Jamie_Harris

• Jamie_HarrisNov 30, 2024, 7:07 PM

  6 points

  3 ∶ 0

  in reply to: Ariel Simnegar 🔸’s comment on: Fund Causes Open Phil Un­der­funds (In­stead of Your Most Preferred Causes)

  Yeah it might be more tractable.

  Focusing solely in EAs has a bunch of weird effects though.

  E.g. I’ve been thinking about some ‘safeguarding democracy’ type interventions for longtermist reasons. If I looked at EA funding I’d presumably conclude that the area was massively underfunded—almost no one working on this. Whereas looking in a global sense the initial impression is that it’s a very large, well-funded area. (Maybe it’s still a useful heuristic though because explicitly longtermist funding and effort might focus on quite different subcomponents of the broad topic?)

  And another one is just that how liberal you are in your definitions of what’s EA or not can make quite a big difference. E.g. plausibly by a factor of 2 in the case of animal advocacy.

  (No need to reply, I’m just musing.)

  Thanks!

• Jamie_HarrisNov 28, 2024, 8:24 AM

  19 points

  1 ∶ 0

  on: Has your or­gani­sa­tion lost fund­ing due to the Good Ven­tures fund­ing shift? Have you man­aged to re­place it?

  Open Philanthropy instituted a policy of no longer funding community-building grants focused on high-school students or minors. The org I have run for several years, Leaf, is currently funding constrained partly due to this. I don’t know of funders who have ‘stepped in’ so to speak (leads welcome!), although I also work a few hours a week at the EA Infrastructure Fund which is happy to make grants in this area.

• Jamie_HarrisNov 26, 2024, 7:56 PM

  4 points

  1 ∶ 0

  in reply to: Luke Moore 🔸’s comment on: Con­sider donat­ing to who­ever helped you

  I had a related concern that (if this funding model became more widespread) it could lead to overinvestment in more legible/​obvious contributions relative to more behind-the-scenes ones. E.g. 80k gets lots of donations but Effective Ventures doesn’t; a local group gets donations but the orgs who created the resources they reused and adapted (BlueDot, CEA, etc) doesn’t.

  (I still think on net I agree that it’d be cool to shift somewhat in this direction though. And I think more community building services should consider charging or making donations an opt-out rather than opt-in default.)

• Jamie_HarrisNov 26, 2024, 7:35 PM

  13 points

  1 ∶ 0

  on: Fund Causes Open Phil Un­der­funds (In­stead of Your Most Preferred Causes)

  I initially found myself nodding in agreement but then I realised a confusion I have:

  Why should a donor/​grantmaker limit their consideration of what is most underfunded to the EA community?

  After all, the EA community is a nebulous community with porous boundaries. E.g. we count Open Phil, but what about The Navigation Fund? Bill and Melinda Gates Foundation? And even if we can define the boundaries, what do we actually gain by focusing on this specific subset of donors?

  If you instead focus on ‘what is most underfunded at the global level’ then the question returns to the same broad question of cause prioritisation (“your value system’s most preferred causes”).

• Jamie_HarrisNov 18, 2024, 9:08 AM

  4 points

  0 ∶ 0

  in reply to: MichaelStJules’s comment on: Weighted Fac­tor Models: Con­sider us­ing the ge­o­met­ric mean in­stead of the ar­ith­metic mean

  Yeah I agree in principle it “might be for good reason”, though I still have some sense that it seems desirable to reduce overdependence on your ratings for one or two criteria. Similar to the reasoning for sequence thinking vs. cluster thinking

• Jamie_HarrisNov 17, 2024, 5:16 PM

  4 points

  0 ∶ 0

  in reply to: MichaelStJules’s comment on: Weighted Fac­tor Models: Con­sider us­ing the ge­o­met­ric mean in­stead of the ar­ith­metic mean

  I sometimes do this, but I wonder if it defeats one of the key benefits of a WFM—that it accounts for multiple criteria and prevents any single consideration dominating.

  (With BOTECs, sometimes the final ranking/​conclusion is very dependent on one or two very uncertain or arbitrary criteria.)

Shap­ing effec­tive al­tru­ism: Light­ning talks on com­mu­nity and field building

• Jamie_HarrisOct 28, 2024, 7:10 PM

  2 points

  0 ∶ 0

  in reply to: John Salter’s comment on: What should EAIF Fund?

  Also just copying unedited a related rough note to self I made on my own list of potential entrepreneurial projects (but which I’m very unlikely to ever actually work on myself)

  Impact-focused red-teaming, consultancy, and feedback marketplace

  Problem I faced as a founder: Often making decisions where I would have loved external input but felt reluctant because I didn’t want to ask people for favours, and didn’t necessarily know who I could ask other than my personal network.

  Solution: a platform like Fiverr or similar where people willing to give feedback and advice are listed with a description of their interests/​background, their hourly rate, plus a bunch of tags (e.g. cause area, experience type) so that people can filter by types of role.

  Probably take a % cut of the fees. E.g. Fiverr system: “Service fees are 5.5% of the purchase amount. For purchases under $100, an additional $3.00 small order fee will be applied.” Maybe also add some minimum amount, in case people want to list their services for free, e.g. we always charge minimum $1 per hour of service performed. (I think maybe we could get away with charging a lot higher rates than Fiverr though, like 10% or 20%. E.g. if someone is willing to pay $50 for 2 hours of review from someone, they’re probably willing to pay $60

  Later could potentially expand into a broader freelancing type platform.

  Misc idea: crowdsource ratings of how well the job was done by a bunch of different criteria, and encourage ACCURACY in the ratings. Batch the ratings in groups of 5 to help anonymise the source of the rating and thereby reduce embarassment for not giving wholly positive ratings.

  Downside: Probably pretty small volumes of money. E.g. a person read-teaming a doc for 2 hours might only charge ~$50, of which we would get ~$5. So we’d need to achieve a really high usage volume for it to become financially self-sustaining and justify the time on the setup.

  [I’m sharing partly because my “Misc idea” seems like it could help with the problem you highlight.]

• Jamie_HarrisOct 28, 2024, 7:07 PM

  2 points

  0 ∶ 0

  in reply to: John Salter’s comment on: What should EAIF Fund?

  Interesting! Could you share one or two examples of the “Templates [that] already exist on-line for this exact type of marketplace”?

  Since you’ve thought about it a bit already, I’d be interested if you have any thoughts on how long something like this would take to setup on technical/​operational side to a high standard, excluding time spent “finding enough early adopters”.

  (Also, I’m guessing this isn’t something you’re interested in doing yourself?)

• Jamie_HarrisOct 25, 2024, 1:15 PM

  8 points

  1 ∶ 0

  on: The cost of (not) tak­ing mar­ket­ing seriously

  Hey James! I’ve heard the claim a couple of times that EA orgs systematically underinvest in marketing. I was wondering if you are able to share (here or privately) any direct evidence that EA orgs are indeed doing that?

  I appreciate the broad point that ‘marketing can have benefits’ (a crude summary of the takeaway from the “5 key pieces of research here”), but is there evidence that EA orgs are (consistently) spending less time and money on marketing that would be optimal?

  E.g. skimming this post I didn’t really see evidence for the claim in the first sentence:

  By not properly addressing the role of marketing in the effective altruism movement, there is substantial impact being forfeited

  A separate but somewhat related question that happens to be relevant to something else I’m thinking about at the moment: do you have a rough take on how much EA meta orgs should be ‘willing to pay’ for newsletter subscribers, vs how much they are WTP? Again, might be easier to discuss with reference to specific orgs, potentially not publicly on the Forum.

  Thanks!

• Jamie_HarrisSep 19, 2024, 10:51 AM

  4 points

  0 ∶ 0

  in reply to: JWS 🔸’s comment on: In the last 2 years, what sur­pris­ing ideas has EA cham­pi­oned or how has the move­ment changed its mind?

  Thanks! That’s helpful.

  • Seems to me that at least 80,000 Hours still “bat for longtermism” (E.g. it’s very central in their resources about cause prioritisation.)

  • Not sure why you think that no “‘EA leader’ however defined is going to bat for longtermism any more in the public sphere”.

  • Longtermism (or at least, x-risk /​ GCRs as proxies for long-term impact) seem pretty crucial to various prioritisation decisions within AI and bio?

  • And longtermism unequivocally seems pretty crucial to s-risk work and justification, although that’s a far smaller component of EA than x-risk work.

  (No need to reply to these, just registering some things that seem surprising to me.)

• Jamie_HarrisSep 17, 2024, 7:34 PM

  4 points

  0 ∶ 0

  in reply to: JWS 🔸’s comment on: In the last 2 years, what sur­pris­ing ideas has EA cham­pi­oned or how has the move­ment changed its mind?

  “Longtermism is dead”: I feel quite confused about what the idea is here.

  Is it that (1) people no longer find the key claims underlying longtermism compelling? (2) it seems irrelevant to influencing decisions? (3) it seems less likely to be the best messaging strategy for motivating people to take specific actions? (4) something else?

  I’m also guessing that this is just a general summary of vibe and attitudes from people you’ve spoken to, but if there’s some evidence you could point to that demonstrates this overall point or any of the subpoints I’d be pretty interested in that.

  (Responding to you, but Peter made a similar point.)

  Thanks!

• Jamie_HarrisAug 15, 2024, 11:11 AM

  9 points

  0 ∶ 0

  in reply to: Jamie_Harris’s comment on: Jamie_Har­ris’s Shortform

  Just dumping this here in case it’s helpful for someone stumbling back across this: Here’s a “Worksheet for choosing the most pressing problem” I made.

• Jamie_HarrisJun 25, 2024, 7:14 PM

  2 points

  0 ∶ 0

  on: 2022 AI Align­ment Course: 5→37% work­ing on AI safety

  So many incredible achievements!

• Jamie_HarrisMay 28, 2024, 7:42 AM

  10 points

  2 ∶ 0

  in reply to: Ryan Kidd’s comment on: Ta­lent Needs of Tech­ni­cal AI Safety Teams

  Yep, I realise that.

  Also feel like a big limitation is that this data comes from asking current orgs. Asking current orgs how many “connectors” they need feels a bit like asking a company how many CEOs they want.

  Nonetheless, still an update! E.g. this bit was slightly surprising to me:

  Funders of independent researchers we’ve interviewed think that there are plenty of talented applicants, but would prefer more research proposals focused on relatively few existing promising research directions (e.g., Open Phil RFPs, MATS mentors’ agendas), rather than a profusion of speculative new agendas. This leads us to believe that they would also prefer that independent researchers be approaching their work from an Iterator mindset, locating plausible contributions they can make within established paradigms, rather than from a Connector mindset, which would privilege time spent developing novel approaches.

• Jamie_HarrisMay 27, 2024, 9:07 PM

  2 points

  0 ∶ 0

  on: Ta­lent Needs of Tech­ni­cal AI Safety Teams

  I found this helpful. It updated me towards the importance of finding/​supporting iterators, relative to connectors. Thank you!

• Jamie_HarrisMay 26, 2024, 6:09 PM

  2 points

  0 ∶ 0

  in reply to: Larks’s comment on: Sur­vey: bioethi­cists’ views on bioeth­i­cal issues

  I agree with the basic point you’re making (I think) and I suspect either:

  (1) we disagree about how much you should negatively update, i.e. how bad this data makes bioethicists look

  Or

  (2) we don’t actually disagree and this is just due to language being messy (or me misinterpreting you)

• Jamie_HarrisMay 26, 2024, 6:01 PM

  2 points

  0 ∶ 0

  in reply to: Richard Y Chappell🔸’s comment on: Sur­vey: bioethi­cists’ views on bioeth­i­cal issues

  That all seems fair /​ I agree.

• Jamie_HarrisMay 25, 2024, 1:14 PM

  19 points

  5 ∶ 3

  on: Sur­vey: bioethi­cists’ views on bioeth­i­cal issues

  Thanks a lot for collecting this survey! I think it’s valuable to solicit ‘external’ (to the EA community) views on important questions that affect our decision-making, especially from plausible expert groups.

  I’m quite shocked at the vehemence and dismissiveness of many of the comments on this post responding to these results. Here are some quotes from other commenters:

  1.

  “Preventing a death is equally important irrespective of age” strikes me as a genuinely insane position… No one would be indifferent between extending someone’s life by an hour, even a very valuable hour, and extending another person’s ordinary life by 30 years. But it’s just really strange to endorse that, but not apply the same logic to saving a 20-year old person over a 100-year old person.

  2.

  Yeah, it’s just transparently stupid stuff like “Each life counts for one and that is why more count for more. For this reason we should give priority to saving as many lives as we can, not as many life-years.” [Caveat, this quote is slightly out of context… it’s actually responding to the comment above.]

  3.

  Results give some support to the notion that bioethicists are more like PR professionals, geared to reproducing common sentiments rather than a group that is OK with sometimes taking difficult stances. Questions 6 & 7 especially seem like vague left-wing truisms… I still can’t get over 40% thinking being blind would be not disadvantaging if society was “justly designed”.

  4.

  It’s really pretty shocking to me how badly this makes bioethicists look.

  Here are some possible explanations for the supposedly crazy results:

  • There are reasons, logic, or evidence that are considered or known amongst (some? many?) bioethicists that you yourself are not familiar with.

  • There are reasons, logic, or evidence that you are familiar with that (some? many?) bioethicists are not.

  • There are multiple conflicting principles or heuristics that apply in a different case, and the respondents just weigh those differently to you. E.g. this strikes me as likely what’s happening with the “It is most important to prevent someone from dying at which of the following ages” question.

  • The respondents have different ethical systems and worldviews to you, e.g. placing more weight on virtue ethics relative to consequentialism. That doesn’t make them insane or unthoughtful; ethics is really tough and probably depends on a lot of things like your upbringing. (Otherwise people would have similar ethical views across cultures, which clearly isn’t true.)

  I share the intuition that many of the results in the survey seem surprising, and very discrepant from my own views. But regardless of whether you understand the reasons, surely after seeing that a group of people holds substantially different views to your own, your all-things-considered belief should shift at least somewhat towards those views, even if your “independent impression” does not? Especially when that group has years of relevant thought or expertise; these facts make it more likely that there are valid reasons underpinning their beliefs. Where there are discrepancies, there’s a chance that they are right and you/​we are wrong.

  I’m worried that some of the quotes above represent something like cognitive dissonance or a boomerang effect. Or at least they seem more like “soldier mindset” than I’d expect here, although I note some exceptions where several commenters (including some of those I quoted above) ask others for input on helping to understand and steelman the bioethicists’ views.

  [Edit: the following paragraph felt true at the time of writing but I regret writing it as it seems pointlessly offensive/​inflammatory itself in hindsight. I apologise to the people I quoted above.] Honestly, seeing the prevalence of these kinds of reactions in the comments makes me feel less confident in the epistemic health of this community and more worried about groupthink type effects. (Maybe some of these commenters have reasons for their vehemence and dismissiveness that I’m missing?)

• Jamie_HarrisMay 24, 2024, 8:02 PM

  4 points

  0 ∶ 0

  on: The suffer­ing of a farmed an­i­mal is equal in size to the hap­piness of a hu­man, ac­cord­ing to a survey

  Very interesting results. They seem surprisingly animal-friendly/​considerate to me.

  “The survey title and introduction didn’t mention objectives of the survey, except that it is about animal welfare”.

  To check, could people see this before deciding whether or not to take the survey? If they could, I’d expect that to skew the participation substantially towards people who already care disproportionately about animals, or think them to be closer in capacity to humans etc.

  If so, did you collect any demographic information that wasn’t filled based on quotas and can be compared to wider Dutch demographic data that might give insight into this? E.g. rates of vegetarianism? Or even if you found that survey participation was filled more quickly by women than by men, I’d take that as a relevant indication, given correlations between gender and various measurements relevant to caring about animals.

  (Even if these aspects are similar to the wider population, or you only shared that it was about animal welfare later on after people signed up, I’d expect social desirability bias to influence the results somewhat.)