I am an attorney in a public-sector position not associated with EA, although I cannot provide legal advice to anyone. My involvement with EA so far has been mostly limited so far to writing checks to GiveWell and other effective charities in the Global Health space, as well as some independent reading. I have occasionally read the forum and was looking for ideas for year-end giving when the whole FTX business exploded . . .
Jason
Karma: 18,925
76% of experts saying it’s “unlikely” the current paradigm will lead to AGI leaves ample room for a majority thinking there’s a 10%+ chance it will . . . .
. . . . and the field are still mostly against you (at the 10% threshold).
I agree that the “unlikely” statistic leaves ample room for the majority of the field thinking there is a 10%+ chance, but it does not establish that the majority actually thinks that.
I would like to bring back more of the pre-ChatGPT disposition where people were more comfortable emphasizing their uncertainty, but standing by the expected value of AI safety work.
I think there are at least two (potentially overlapping) ways one could take the general concern that @Yarrow Bouchard 🔸 is identifying here. One, if accepted, leads to the substantive conclusion that EA individuals, orgs, and funders shouldn’t be nearly as focused on AI because the perceived dangers are just too remote. An alternative framing doesn’t necessarily lead there. It goes something like there has been a significant and worrisome decline in the quality of epistemic practices surrounding AI in EA since the advent of ChatGPT. If it—but not the other—framing is accepted, it leads in my view to a different set of recommended actions.
I flag that since I think the relevant considerations for assessing the alternative framing could be significantly different.
Given EA’s small share of the total global health/poverty funding landscape, the most likely effect of its investment on an expensive-but-permanent project is to speed the timetable up. So, for instance, perhaps we would get a hypothetical vaccine a year or two earlier if there had been EA investment. So, in comparing the effects of a yearly intervention vs. an expensive-but-permanent one, we are still looking at near-term effects that are relatively similar in nature and thus can be compared.
I don’t suggest that is true for all “permanent” interventions, though, so it isn’t a complete answer. It also might not scale well to a field in which EA funding is a large piece of the total funding pie.
Some supporters of AI Safety may overestimate the imminence of AGI. It’s not clear to me how much of a problem that is?
It seems plausible that there could be significant adverse effects on AI Safety itself. There’s been an increasing awareness of the importance of policy solutions, whose theory of impact requires support outside the AI Safety community. I think there’s a risk that AI Safety is becoming linked in the minds of third parties with a belief in AGI imminence in a way that will seriously if not irrevocably damage the former’s credibility in the event of a bubble / crash.
One might think that publicly embracing imminence is worth the risk, of course. For example, policymakers are less likely to endorse strong action for anything that is expected to have consequences many decades in the future. But being perceived as crying wolf if a bubble pops is likely to have some consequences.
Hot take behind a semi-veil of ignorance on this year’s results: I submit that next year, there should be a modest allocation (~10%) for the best finisher in certain categories if no org in that category makes the top three:
Last org standing in a major cause area grouping (i.e., GHW, AW, LT, meta/cross-cause/other)
Last small org standing (i.e., annual budget less than . . . 300K? 500K? 750K??)
If the main value of the election is eliciting / signaling community preferences, then I think it’s helpful to have good information available for each major cause area and also to signal-boost a small (usually upstart) org or two. Guaranteeing a modest pot of money for each sub-winner should improve the quality of the signal.
If the main value of the election is driving engagement, then I think it’s helpful to give (almost) everyone one race in which they are more invested in the outcome / feel like there’s an option to meaningfully support one org in their preferred cause area.
Re: Nestle in particular, I get the spirit of what you’re saying, although see my recent long comment where I try to think through the chocolate issue in more detail. As far as I can tell, the labor-exploitation problems are common to the entire industry, so switching from Nestle to another brand wouldn’t do anything to help??
That could be correct. But I think the flip side of my individual chocolate purchasing decisions aren’t very impactful is that maybe we should defer under some circumstances to the people who have thought a lot about these kinds of issues, even if we think their modeling isn’t particularly good. Weak modeling is probably better, in expectancy, than no modeling at all—and developing our own models may not be an impactful use of our time. Or stated differently, I would expect the boycott targets identified by weak modeling to be more problematic actors in expectancy than if we chose our chocolate brands by picking a brand out of a hat.[1] (This doesn’t necessarily apply to boycotts that are not premised on each additional unit of production causing marginal harms.)
- ^
Of course, we may not be picking a brand at random—we may be responding to price and quality differences.
- ^
The most likely problem is that a donor must reduce the amount of their deduction by the amount of the personal benefit they received as a result of the donation. That the benefit is bestowed by a third party doesn’t change the result. Here, the donor is receiving a very real and substantial personal benefit (release from their personal indebtedness). So after the reduction, their donation amount is $0.
I agree with the view that this starts in 2026:
SEC. 70424. PERMANENT AND EXPANDED REINSTATEMENT OF PARTIAL DEDUCTION FOR CHARITABLE CONTRIBUTIONS OF INDIVIDUALS WHO DO NOT ELECT TO ITEMIZE.
(a) In General.--Section 170(p) is amended-- (1) by striking ``$300 ($600″ and inserting ``$1,000 ($2,000″, and (2) by striking ``beginning in 2021″.
(b) Effective Date.--The amendments made by this section shall apply to taxable years beginning after December 31, 2025.
(emphasis mine).
https://www.congress.gov/bill/119th-congress/house-bill/1/text/enr
This is particularly relevant because the reader is dependent on trusting Snowdon to have conducted a balanced review of the literature. I was already skeptical that he had done so even before considering that he has an ideological axe to grind and now a potential financial conflict of interest.
The article has value; it reminds us that local context matters and that we should exercise caution in blindly applying a general finding to specific circumstances. But it would have been a lot stronger if it had grappled with the arguments and evidence that led WHO to deem these interventions best buys, and had acknowledged the limited scope of his own data analysis more.
I find it paradoxical that the signature strategy of a major cause area—threatening a “haze of bad vibes and negative associations” if a corporation doesn’t somewhat clean up its animal-welfare record—probably would be ineffective if everyone acted like EAs. The mechanism of action is still dependent on individual consumer choice (“tilt you towards going to a nearby mcdonalds or panda express instead that day”) and the commentariat can be read as implying to OP that making individual consumption decisions based on such considerations is too low-impact to pay attention to.
There’s something that feels vaguely non-cooperative about this—we’re dependent on other people responding to our threatened PR campaigns regarding animal welfare (or at least on corporations believing other people would respond), but seem not interested in cooperating with other people’s altruistic PR campaigns. I’m not sure there is anything practical to do with with this mood—other than encourage OP to present clearer models of impact for the interventions they mentioned—but I think it is worthwhile to acknowledge the mood. And maybe if you have a choice in the supermarket between a Nestle and non-Nestle chocolate product, consider purchasing the latter?
The underlying principle seems similar to the reasoning behind veganism: purchasing animal products contributes, at least marginally, to the continued production of those products. Please correct me if I am mistaken in drawing this parallel.
There is likely similarity in at least some cases, but it may be attenuated.
Given that Nestle presumably has significant fixed costs, it is unclear what the effect of a small group of consumers boycotting would be. You presumably have Nestle and non-Nestle chocolate. Most consumers are indifferent between the two, while some consumers refuse to buy the Nestle product. If some consumers start refusing the Nestle product, it has two choices. One, it can cut back production. Two, it can sell its “excess” production to consumers who don’t care where their chocolate comes from (or to other companies which then sell cocoa-based products to consumers). I don’t have a good sense of what the actual reduction in Nestle’s production of a consumer boycotting Nestle is, but I suspect it is far from a 1:1 reduction.
(If you’re wondering whether this logic also applies to vegan consumption decisions—yes, it does to some extent even though the consumer is reducing industry-wide demand. For instance, reducing one’s consumption of chicken by 1 pound is expected to reduce chicken production by about 0.76 pounds.)
My understanding is that soda is produced locally, so I’d be looking for evidence that my Coca-Cola consumption counterfactually affected water availability outside my community (which does not, to my knowledge, experience water shortages). I’d also have to consider the adverse effects of what I drank instead—maybe drinking Pepsi has less negative environmental effect, but it’s not plausible that it has no effect. And switching to quality apple juice (not from concentrate) might be worse, due to the distances the juice would travel.
Figuring out the actual impact of these boycotts would require a lot of modeling (and perhaps some data that would be difficult to obtain). But I suspect the actual impact is significantly less than many people participating in the boycott assume.
However, I am interested in your view on which kinds of actions should be considered morally permissible and which should be regarded as morally obligatory.
I don’t know. Philosophers have spilled much ink on that question, and I’ve never found an answer I find satisfying.
I also find it notable that issues such as widespread contempt for Nestlé or the extensive discussions about ethical and fair-trade chocolate seem largely overlooked in this forum.
That they are in widespread circulation is arguably a good reason to not focus on them here! I knew fair-trade chocolate is a thing, and I suspect most readers here did too. We know where to go for information on the subject if we want to change our consumption patterns. But information about shrimp welfare, there aren’t too many other places to hear about that.
Attention is a limited resource, and there is much evil and suffering in the world. All altruistically-minded communities pick and choose what their focus issues and methods are. They would collapse from incoherence otherwise. Individuals are likewise constrained lest they experience burnout.
Moreover, giving focus to things that aren’t very effective would be mildly corrosive to the spirit of EA. In part, the movement grew out of a belief that traditional charities and donors can put far too much stock in what makes the charity/donor feel righteous/ethical (irrespective of actual magnitude of results), or what gives the charity/donor higher status in society. I do think it is important to maintain strong boundaries against those impulses.
In the end, my own lightly-held assumption is that most boycotts of the sort you describe are not effective enough in producing significant enough changes in production to be worth diverting community attention from more effective actions. The lack of attention to them here suggests most of the user base would agree. But the beauty of the Forum is that you can run the numbers and present a model explaining why that tentative view is incorrect.
I think this highlights why some necessary design features of the karma system don’t translate well to a system that imposes soft suspensions on users. (To be clear, I find a one-comment-per-day limit based on the past 20 comments/posts to cross the line into soft suspension territory; I do not suggest that rate limits are inherently soft suspensions.)
I wrote a few days ago about why karma votes need to be anonymous and shouldn’t (at least generally) require the voter to explain their reasoning; the votes suggested general agreement on those points. But a soft suspension of an established user is a different animal, and requires greater safeguards to protect both the user and the openness of the Forum to alternative views.
I should emphasize that I don’t know who cast the downvotes that led to Yarrow’s soft suspension (which were on this post about MIRI), or why they cast their votes. I also don’t follow MIRI’s work carefully enough to have a clear opinion on the merits of any individual vote through the lights of the ordinary purposes of karma. So I do not intend to imply dodgy conduct by anyone. But: “Justice must not only be done, but must also be seen to be done.” People who are considering stating unpopular opinions shouldn’t have to trust voters to the extent they have to at present to avoid being soft suspended.
Neutrality: Because the votes were anonymous, it is possible that people who were involved in the dispute were casting votes that had the effect of soft-suspending Yarrow.
Accountability: No one has to accept responsibility and the potential for criticism for imposing a soft-suspension via karma downvotes. Not even in their own minds—since nominally all they did was downvote particular posts.
Representativeness: A relatively small number of users on a single thread—for whom there is no evidence of being representative of the Forum community as a whole—cast the votes in question. Their votes have decided for the rest of the community that we won’t be hearing much from Yarrow (on any topic) for a while.[1]
Reasoning transparency: Stating (or at least documenting) one’s reasoning serves as a check on decisions made on minimal or iffy reasoning getting through. [Moreover, even if voters had been doing so silently, they were unlikely to be reasoning about a vote to soft suspend Yarrow, which is what their votes were whether they realized it or not.]
There are good reasons to find that the virtues of accountability, representativeness, and reasoning transparency are outweighed by other considerations when it comes to karma generally. (As for neutrality, I think we have to accept that technical and practical limitations exist.) But their absence when deciding to soft suspend someone creates too high a risk of error for the affected user, too high a risk of suppressing viewpoints that are unpopular with elements of the Forum userbase, and too much chilling effect on users’ willingness to state certain viewpoints. I continue to believe that, for more established users, karma count should only trigger a moderator review to assess whether a soft suspension is warranted.
- ^
Although the mods aren’t necessarily representative in the abstract, they are more likely to not have particular views on a given issue than the group of people who actively participate on a given thread (and especially those who read the heavily downvoted comments on that thread). I also think the mods are likely to have a better understanding of their role as representatives of the community than individual voters do, which mitigates this concern.
You may want to edit this to include a more descriptive title.
Keeping a ban in place for a median of ~10 years (based on the four mostly historical examples you cited) strikes me as a pretty successful failure. There are reasons to have some skepticism about a ban, but I don’t see this as one of them.
The published comment doesn’t display your votes, so you might want to edit it if you want the public to know to which organizations you were referring.
That’s certainly a relevant question, and I think some of it would depend on the contours of the attempting quashing. It’s quite plausible—maybe even probable—that some combination of demand reduction (without outright banning), harm reduction, and other softer techniques would achieve superior results as a practical matter.
That being said, the volume of tobacco consumed by an active smoker (e.g., a pack a day) is many times greater in volume of (e.g.) street opioids where even 2 mg of fentanyl can be a fatal dose. There are about 425MM pounds of tobacco harvested in the US on about 200K acres. I’d expect the volume involved to make smuggling enough tobacco-leaf products significantly more challenging than it is for most drugs. (Or alcohol, which can be produced quite easily using basic equipment.) And so a ban on tobacco leaf might stand a better chance than the median drug ban. On the other hand, it may not be particularly hard to synthesize and smuggle nicotine itself.
Another possible difference might be how society manages addiction in this hypothetical new world. There’s already broad acceptance for treating nicotine addiction as a health problem that warrants professional treatment. If there were a safe, legal, inexpensive way for people with nicotine dependence to get treatment (including nicotine replacement therapy) in a post-quash world, that would take a lot of the wind out of the sails of the black market.
I’d be interested in understanding what you mean by “curing addiction” in this context—would it be something like never having known the pleasure of smoking, adding a moment of mindful reflection before each cigarette, or something else? And how can we tell whether someone is making the choices they are making because they are addicted vs. because they have weighed the pros and cons sufficiently for us to believe they are making a free choice?
I think you’re right that this is a significant crux. It’s a thought experiment (hence “magic wand”), so I don’t think it’s necessary to have a perfect operationalization or measurement criteria here. That being said, I was envisioning that the wand would put smokers in the same position as people who have experienced pleasurable but generally non-addictive events in the past. For example: sex is generally pleasurable, people know it is generally pleasurable, but people aren’t generally addicted to it. If they choose to have sex, even under circumstances where I might think it isn’t the best decision for them, I don’t generally question that this is their free choice.
I expect that, at a minimum, the magic wand would reverse neurobiological changes commonly seen with addictive behavior and would leave the subject with the firm conviction that they could choose to stop without any withdrawal symptoms, cravings, and so on.
The results would be a 2x2 matrix: did the smoker decide to return to smoking after the wand treatment, and did they decide correctly (based on what would make them happier given their own values and preferences?) People make suboptimal decisions all the time—or at least I do! So I do not presume that their choice in thought experiment was the correct one for them. On the other hand, I do want to give their choice significant weight.
Groups 1C and 1E: People who decided to quit smoking. This is estimated as: the 67% who said they wanted to quit, less a downward social-desirability adjustment, plus an upward freedom-from-addiction. One could only speculate on the size of this group, although I think the latter adjustment would swamp the former. I also think very few members of this group would err (by their own values and preferences) in deciding to quit—we will put those people in Group 1E instead of Group 1C.
Group 2E: People who decided not to quit, but would have been happier (by their own values and preferences) had they done so. I would not be inclined to credit an intervention that abolished the tobacco industry as helping members of this group; that strikes me as too paternalistic. But I am also not inclined to count them as having been harmed as a side effect of abolishing the tobacco industry either; by definition they are better off without it. I’ll treat Group 1E the same way for consistency (i.e., weighing the smoker’s actual decision equal to whether they’d actually be happier, and only assigning net harm or benefit when both are in the same direction).
Group 2C: People who decided not to quit, and who would be happier with that decision. These are the only individuals clearly harmed by abolition.
Unfortunately, magic wands don’t exist, and so we can only speculate about the relative sizes of these groups. But if one believes that Group 1C is much larger than the rest (and especially if one believes that Group 2C is pretty small), then accounting for the net costs to Group 2C actually isn’t that important in generating a model.
Do you think there’s a known realistic way to abolish the industry as a whole when more than a billion people in the world smoke daily?
I think SMA’s stated goal is somewhat hyperbolic in the short/medium run, although it might be a realistic goal over something like 75-100 years. For example, public support for proposed generational-ban legislation in the UK appears strong.
If yes, what’s a tractable path given the track record of prohibition of psychoactive substances more generally, and tobacco specifically (e.g. recent bans and subsequent reversals in South Africa and Bhutan)?
Even assuming that a generational ban would be dead on arrival, running an abolitionist campaign has some helpful potential failure modes. First, it could move the Overton window and make it earlier to establish policies that reduce usage levels and/or move usage toward less harmful forms. Second, pushing for a generational ban (e.g., raising the age of legal purchase by one day each day) could get watered down in the legislature to (e.g.) gradually increasing the legal age to the mid-20s, or to imposing a generational ban for only higher-harm products. Those would be massive wins.
In contrast, “educating the public about the risks of smoking and assisting those who desire to quit” sounds a lot like the status quo approach—which has led to over a billion people smoking, most of whom wish they could quit. Governments have already spend quite a bit on educating the public, and the quitting-nicotine market is already quite large (even if its offerings leave much to be desired in terms of efficacy). There’s no reason to think SMA folks would be more skilled at educating people about the harms of nicotine than public health experts, or that they would be better at designing anti-addiction drugs than expert psychopharmacologists whose firms have a massive financial incentive to succeed.
It’s not yet sufficiently clear to me that WAW has the readily accessible levers to magnify impact that FAW has. Animals are farmed by humans, who are employed by a manageable number of corporations and answer to a manageable number of governments. That means that changing a single corporation or government’s stance can have a massive effect. Moreover, public support to secure a change may be easier to obtain for the average farmed animal.
My view is lightly held, but for now I think I need more evidence that highly-leveraged wins are practically achievable before moving to the WAW side of the ledger.
There’s a rule that the organization needs to contribute a comment or post during Marginal Funding Week to be eligible to opt-in to the election. So the first step would be encouraging them to do so next year.
non-tax-deductible opportunities (e.g. 501c4. . ., even future 501c3s awaiting their 501c3 determination) . . .
I’d note that when an organization files for exempt status within 27 months of its creation, the approval of 501c3 status is retroactive to the organization’s founding. If the approval happens after the donor files their return for the current year, the donor would need to file a 1040-X amended return. So it’s more accurate to say these donations pose a risk of non-deductibility (although I don’t think the base risk is that high).
So people who are willing to file 1040-X, which tax software can do, shouldn’t discount much for application-pending status even if they highly value 501c3 status. This is probably a barrier for people donating through DAFs, employer matching programs, etc. and so I think your broader point that there’s a higher risk of neglected less here is correct.
Even if the organization files late, approved status is at least retroactive to the date of filing.
https://www.irs.gov/instructions/i1023
[edited 12⁄2 PM for formatting]
Also, if I were on the low probability end of a bet, I’d be more worried about the risk of measurement or adjudicator error where measuring the outcome isn’t entirely clear cut. Maybe a ruleset could be devised that is so objective and so well captures whether AGI exists that this concern isn’t applicable. But if there’s an adjudication/error error risk of (say) 2 percent and the error is equally likely on either side, it’s much more salient to someone betting on (say) under 1 percent odds.