was in charge of IT and software at Alvea
ex-boardmember of EA Germany
co-founded two small startups
ex-human rights observer with Peace Brigades International
Flo šø
Karma: 693
One costly public signal they could use during those days would be to give away inference at an extreme discount or even for free to demonstrate that they donāt use the compute for training or research.
Yes, sorry if Iām coming on to strong, it just feels like more āeverythingās vibes, we just do whateverā for me
Wait, so pledges from you personally on the blog revival project site can just be retroactively declared a joke with no intention of paying out?
Iām happy that people overwhelmingly think that this is a bad idea. When I posted the comment, this wasnāt clear to me. I posted this because all I had seen at that point was a stream of positive comments on substack which all celebrated that Caroline was given a second chance in this way. I wanted there to at least be one comment expressing why this is bad.
I may have overestimated Manifunds importance since it was regularly recommended to me as where I should apply to fund a side project of mine, and I seriously considered it.
So no, I donāt reconsider my involvement after every bad action by any org, only when something seems to me like one of the worst moves, while seeing people I respect celebrating it, with little to no pushback (which now is clearly happening).
Look, Austin made the comparison
I wish that the public would see āEAā as something like, idk, āCatholicā, where one bad or disagreeable act doesnāt make a particular statement about others.
I may have been more short tempered and aggressive than necessary, and arguing the specifics of what harm the Catholic Church has done is indeed beside the point. What I was gesturing at was: Wishing that EA were the same as a 2000 year old world-spanning religion in terms of public perception when scandals happen has some obvious problems.
I donāt expect you to; this is mostly about the optics, no internal details needed.
It took the Catholics 2000 years to get to the point where they could systematically protect and enable paedophiles and still survive the blowback. It should be quite obvious why EAās reputation is far more brittle, especially when it comes to funding organisations.
I think you have every right to hire any cofounder you like if you run a startup, whatever their history, but then it shouldnāt be a 501c associated in such a big way with Effective Altruism.
From the perspective of someone who spent years in tedious discussions about being extremely careful with public outreach to not damage the perception of EA as a brand when we were building EA Germany, this feels like a slap in the face.
After researching this a little more, Iām quite taken aback that apparently you still didnāt return the 500.000$ FTX investment, and instead let it go to litigation. So the Trust that is making the victims whole is currently suing you (Manifold for Charity) for that Money, while you hire one of the perpetrators.
Surely, even though you may not consider these facts related, you must understand how this looks from the outside. Let me spell it out: It looks like Caroline is continuing to benefit from the fraud by being paid by an organisation that kept the money she helped steal.
While I am thankful for the transparency and understand that you donāt want to take part in discussing this in public forum threads, I, and maybe others, want to.
I have to say, I think this is harming what remains of the Effective Altruism name/ābrand and is making me reconsider openly associating with it.
If we ever hope to rebuild the trust that was lost and construct a better public image of EA, we need to have strong community norms that people who commit wire fraud, securities fraud and money laundering have no place in any type of financial role in organisations that make up this community.
I get that you have personal ties, and I can understand wanting to act from a position of forgiveness towards a friend and uphold norms of redemption. From an outside perspective, though, these same personal ties can be perceived as more of the same mixing of professional and personal lives that infamously happened at FTX.
I, too, believe people can redeem themselves and, after completing their sentence, should be able to rejoin society. And I also think that the EA community can only ever gain a positive public image if they beat the accusations of fraud and favouritism through an abundance of care about these things and strict professional norms. Background checks for financial roles that exclude people with felony convictions for financial crimes are, in my opinion, justified.
Thank you so much for writing about this. This insight felt especially important to me:
āIf you decide, by default, that everyone in your group is self-aware, well-intentioned, and telling the truth, and you then make their intent a central factor in assessing misconduct, you will almost never find abuse.ā
So often EAs are used to everyone cooperating that they struggle to notice when someone is actually doing something bad.
Thank you for that addition. I think my post is due for an update soon. I still hold mostly the same concerns as described in the post, but by now the trade-offs between convenience and risk have shifted for many people due to
the more obvious rise of fascism
anyone having the capability with LLMs now to extremely cheaply evaluate each person whoās data they have access to
Wondering how much of the higher happiness and general nice things in Scandinavian countries can be attributed to really high taxes on alcohol.
It is possible to rationally prioritise between causes without engaging deeply on philosophical issues
Even what we think of as our terminal values may change once we examine them closer.
itās easy to be wrong about what futures are more valuable, but itās pretty clear that extinction closes off all future options of creating value.
Judging from your description and the images Ameliorate or DebateMap may come close. The name for this tool category I have encountered is either argument mapping or concept mapping. You should also have a look at Whysaurus, the way it handles contributions seems close to what you describe.
Iām also building an argument analysis tool, but itās probably not what youāre looking for, since it focuses more on identifying cruxes between two users: Fences. As part of that work, Iām also creating a page that compares all these tools, Iāll comment here once that is finished.
Thanks for writing up your thinking on this, and also more in general for doing the hard work of community building. š
I think Tess makes good points in their comment about the huge amounts of uncertainty contained in that impact potential factor, and I would go a step further and say:
The factor i is a red herring, and it is harmful for community building to try and predict it.
āintuitivelyā predicting this is extremely succeptible to bias and rationalizations
overconfident community builders using this to justify their decisions will come off as arrogant and elitist
it will increase impostor syndrome (āmy local group doesnāt have a high enough impact potentialā)
itās dependent on which ethical frameworks are correct
core EA orgs werenāt able to predict the hugely negative impact of SBF
Using it to decide who is āmost worthy of outreachā is in direct conflict with core strenghts of the EA community: cause-neutrality and openness to neglected approaches to making the world a better place.
What should be done instead?
Aim to build a broad, diverse community that is welcoming and non-judgemental and accurately communicates EA ideas to a large audience. Build a community that you would want to be part of. Does this mean abandoning quantitative measures? No, we can still measure if we achieve our goals:
measure participantās understanding of core EA ideas
measure diversity
measure people reached
measure giving pledges
measure money donated
The high impact potential people (those mythical creatures) will find their way once they have come in contact with an accurate representation of EA ideas. Itās not the job of community builders to identify and guide them (because they canāt), their job is to build an awesome community that communicates EA ideas with high fidelity so that many people come in contact with good first hand accounts of these ideas instead of simplified straw-man versions of EA in take-down articles about scandals in the EA community.
Yeah, I mainly posted it because itās good comedy. That doesnāt mean I think Give Directly should adopt it.
As inspiration, hereās a quote by Scottish stand-up comedian Frankie Boyle on how patronizing the teaching-to-fish idea is:
https://āāwww.youtube.com/āāwatch?v=4X5KmQxvO5U&t=1537s
Give a man a fish, and he can eat for a day. Give him a fishing rod, and he can feed himself. Alternatively, donāt poison the fishing waters, abduct his great-grandparents into slavery, then turn up 400 years later on your fucking gap year talking a lot of shite about fish.
A [ā¦] report linking deforestation in the Amazon to soy and beef production was published in Science [ā¦], Forbes and BBC then mistakenly announced that 1ā5 of the entire national exports were so contaminated.
Sadly this kind of misrepresentation of scientific papers or overgeneralization of their actual results seems to be a common pattern.
Have you tried to contact the authors at Forbes and BBC? Maybe writing letters to the editor could be a quick way to curb some of the harms of sub par science reporting? (at least in cases where the article misinterprets the underlying paper)
Yeah, pacing weekends could be a similar success as Anthropicās end-of-year 2025 2x usage + free invite passes promo. That kicked off a wave of developers suddenly realising how far agentic capabilities had come, because they had both the free time and the extra tokens to actually build something. So the PR/āmarketing effects of this could offset some of the costs.