As someone who hires people (I am hiring right now, see bottom of the post), I often get emails asking whether someone should apply for a role I've listed.
If you're considering contacting a hiring manager to ask if you're a good fit for the role, you should just apply.
Why?
Notable exceptions, where you should email:
To be clear, the urge to send an email to check if you should apply is totally normal (I have done it before), and anyone who has done this has nothing to worry about. However, I would strongly recommend applying despite your reservations. You can just do things!
Also, if you want to apply to become Access to Medicines Initiative's Head of Monitoring, Evaluation, and Data, JOB DESCRIPTION AND APPLICATION HERE: https://docs.google.com/document/d/1nqOz2a5bTQs5W361G5gpJ3brLEf6xuG0
⬆️ The future we are all dreaming of!
(I hope sharing of what is essentially a meme is allowed here)
Here is a claim I made to my friend yesterday that I think is true and believe people continually underrate: given reasonable assumptions about the influx of incoming capital, getting into (ie working in/thinking about) AI/ AI Safety early has probably been better for GHD or AW than lots (maybe all, tho I haven’t crunched the numbers) of direct work in GHD and AW.
I think this could even be true for EA CB (with much less confidence). AIS groups grow a lot faster than EA groups historically do and while the groups are often not that EA aligned, the leaders (who are, in my view, often very talented) are to various degrees. This is especially true for the ones who end up working full time in AIS: they often have proto-EA dispositions and these become intensified when they work around a bunch of EAs full time (though I can see two ways in which this is not as good as normal EAs: (1) being EA for social reasons and (2) being EA in stated but not revealed preference). TBC, I think having an EA group at a school make it much more likely that the leaders of the AIS group are aligned.
I can also see worlds (though def not my mainline at the moment) where the best thing one could have done for quality adjusted EA CB is actually just supporting AIS groups.
I don’t know how we should update/ by how much, but I think a reasonable way is the following: being ahead on some big emerging issue might just be more important than what seems like the obviously important thing that nobody does for bad reasons (ie GHD or AW).
After the initial experiment period, we've decided to keep the featured page.
One of the key graphs in making this decision is below:
Now that the featured page is likely to stick around long term, I'd like to start collecting ideas for V2[1]. Do you use the page? Do you pointedly not use the page? I'd love to get any and all feedback, in the comments here, or via this form.
Please share small things like aesthetic tweaks as well as larger issues such as a disagreement on the kind of posts that get featured, or how featuring is done. Happy to answer questions too.
I.e., the current version is a prototype, and I'm ready to do a full redesign if necessary. Because of some protracted OOO I might have to take soon - V2 might land near the end of September, or even early October.
What if you ran an effective altruism group where, every week, people congregated to contemplate topics such as:
What if your group satisfied a deep need that many young people feel for an impartial, calm, and evidence-oriented environment to expl...
I recently hosted a conversation between Imelda and Jack on the implications of the third wave of philanthropy on global health financing in Africa. Imelda is the founder of Philanthropy Reform Alliance, a nonprofit that examines philanthropy through the lens of power and stewardship. @Jack Lewars, as we all know, is the founder of Ultra Philanthrop...
Yes this is the hard part. AI can make it much cheaper for a donor to assess an unfamiliar organisation once there is enough evidence to work with, but it cannot manufacture information that is not visible in the first place. And there is probably a second-order problem here: once AI starts doing more of the assessment, organisations that are easier for machines to identify, document and compare may get an advantage simply because they are more legible to the system.
I have been thinking about this as a kind of “machine-legibility privilege.” It is one of the things I am trying to test with zooidfund: whether AI assessment can actually broaden the set of needs donors can consider outside their existing networks, without simply shifting the advantage toward whoever leaves the best digital trail.
Recently, a friend asked me whether it seems like a good idea to advance AI's introspective qualities. In this exercise, I discovered how to elegantly compress most of my strongest views on AI consciousness research strategy.
Currently, the correct theory of consciousness is extremely uncertain. However, in the future, we are very likely to understand consciousness fully.
Consciousness is not just "one binary thing". Across people and situations, it varies...
Context: Effective Mental Health is a new group working to improve the field of global mental health along EA principles (and we're running another round of our Global Mental Health Fellowship, see info / apply here!).
Here, we outline the quick & basic case for why field-building work in effective global mental health seems especially cost-effective.&...
I'm not sure this is correct - a DALY weight is not a literal year of human life requiring sustenance. It's saying that a year with depression is like losing some proportion of healthy life. But the actual human years of life would remain the same (e.g., if the weight of depression is .5, then curing depression is not leading to twice as many human years actually lived and thus twice as much eating of meat)
This competition entry has been selected for publication by the Forum team.
Here are some comments on the Summary of the argument.
What would justify preferring action A o...
I'm frustrated that this answer got downvoted to zero.
- Your "q<p and Q<P" example is a case where we can compare imprecise "EVs".
We do not have to compare imprecise "EVs" and hence coarseness doesn't come into it. We are directly assessing the preferability of one over the other. In the two images below the "EV" of each option is completely undefined (or +-infinite)
This post will summarize the trajectory of Wild Animal Initiative’s field-building for wild animal welfare science so far, along with patterns and indicators we want to see more of going forward. Beginning with an initial phase focused on establishing credibility and awareness of the field, through a second phase that has placed more emphasis on the distinctiveness and broad priorities of the field, we are now moving into a third phase designed to co...
I think I’m getting lost in the intermediate outcomes. I understand there are signs that you’re succeeding at building the field. But I keep coming back to a basic question: what is the causal pathway from this work to animals actually being better off? I’m struggling to identify it beyond 'a stronger field may eventually discover interventions', and I find that difficult to evaluate against alternative uses of the funding.
I’m also not sure the examples you give make the case for building an entirely new academic field as strongly as you say. Avian botulism, wildlife contraception, fish welfare etc appear to already be established areas of research to varying degrees; what seems novel is applying a welfare-oriented lens. That makes me wonder why the right response is to build a new field at substantial cost, rather than fund targeted intervention development using knowledge that already exists in adjacent disciplines.
So no, the key question for me isn't how important I happen to think WAW is in general. It's 'how have animals benefitted so far, and if not at all yet, then on what sort of timeline can we expect concrete outcomes, and at what further cost?'
Epistemic status: Speculation from two decently informed advocates armed with anecdata.
Note on process: After having some version of this conversation several times and saying, “we should probably write about this publicly,” we took the less heroic route: we recorded one of our conversations, fed the transcript into an LLM, and then substantially revised the structure, substance, and framing ourselves. We will not be sharing the transcript, as it is in...
Thank you for writing this, Tom and Rocky. I think this is an important caution, and I appreciate the way you are framing it.
As someone directly involved in a welfare-tech-style intervention, I actually agree with much of your core argument. I also strongly believe that good tech interventions can be extremely impactful, partly for some of the reasons you outline: they can sometimes bypass certain barriers that exist for other welfare work, and they can make it possible to help very large numbers of animals relatively quickly. But I also agree that, when looked at from the outside, they can be falsely perceived as silver bullets, so I want to share three insights from SWP about electrical stunning for shrimp.
I’ve also posted a more direct response to “Animal Welfare Has an Evidence Problem,” which you refer to in the post, to clarify some of these points.
First, at SWP, our work on shrimp stunning has been much more complex than simply deploying stunners at scale. The Humane Slaughter Initiative does involve getting electrical stunning equipment used on farms, but in practice, we are doing much more than that: coordinating and funding welfare research and R&D, collecting field data, working with scientists, developing and refining slaughter protocols, iterating on the technology with existing equipment providers, negotiating with industry, engaging corporate stakeholders, building and maintaining producer relationships, training staff, educating farmers… Much of this work is not visible from the surface, but it’s central to making the intervention possible. The tech matters, but it does not operate in a vacuum.
Second, on shrimp stunning specifically, I think you captured something important. SWP entered this space aware that there were uncertainties, and we’ve been proactively trying to reduce them. Some of that work is public, but much of it happens through industry collaborations, field implementation, and producer relationships, which are often necessarily behind closed doors or covered by NDAs. That can make the work less visible from the outside, even though it is central to what we are doing.
This is one reason I think treating our intervention too simplistically, as merely putting stunners on farms, can lead to exactly the kind of disillusionment and backlash you describe.
Third, I feel some of the issues you describe are not limited to tech interventions. Many ambitious attempts to change existing systems seem to have an early phase where progress looks relatively fast, followed by a harder phase where problems become more complex, uncertainty can increase, coordination costs become bigger, and resistance from affected stakeholders becomes stronger. That does not necessarily mean the intervention – tech or otherwise – was a mistake; it can be a natural progression of an ambitious project.
Thanks again for writing this! It’s an important reminder that welfare tech can be genuinely promising and, in some cases, extremely impactful, but it should not be treated as a silver bullet.
I think the picture is (mostly) bleak if you look at the denominator (total animal suffering), and how little we can do to end suffering of all farmed animals. If you look at the numerator, there are millions and sometimes even billions of individuals whose lives you can help, even in expectation.
If you end up reading it I'd particularly value your feedback on how I'm striking this balance.
This is a linkpost for Is Extinction Risk Mitigation Uniquely Cost-Effective? Not in Standard Population Models by Gustav Alexandrie and Maya Eden, which was published on 18 August 2025 as chapter 20 of the book Essays on Longtermism: Present Action for the Distant...
Thanks for the discussion which prompted me to look into it.
This week is Cluelessness Critiques Week on the EA Forum.
We'll be publishing eligible entries[1] to the Cluelessness Critiques Essay Competition on the Foru...
Yeah, I think it's easy to find a counterexample to "unconditional cluelessness" (i.e., even in the box situation), but much harder to find one relevant to what you and I should do in our present non-simplified situations (which is presumably what Anthony meant for us to discuss).
Oops sorry for the GW-GD confusion, but yeah, this changes nothing to my point I think.
I agree that it is possible for them to occur, but there is no probability distribution I would include in my representor under which they are sufficiently likely to occur that the comparison between actions is indeterminate.
This seems hardly defensible.
- The average human being (including in poor countries) contributes to the farming of so many animals throughout their lifetime (at least in expectation). You would have to be astonishingly confident that the welfare of the animals we eat do not significantly matter for your above conclusion to follow.
- The number of far-future lives we indirectly influence might be astronomical such that long-term effects (almost) always dominate. I don't see on what basis you can exclude this possibility from your probability distributions, given the arguments given here and refs therein.
(I don't recall this specific example from Mogensen and don't have time to dive back into it, so I won't comment on that, sorry.)
Hi Jim. I agree trying to come up with counterexamples is useful. Below are 3 potential counterexamples I have given. @Anthony DiGiovanni 🔸 does not consider them counterexamples (see Anthony's replies for details).
1st example, which Elliot already quoted in this thread.
Consider these 2 options for what I could do tomorrow:
- Torturing my family, and friends, and then killing myself. I would never do this.
- Donating 100 $ to the Shrimp Welfare Project (SWP), which I estimate would be as good as averting 6.39 k (= 639/10*100) human-years of disabling pain.
My understanding is that you think it is "irreducibly indeterminate" which of the above is better to increase expected impartial welfare, whereas I believe the 2nd option is clearly better.
2nd example.
given any 2 objects, I believe my best guess should be that the expected mass of one is smaller, equal, or larger than that of the other.
3rd example.
Hi Anthony. Do you think the expected welfare of 2 states of the world which only differ infinitesimally can be incomparable? I do not see how this could be possible. For example, it feels super counterintuitive to me that, given 2 identical states, moving an electron by 10^-100 m in one of the states would make their expected welfare incomparable. I guess one can get from any state of the universe to another in an astronomical number of infinitesimal steps, and I believe any 2 states which only differ infinitesimally are comparable. So I conclude any 2 states are comparable too, even if it is very hard to compare them, to the point that I do not know if electrically stunning shrimps increases or decreases welfare in expectation.
Here is a 4th example. Consider these 2 actions:
I think we are justifyed in c-preferring the 2nd action. I believe Anthony disagrees
Under maximality, we can't even say that [1, 10 000] is better than [-10 000, 1.0001].
That's not quite right. Maximality says an action A is impermissible when some alternative B has higher EV on every probability function in your representor. And that can be true even when A's and B's EV ranges overlap.
Example:
If A had the same range but sloped the other way, then it would be permissible by maximality:
So to figure out what's permissible under maximality, we can't just look at ranges. We need to look at the representor.
Hi Amanda, I have a few points of clarification regarding the evidence on shrimp stunning and Shrimp Welfare Project’s Humane Slaughter Initiative.
We acknowledge it is true that academic peer-reviewed evidence is limited (although we’re excited to report that the Stirling study is no longer a pre-print and has been published in Aquaculture). That being said, our intervention is informed by more evidence than just the peer-reviewed literature. We (and, sometimes, the industry) have access to more evidence than what’s published or able to be communicated publicly.
It may be useful to clarify what we mean by “research” and “evidence-building” in this context. Beyond peer-reviewed studies, evidence-building for humane slaughter means reducing uncertainty across several linked questions, including: whether electrical stunning can quickly render shrimp insensible under controlled conditions; whether it works reliably during commercial harvests; what parameters, protocols, and equipment designs improve outcomes; whether producers are using stunners as intended; and how monitoring systems can make that implementation more verifiable over time.
In practice, that means the evidence base is built from multiple sources: lab studies, field observations, stunner data, producer feedback, behavioral indicators, equipment iteration, implementation protocols, and MEL systems. This is not a linear process where all the research happens first and implementation only begins later. Field implementation is part of the learning which often reveals the constraints, failure modes, and practical questions that the next round of research needs to answer.
Recently, The Center for Responsible Seafood (TCRS) published a summary of a study conducted by the University of Stirling[1]. In a field visit in April 2026, researchers compared electrical stunning and ice slurry on an Indian farm. An excerpt reads:
“In summary, to meet ethical and welfare concerns, shrimp benefit from rapid stunning immediately after harvest. Electric stunning provides a rapid and effective stun as shown by the lack of response to subsequent cold exposure, but the process triggers a strong muscle contraction and elevated blood lactate. Cold shock does not deliver an immediate stun.
The report found that, with ice slurry alone (referred to as cold shock or CS), “shrimp are highly likely to maintain neural sensibility during the CS stunning process,” and those animals “did incur an initial ca. 30 sec tail flip response (since they are not insensible).”
“Therefore,” the researchers wrote, “best practice would suggest that ES needs to be followed by CS (ES+CS protocol).”
For those unfamiliar with TCRS, it’s quite an industry-aligned organisation, so their study coming to this conclusion is noteworthy.
Through our Humane Slaughter Initiative, Shrimp Welfare Project’s team members have seen dozens of harvests using stunners, and we have collected thousands of photos and videos that have informed our intervention (e.g., which parameters work best in certain contexts, how the use of a pump/harvester could minimize time shrimps spent out of water, etc.). We can’t share most of that content publicly because our industry partners (and therefore, we) operate in a competitive market environment where confidentiality, trade secrets, and proprietary information are very important. We understand that many folks in the EA community would appreciate greater transparency, and we’re always trying to facilitate a productive exchange of knowledge about our work – hopefully, this comment will accomplish some of that.
To respond to some of your specific points:
In 2021, Tesco and Hilton Seafood published a report on their use of a modified Optimar stunner in Vietnam. … This gives some reassurance for this particular machine, although I know from talking to people in the field that this machine was a modified version of the commercially available one.
Shrimp Welfare Project has visited that farm and observed the stunner in operation, and it is an early (yet almost-standard) Optimar model. It's a fish stunner that had its electrified "fingers" (which, when they touch animals on an electrified conveyor, complete the electric current and produce the stunning shock) lowered to reach much smaller animals, rather than a completely bespoke modification that could have fundamentally changed the stunner's parameters or operation.
We've worked with Optimar to implement improvements over the years (we’re currently on version ~6), so the shrimp-specific finger-lowering adjustment is no longer a post-manufacturing modification, but a feature of the commercially-available standard model. Perhaps this is the source of the confusion?
Regardless, the important takeaway here is that the 2020 case study used a stunner that is representative of the kind of stunners we support via our HSI program at present, and it concluded that electrical stunning at the recommended parameters was able to “deliver >97% stunning efficacy,” and followed by “immediate immersion in ice slurry achieving 100%” efficacy.
Going into your next point about sample sizes, I wanted to note that this case study used 50kg samples per voltage tested. Since shrimps were 25-30g each at harvest, that’s ~1,700-2,000 animals per 50kg sample, though the report doesn’t state this explicitly. There were also 100 shrimps checked for stunning effectiveness at start-up.
[re: Weineck et al.] The study uses a very small sample size (N = 6 for each intervention) which makes me uncomfortable recommending any action based on it. … [re: Stirling study] This study is also small (N = 4-6 per intervention) which again leads me to have limited confidence in conclusions.
Many of the studies that involve animal testing like this use small sample sizes. In speaking with scientists who specialize in invertebrate biology, it’s understandable why these studies would be designed with small sample sizes. If the desired effect size is large for the variables tested (as is often the case here) and the standard deviation among individuals is only moderate (as is expected to be the case here), this is not necessarily a limitation. It’s statistically appropriate, and even often ethically mandated – by institutional standards and/or the Three R Principles of animal research. This is only to say: we should not have a knee-jerk response to sample size.
That said, we recognize that the actual findings are not a clean on/off, but rather a subtle, dose-dependent difference between responders and non-responders. Our confidence would increase if there were more animals tested to generate more data to support precise parameters for effective stunning. At the same time, the direction of the findings is still informative, especially as they converge with other evidence, including the TCRS study and our own field data.
Based on this data, it is unclear if electric shock followed by ice slurry provides any benefit over ice slurry alone, provided the animals are kept in ice slurry until they are fully dead. (It is unclear how long that would take, though.) … A sufficiently strong electrical shock with proper ice slurry (which is hard to implement in practice) does not provide much improvement over proper ice slurry alone.
I don’t think the Stirling paper supports this.
At 2.5–5 °C (which would be a high-performing slurry temperature by current commercial standards) cold shock was slow and unreliable as a route to neurological insensibility. Using the paper’s threshold of Ptot <10% of baseline brain activity, only 4 out of 5 shrimp reached this threshold, and this occurred after 28 minutes. One animal did not appear to reach the Ptot <10% threshold within the 30-minute observation period.
At −2.5 °C, all shrimp reached Ptot <10% within 4 minutes. However, achieving and maintaining a −2.5 °C slurry consistently during commercial harvests, especially in tropical climates and under high biomass loading, is likely to be extremely difficult. We are exploring this avenue, including deepchill-type technology, but at present we are only moderately optimistic about its practical feasibility. At the moment, maintaining these kinds of temperatures is far from “proper ice slurry,” and more appropriately understood as a rare exception in real-world production settings.
The best-supported approach in the paper is electrical stunning (ES) followed by cold shock (CS). In the effective ES + CS group, all animals that did not tail flip after cold exposure reached Ptot <10% after 3 minutes of cold shock. There’s limited research on the relationship between tail flipping and neural insensibility / loss of consciousness[2], so our confidence in this correlation would increase with more data. But it’s important to reiterate that, based on the best available knowledge from the Stirling lab (which they clearly state in the paper), absence of tail flipping after cold exposure is aligned with neurological insensibility[3].
Here is a video that shows shrimp in ice slurry – the first group has been electrically stunned, and the second group has not. The second clip shows a nearly ideal ice slurry in a production setting, with a temperature around 1 degree C. What you see in that video is consistent with what we’ve seen at shrimp farms around the world: that even when in a highly controlled ice slurry, shrimps flip their tails for multiple minutes.
Insufficient electrical stunning with proper ice slurry may be worse than ice slurry alone.
This would be true for basically all stunning methods, across land and aquatic animals. If a percussive stunning bolt is insufficiently administered to a cow’s head, then the slaughter process would be more painful than if a stunning method hadn’t been poorly attempted.
Electrical stunning without proper ice slurry slaughter poses real potential for causing harm.
A few things here:
Additionally, this line of reasoning would seem to contradict your argument in your post, that:
“Many shrimp harvests do not use ice slurry, or do not use it properly (not cold enough, not long enough, etc.).”
There’s a tension here worth flagging: on the one hand, your comparison of electrical stunning vs. ice slurry assumes an idealized version of “proper” ice slurry – cold enough, maintained long enough, carefully managed. But when it comes to what electrical stunning would look like in real-world harvest conditions, your concern raised is precisely that ice slurry is often not implemented properly. If poorly implemented ice slurry undermines the case for electrical stunning, then the same realism has to apply when ice slurry itself is proposed as the alternative. Either way, the fair comparison is between the two methods as they would actually be implemented – and in our view, that comparison favors electrical stunning followed by ice slurry as the kill method.
The science is clear that any stunning method can be reversible if it’s not followed by an adequate kill step. This is true for both electrical stunning as well as ice slurry that’s used for stunning. In the Humane Slaughter Initiative, our protocol calls for ice slurry as the kill step; importantly, we’re not using ice slurry as a stunning method, but rather a method to prevent the shrimps from recovering after the electric shock. We guide farm staff on how to do this ice slurry properly, including monitoring the temperature and replenishing the ice, etc. In our field experience, this process usually involves putting shrimps through the stunner, then directly into an ice bath for several minutes (as in this video and this video), and then into transport crates where they are packed with ice. In this slaughter method, the ice slurry prevents the shrimps from re-warming and recovering, rather than being the step that also stuns them.
Doing a well-implemented ice slurry for stunning is much harder to do effectively in practice than an ice slurry for prevention of recovery/slaughter after electrical stunning. So, although these practices share the same name, their different purposes and application create a major distinction. We are more confident in ice slurry as a slaughter and recovery prevention method after electrical stunning, than ice slurry as a stunning method itself.
Overall, the evidence base for shrimp stunning is still developing and Shrimp Welfare Project acknowledges this and is actively working to develop it further, both through our own fieldwork and by coordinating with researchers, manufacturers, producers, and other industry actors. Given the scale of the problem, we do not think the right response is to wait for full certainty, as it would mean accepting a harmful status quo for billions of animals while the evidence accumulates. The direction of the evidence, our field experience, and the practical realities of commercial harvests so far all point the same way: electrical stunning is faster, more consistent, and more monitorable than industry-typical ice slurry stunning.
More importantly, it is somewhat illusory to think that all the necessary research and R&D can happen first, in isolation, and only then be translated into the field. In industry-facing work, implementation is often what makes the most useful research possible. Without producer buy-in, farm access, commissioning data, firsthand observations of real harvests, and tests in farm conditions, we would be left with a largely theoretical understanding of the problem. We would not know which constraints actually matter, which failure modes appear in practice, or what kinds of equipment, protocols, and monitoring systems can realistically work.
This is especially true in an industry that is often cautious about disclosure around its know-how. Building trust with producers and the wider industry is not incidental to the intervention; it is part of what makes practical progress possible. That is why we think collaborating with the industry is the most effective path: build trust, implement carefully, learn from the field, improve the technology, and strengthen the evidence base as we go. The field implementation work is what creates the conditions for practical, relevant R&D to happen.
Humane slaughter for shrimp is emerging as a higher-welfare standard, and the science on effective stunning is not fully settled. As I’ve outlined, there are things that would increase our confidence levels. But the available evidence indicates that well-implemented electrical stunning beats what is happening on most farms today. And it’s my view that responsible implementation now, alongside continued research, R&D, and monitoring, is one of the strongest bets we can make for reducing suffering at scale.
This is an example of the fact that we have access to more data than that which is publicly available. In this case, the research summary is published online, but we also have access to the full scientific report (although we’re not at liberty to post or share it). There have been many instances like this over the years.
Behavioral indicators are tricky because they’re not always a reliable indicator of unconsciousness. It’s like proving a negative: you can often infer consciousness from the presence of certain behaviors, but you cannot reliably infer unconsciousness from a lack of those behaviors alone. However, Stirling’s team measured behavioral indicators and investigated which of them correlated with EEG results, to determine which might be reliable.
Generally, when shrimps are electrically stunned, there’s one big tail flip when shrimps go through the stunner (likely a reflexive response), followed by immobility. The Stirling study found that this correlates with shrimps having one big EEG spike and then a drop, likely indicating a seizure analog followed by absence of brain activity (a loss of consciousness).
(I manage the FP Climate Fund, so my incentives run counter to my take). I hadn't had the time to examine this paper in detail, but I am quite skeptical of this Rethink Priorities paper. I think it is quite easy to string together assumptions that yield a high social cost of carbon, but I wouldn't treat this as an unbiased estimate.
For example, if I understand this correctly based on your description, they use the Rennert et al (2022) paper to derive the SCC from which they make adjustments.
https://www.nature.com/articles/s41586-022-05224-9/figures/1
The assumptions of that paper are clearly extremely pessimistic, probably by 2022 standards, but definitely by what would now be the consensus view.
For example, they assume close to 20/Gt annual emissions in 2100 as their median scenario and high emissions continue well into the 23rd century. In other words, we are more than a 100 years late in achieving net-zero in their median scenario despite all technological trendlines rendering this quite implausible.
Combining this with a low discount rate will give a high SCC, but I don't think this is close to a reasonable baseline for what a median expectation should be. (Obviously good to have a low discount rate from an EA perspective, but this requires that the modeling of the future is a bit more careful). Essentially, this means that most marginal carbon reduction modeled for the SCC will happen in worlds where this is implausibly valuable thereby inflating the SCC value.
This alone probably leads to an overestimate of the SCC of a factor of 5x or more and this came up from looking at the paper for 5min.
Thanks for developing on this, Richard. I think we can now identify more fine-grained cruxes than we did in our previous discussion, a bit.
Philosophical cluelessness, by contrast, commits one to the much stronger claim that nobody can (realistically) form any reasonable judgments or expectations about this question, no matter how carefully they investigate it. That’s an extremely strong skeptical claim!
You're claiming that you have a truth-tracking expectation of whether doing action A rather than B has an overall positive impact, considering, impartially, all their possible effects on all sentient beings from now until the end of time. I don't see how that's any less strong than suspending judgment on this very question, especially before being given any substantive argument for why I should c-prefer any A over any B (which your post does not do). I agree with Jo's comment.[1]
We can, for example, reasonably expect that starting a nuclear war would be impartially dispreferable in prospect.
I'm sure you'll agree that this could overall decrease x-risks because humanity could survive and be more peaceful/resilient afterwards. Or that human extinction might actually be good because there may be aliens with better values who'd take over if we're not there.
Say an alternative version of you (let's call them Anti-Richard) comes to us and says "my best guess is that starting a nuclear war is impartially good" because of the above considerations, and gives you an incredulous stare for thinking otherwise. You both agree on what the cruxes are but just make different opaque judgment calls. Why should I trust you any more than Anti-Richard? Why should you trust you any more than Anti-Richard? Why should I trust any of you any more than a coin flip?
in many cases we can reasonably—albeit tentatively—expect “our idealized self’s EV for A to be higher, lower, or equal to B’s,”
If you really mean in many cases and not always, then you're objecting to P3 and not P2, I think.
(i) one option can be better in prospect than another, even when you know that, given more information, you would (correctly) regard the two outcomes as incommensurable or on a par; and (ii) in such cases, it’s rational to pick the better prospect rather than deferring to the idealized perspective.
This violates the reflection principle. I think you're going to need to at least give an argument if you want us to give up on this. Why would you believe something your idealized self tells you should not be believed?
since we’ve no grounds for expecting the balance of unknown reasons to count against rather than for our currently-preferred action, the fact of cluelessness is normatively inert: it makes no difference to our reasons for action.
Two possibilities. Either that's a version of the old "canceling out" objection to cluelessness that says the summed EV of the unknowns is exactly 0, and you don't engage with the many compelling rebuttals that have been given thus far (see, e.g., this, this, this, that, and references therein).
Or that's some form of wager based on bracketing out clueless worldviews (see, e.g., DiGiovanni's metanormative bracketing), which has important problems you're not addressing.
Value Correlation: An action’s short-run/visible value is a positive predictor of its long-run/invisible value. Actions that look good on the visible margin are more likely to be good than bad in the long run.
You don't seem to be advocating for human extinction because of humans' current massive negative impact on numerous farmed animals, which I'd bet you believe dominates humans' current welfare-relevant impact on themselves.
So presumably, you think the value correlation thesis is only one consideration among many, and not a slam-dunk one. This is just one thing in the pile of conflicting considerations that may make DiGiovanni and others clueless. This does not help them.
Relatedly, here's a comment with links to texts that clarifies why we should arguably suspend judgment on impartial goodness but not be, e.g., Pyrrhonian skeptics, and that the arguments for the former (those DiGiovanni endorses, at least) are very different from the arguments for the latter, contrary to what you seem to assume.
EA groups have been and continue to be overpowered.
Running an EA group is among the most impactful things you can do when you are at University.
The other day my previous co-organizer sent me a list of 15 (!) people who were in our EA group who are doing exciting and important GCR work.
In the interest of getting this quick take out, I am not reaching out to everyone to see if I can list their name/quote them but I want to stress the “exciting and important” aspect of this. Some of these people are perhaps within the top 10 people affecting American AI policy from a GCR lens. And they only graduated within the last 5 years - many of whom started in very impactful roles in less than two years(!). People are advising on 100s of millions of dollars, in political office, and founding exciting new orgs in the ecosystem (like, actually awesome orgs!).
And that is only GCR work - a previous member of our group is now CEO of one of the largest EA-affiliated global health orgs and another, a researcher at GiveWell.
When I was a group organizer, I was often stressing about whether I was having an impact and I often hear organizers stressing about whether the payoff will be quick enough. It is true that my group’s outcomes are not the median and that sometimes it does just take many years to pay off - but there definitely are existence proofs of massive impact on a pretty quick timeframe.
I have been involved in EA groups since 2017. Almost a decade now!
And in the meantime I’ve seen so many exciting impact stories coming out from groups all over the world.If you are interested in more of my takes on groups you can check out some of my past talks: Hyping them up and Sharing potential pitfalls
If you are interested in running a group there are many resources available to you!