot (normal-form) version: a single choice among the four consequences directly. Apply maximality to the static menu. Only NEITHER is dominated (BOTH beats it statewise); every other pair crosses. So the acceptable set is {BOTH, A-only, B-only}.

  

  Tree 2 — the sequential version we’ve been working with, which has the same four consequences reachable at the root (accept-accept → BOTH, accept-reject → A-only, reject-accept → B-only, reject-reject → NEITHER), so F [the set of feasible consequences] is identical. But Michael’s backward induction pruned reject-A. Reaching B-only requires rejecting A, so with reject-A gone, B-only is now unreachable. The acceptable set collapses to {BOTH, A-only}.

  

  Now put the two verdicts against the shared feasible set. Same F, different Φ [set of consequences from following a given behaviour] — and the culprit is B-only, admissible in the one-shot menu but excluded once the same consequences are reached through sequential structure.

  

  ---

  Conditional on being in town, you wouldn’t pay, because this is lower EV.

  If the driver is able to predict what I will do in town with perfect accuracy, and they will only drive me to town if I pay, then I will pay in town. Not paying would go against the perfect accuracy of the driver, and therefore the thought experiment would not be coherent, as the driver cannot have and not have perfect accuracy? As far as I understand, if I was driven to town, paying would be the only option I would consider.

  Are you saying you believed in sharp credences primarily because of Elga’s arguments?

  I already believed in sharp expected values before knowing about Elga’s paper. My intuition has been that it makes a lot of sense to represent uncertainty with distributions, but that only a single distribution is right in principle, even if I believe many different ones are reasonable in some sense.

• Vasco Grilo 23 Jul 2026 15:47 UTC

  1 point

  0

  in reply to: MichaelStJules’s comment on: an­ti­monyan­thony’s Shortform

  Why do you believe behaviour should be consistent in subtrees? Why can’t just the fact that the trees are different be enough to permit (but not necessarily require) different behaviour? I think this premise is basically assuming away our responses, so it’s doing a lot of work.

  I just have a strong intuition that what matters are the consequences of my actions given the current state of the world regardless of how I arrived at the current state.

  And what do the premises together imply for Parfit’s hitchhiker and this decision problem with St Petersburg lotteries?

  I think my thoughts are compatible with the premises.

  Why did you believe in sharp credences when you linkposted Elga’s paper? Do you no longer endorse those reasons as important for your beliefs?

  Elga’s paper supports my belief in sharp expected values (EVs), including EVs of distributions describing probabilities. However, my belief in completeness is supported more by Hammond (1988), and I understand lack of completeness matters more than unsharp EVs for Anthony’s unawareness argument for “no impartial altruistic justification for preferring any action over another”.

  So when you’ve quoted Claude to us (including the summary post of Elga’s paper itself), say the answer “makes sense” and/​or ask for our reactions, you didn’t endorse any of what it says?

  If I say something like “Claude’s objections make sense to me”, or “I endorse Claude’s objections”, I mean it based on the conventional meaning of the words involved. However, the meaning of “I endorse X” depends on the context. All else equal, the endorsement is stronger if I have thought more about the topic, and have written about it more.

• Vasco Grilo 22 Jul 2026 16:18 UTC

  1 point

  0

  in reply to: MichaelStJules’s comment on: an­ti­monyan­thony’s Shortform

  Can you spell out these 3 premises explicitly? I’m not sure what you’re referring to.

  I had in mind these 3 sentences from the abstract of Hammond (1988).

  Terminal nodes have consequences in a given domain [unrestricted domain]. Behaviour is required to be consistent in subtrees [dynamic consistency]. Consequentialist behaviour, by definition, reveals a consequence choice function independent of the structure of the decision tree [consequentialism].

  I have not read the paper, but I understand the 3 premises above lead to completeness and independence as defined in the premises of Von Neumann–Morgenstern utility theorem.

  It implies that behaviour reveals a revealed preference ordering [“a complete, transitive, binary relation”] satisfying both the independence axiom and a novel form of surething principle.

  I wrote “a complete [sharp]” in my last comment. I do not think completeness alone implies sharp probabilities. One can collapse unsharp probabilities to sharp ones using something like the midpoint rule discussed in Adam Elga’s paper. However, assuming one rejects rules like this, completeness leads to sharp probabilities?

  And did you know about and endorse these 3 premises when you posted Elga’s paper on the EA Forum?

  I did not know about Hammond (1988) before linkposting Adam’s paper. I found it thanks to this comment from Simon on my linkpost. I would still have endorsed the premises if I had read about them before publishing the post.

  What exactly is the position you’re defending? Has it changed throughout these discussions (in this thread or on the EA Forun post)?

  I mostly want to defend completeness, which I understand is a key premise of Anthony’s unawareness argument. However, I have been defending sharp probabilities (and by this I have meant sharp probability distributions, which imply their expected values are sharp too).

  Are there any cases in these discussions where you think an argument you made or endorsed turned out wrong, or a counterargument was correct? And if so, how has that affected your credences in the position you’ve been defending?

  I do not feel like I have been making arguments. My recollection is that I have mostly been quoting Claude, and asking questions.

  I lightly endorsed Adam’s claim that “Sharp does not entail Uniqueness” before linkposting the paper. I did not think much about it before linkposting. A comment from Jim Buhler prompted me to think a bit more about it, and come to the view that “I endorse sharpness and uniqueness. As far as I can tell, the issues of unsharp probabilities would apply in the same way to non-unique probabilities”. There is more context about my views about uniqueness in my reply to Jim.

• Vasco Grilo 22 Jul 2026 14:44 UTC

  1 point

  0

  in reply to: MichaelStJules’s comment on: an­ti­monyan­thony’s Shortform

  After all of these discussions, in your own words, what are your reasons for believing in sharp credences (or whatever position you’re defending)?

  My answer was supposed to relate to this question, although I practically did not use my own words.

  Would replying to other questions be useful to you? If so, why? I may reply if it would be useful to you.

• Vasco Grilo 22 Jul 2026 14:04 UTC

  1 point

  0

  in reply to: MichaelStJules’s comment on: an­ti­monyan­thony’s Shortform

  Hi Michael. I believe in sharp credences because they follow from 3 premises I find very intuitive. From the abstract of Hammond (1988):

  Behaviour norms are considered for decision trees which allow both objective probabilities and uncertain states of the world with unknown probabilities. Terminal nodes have consequences in a given domain [unrestricted domain]. Behaviour is required to be consistent in subtrees [dynamic consistency]. Consequentialist behaviour, by definition, reveals a consequence choice function independent of the structure of the decision tree [consequentialism]. It implies that behaviour reveals a revealed preference ordering [“a complete [sharp], transitive, binary relation”] satisfying both the independence axiom and a novel form of surething principle.

• Vasco Grilo 13 Jul 2026 12:50 UTC

  1 point

  0

  in reply to: Anthony DiGiovanni’s comment on: an­ti­monyan­thony’s Shortform

  I agree with Claude’s 4th point.

• Vasco Grilo 13 Jul 2026 10:00 UTC

  1 point

  0

  in reply to: Anthony DiGiovanni’s comment on: an­ti­monyan­thony’s Shortform

  That makes sense. However, if one cannot make such a commitment, or finds its implications undesirable, Elsa’s argument should update one away from unsharp credences (even if one ends up preferring these all things considered)?

• Vasco Grilo 13 Jul 2026 8:04 UTC

  1 point

  0

  in reply to: Anthony DiGiovanni’s comment on: an­ti­monyan­thony’s Shortform

  “The imprecise credence itself doesn’t guide the action w.r.t. the second bet” does not imply “you shouldn’t have imprecise credences in general”. Elga’s argument doesn’t tell us at all that we should, say, “go with our best guess” about altruistic interventions.

  One’s best guess for the intervention with the highest expected marginal cost-effectiveness (EMCE) may be wrong. However, one should still support it under precise credences? Greater uncertainty about which intervention has the highest EMCE will tend to make interventions decreasing that uncertainty rank higher.

• Vasco Grilo 12 Jul 2026 14:21 UTC

  0 points

  0

  in reply to: Anthony DiGiovanni’s comment on: an­ti­monyan­thony’s Shortform

  I see. Below is how Claude thinks Adam would object. Any reactions?

  Elga’s strongest corrected replies

  [“corrected” because Claude initially said something that did not make sense. I pointed this out, and asked Claude to update what would be Adam’s objections. The updated objections below make sense to me.]

  1. The commitment doesn’t pay the original bill; it relocates it. Elga’s challenge was: how do unsharp credences themselves constrain rational action? Notice what C concedes. At the B-node the agent’s credences-plus-maximality still say rejecting B is permissible — that is why she has to lock the option out in advance. So the imprecise state still delivers the wrong local verdict; C works by overriding it. What constrains action, then, isn’t the unsharp credence at all — it’s a self-binding policy bolted on top, one that mimics a determinate dominance-avoiding disposition. So the credence is idle with respect to the very choice that saves the agent, and indeed must be countermanded. That is confirmation of Elga’s thesis (no acceptable account of how unsharp credences constrain action), not a refutation of it.

  2. Required or merely permitted? For C to rescue UNSHARP, reject-both has to become impermissible. But is enacting C rationally required? If it’s only permitted, then not-committing is also permitted, and DiGiovanni’s own horn 2 says that a non-committing agent who then rejects both has done nothing locally irrational — so the manifestly-bad outcome is still reachable by a permissible route, and UNSHARP still licenses it. So C must be required. But then the requirement has to come from somewhere, and it can’t come from the local maximality verdicts (which permit not-committing and permit rejecting B). It comes from a dominance-avoidance-over-policies norm — i.e., from evaluating the sequence ex ante. That is just the policy-level /​ SEQUENCE evaluation. C’s distinctive contribution is only that, by physically removing the later choice, it dodges the identical-choices objection that Elga levelled at policy-level evaluation. So the commitment isn’t an independent third option; it’s ex-ante maximality plus a lock-out whose sole job is to keep Sally from getting a grip.

  3. And that lock-out is itself the concession. Here is the line that doesn’t depend on any of my earlier missteps. Elga’s target is UNSHARP: it is consistent with perfect rationality to have unsharp credences. DiGiovanni’s defense requires that the perfectly rational unsharp agent be a resolute self-binder who must delete options from her own future choice set to avoid a foreseeable dominated outcome. But an agent who has to handcuff herself against what her own beliefs-plus-decision-rule count as permissible is displaying a belief state so defective that it needs handcuffs — and a sharp agent needs none. The need for C is a symptom of the defect, not a vindication of its compatibility with perfect rationality. DiGiovanni’s footnote reply — that adopting sharp credences is merely “more psychologically tractable,” and resolute choice achieves the same thing — misses this: the point isn’t tractability, it’s that the resolute imprecise agent’s good behavior is entirely explained by the commitment (which behaves like a determinate disposition), while the imprecision contributes nothing positive a sharp state couldn’t and must be overridden at the crux. A state that must be overridden to avoid disaster, and adds nothing a sharp state lacks, has no claim to being what ideal rationality delivers.

  4. The “you’re just coercing the future self too” charge fails on a disanalogy. DiGiovanni’s sharpest jab is that Elga’s own prescription — “adopt sharp credences” — is itself a way of making your future self’s beliefs deviate from what they’d count as locally rational, so it’s no more legitimate than demanding C. But there’s an asymmetry. “Adopt sharp credences” is not an instruction to the B-node self to act against its own current verdict; it’s a claim about which belief state to occupy. The sharp agent at the B-node does not override anything — her local verdict, given her precise p, already tells her what to do, and it already coheres with her A-node verdict, automatically, with no binding. The imprecise agent’s trouble is that her local verdicts are not automatically time-coherent, so she must either coerce herself (commit) or behave incoherently across time. Elga’s SHARP diagnoses the fault as lying in a belief state whose local verdicts fail to cohere, and the fix is precision, not coercion. So “adopt sharp credences” means “have a state that doesn’t need forcing” — which is the opposite of forcing.

  The honest bottom line

  Once the strict-rules horn is off the table, Elga doesn’t have a clean knockdown of the committed imprecise agent. What he has is (2) and (3): the demand that C be required rather than permitted, and the charge that the very need to self-bind is a mark against perfect rationality. DiGiovanni can keep resisting by insisting that “imprecise credences + resolute choice” is simply a coherent rival package to “sharp credences,” and that the Dutch book indicts only the third combination — imprecision plus inability to commit plus act-by-act choice. At that point you’re back at the bedrock intuition Elga admits he can only pump: whether a fully-informed, money-only agent who ends up foreseeably dominated has thereby shown a rational defect, or has merely done something a differently-organized agent could have avoided. This is exactly the resolute/​sophisticated/​myopic fork from the money-pump literature you already know from Gustafsson — and DiGiovanni’s own note that his argument “has a very similar structure” to money-pump arguments is him planting his flag on the resolute side.

  So the corrected verdict is narrower than what I gave before: the commitment move genuinely escapes both the Sally objection and the strict-rules horn, and Elga’s live reply is not “that collapses into sharp betting” (it doesn’t) but “requiring self-binding to avoid dominance is itself incompatible with the perfect rationality UNSHARP asserts — and the credence is doing no work in the rescue anyway.”

• Vasco Grilo 11 Jul 2026 13:57 UTC

  3 points

  0

  in reply to: Anthony DiGiovanni’s comment on: an­ti­monyan­thony’s Shortform

  Hi Anthony. Here is a summary of Adam Elga’s argument. Below is an outline of Adam’s argument against the plan strategy. Which part of it do you reject?

  2. PLAN — When you act, you simultaneously form a plan binding your later choices to cohere with it (reject Bet A → plan to accept Bet B → follow through), but without changing any beliefs. Elga refutes this with the case of Sally, who cares only about money and has a highly unsharp credence about rain. Compare two scenarios: in the first she rejected Bet A and is now offered Bet B; in the second she’s offered Bet B alone. PLAN permits rejecting Bet B in the second but not the first. Yet the monetary consequences of accepting and of rejecting Bet B are identical across the two scenarios, and her beliefs are identical, and money is all she cares about—so the situations are alike in every respect she cares about. Rationality can’t impose different requirements on choices that are identical in all relevant respects. To the rejoinder “but rejecting Bet B would break her plan,” Elga replies that either plan-breaking is something Sally finds costly (contradicting the stipulation that it’s costless for her), or it isn’t—in which case “Don’t break plans!” is as groundless a constraint as “Don’t break mirrors!” He flags but sets aside the resolute-choice tradition (Gauthier, McClennen) that would defend plan-following.

• Vasco Grilo 3 Jul 2026 10:02 UTC

  1 point

  0

  in reply to: habryka’s comment on: Don’t Eat Honey

  Hi Oliver. Do you have any thoughts on how to decrease uncertainty about welfare comparisons across species? I would agree it is super large. For sentience-adjusted welfare ranges proportional to “individual number of neurons”^“exponent”, and “exponent” from 0 to 2, which covers the best guesses that I consider reasonable, the sentience-adjusted welfare range of shrimps is 10^-12 to 1 times that of humans, whereas it is tentatively 8.0 % in Bob Fischer’s book about comparing welfare across species.

• Vasco Grilo 23 May 2026 19:13 UTC

  1 point

  0

  on: You can only build safe ASI if ASI is globally banned

  Hi Connor. I am open to bets against short transformative AI timelines, or what they supposedly imply, up to 10 k$. Do you see any that we could make that is good for both of us under our own views?

• Vasco Grilo 23 May 2026 19:05 UTC

  1 point

  0

  on: How quick and big would a soft­ware in­tel­li­gence ex­plo­sion be?

  Hi Toms. I am open to bets against short transformative AI timelines, or what they supposedly imply, up to 10 k$. Do you see any that we could make that is good for both of us under our own views?

• Vasco Grilo 23 May 2026 19:04 UTC

  1 point

  0

  on: Dave Kas­ten’s AGI-by-2027 vignette

  Hi Dave. I am open to bets against short transformative AI timelines, or what they supposedly imply, up to 10 k$. Do you see any that we could make that is good for both of us under our own views?

• Vasco Grilo 5 Jan 2026 18:03 UTC

  1 point

  0

  on: Ap­pendix: Ham­ming Questions

  Thanks for the great questions!

  Though somewhat annoying, these questions are also extremely helpful for focusing one’s attention on problems that are worthy of one’s efforts. (See Hamming’s talk “You and Your Research” for more.)

  Here is the talk. I liked it!

• Vasco Grilo 31 Dec 2025 15:15 UTC

  1 point

  0

  in reply to: Steven Byrnes’s comment on: Four ways learn­ing Econ makes peo­ple dumber re: fu­ture AI

  And I can’t bet on “what they supposedly imply” because if my beliefs about AGI are right, then I expect we’ll both be much too dead for a bet to pay out in my favor.

  I could pay you in advance as I did with Greg Colbourn, although this would only work for relatively short timelines. Otherwise, I had better keep my money invested.

• Vasco Grilo 31 Dec 2025 14:09 UTC

  1 point

  0

  on: Four ways learn­ing Econ makes peo­ple dumber re: fu­ture AI

  Hi Steven. Would you like to bet on short AI timelines, or what they supposedly imply?

• Vasco Grilo 1 Oct 2025 13:52 UTC

  1 point

  0

  on: Es­ti­mat­ing the Cur­rent and Fu­ture Num­ber of AI Safety Researchers

  Hi Stephen.

  To create the estimates in the tables below I used the following sources (ordered from most to least reliable):

  • Web pages listing all the researchers working at an organization.

  • Asking people who work at the organization.

  • Scraping publications and posts from sites including The Alignment Forum, DeepMind and OpenAI and analyzing the data ^[2].

  • LinkedIn insights to estimate the number of employees in an organization.

  I find it interesting that you ranked LinkedIn last. Is this because many people working at the target organisations do not add their roles or organisations to their LinkedIn profiles?