Hide table of contents

Epistemic status: These are exactly part of the non-performative speech I talk about here. Think about what value these actually add, if I put little effort into the post, why should you trust me saying otherwise? I might have pulled this out of my ass and claim to have spent a year thinking about these things, or I might have spent a year thinking about it and still be uncertain. I obviously think it's worth reading and believe in it. But you need to skim and decide yourself.


I'm writing this as a short awareness raiser of a few concepts that were touched upon in a recent PhilosophyTube Video (see below) which explores the concepts of System Design through the lens of transgender treatment in the NHS. I recommend it for anyone who would like to really grok what the rest of this post is talking about, as Abigail has a knack for Brechtian conveyance of concepts beyond mere definitions.

The TL;DR; I'd like you to take away from this post and the video are

  • Sara Ahmeds "On Complaint" and other work on how institutions Jiu-Jutsu away real accountability and change and Stafford Beers "Designing Freedom" and related work on System Design and interpretation are worth reading
  • the past and current responses to criticism and calls for accountability have, in my eyes, been mainly met with non-performative speech and served, to paraphrase Ahmed "to do the document, not do the doing"
  • In particular, Williams,  Holdens and others initial responses are, as of now, mainly non-performative speech. THIS IS FINE, I would not trust or recommend committing to rash actions now and I don't want to shit on them as people. But I want to caution against leaving things at that, and I do want to point out towards them as powerful agents of a culture.
  • Whistleblower systems, auditing and legal aid as has been discussed in https://forum.effectivealtruism.org/posts/HyHCkK3aDsfY95MoD/cea-ev-op-rp-should-engage-an-independent-investigator-to and the cited posts are nice, but should not be the end. A culture of accountability starts with documenting and taking responsibility in the task of uncertainty and handing real power to people even if it means they can use it against you when you think you haven't done wrong. Whistleblower systems can be a part of this, but changing the power dynamics by, for some non-exhaustive examples:
    • changing the ways grants are paid out to escrow systems of hard cash (as opposed to apparently just committing to pay for people using an endowment of crypto monopoly money if reports of the FFF grants being dropped are true?)
    •  creating a system in which the BFDL style leadership of EA can be replaced by the community (does not need to be immediate, but as of now, if the trusted voices of EA all get COVID-22 and die or turn out to have been in cahoots with SBFs alleged fraud, what exactly would be the orderly "order of succession" or replacement procedure?) 
    • clearly disclosing, in detail, money flow and organisational team overlap within the EA ecosystem in regular intervals and generally inviting  "hostile" journalists to do their thing instead of the defensiveness I generally see

I'll expound on these points a bit more below the video.


What complaints and criticism reveal about institution

Abigail uses her experience of fighting through the complain-swamp of the NHS to illustrate the institutional techniques of 

  • expressing concern (and doing nothing)
  • launching investigations (which bear no consequences)
  • create official avenues of complaints (with no guaranteed consequences)
  • promising guarantees (with no possibility to collect on them)
  • individualising and focusing on personal empathy and trust (while leaving a broken system in place).

In the EA world, in my experience, this looks like the following:

  1. Somebody raises a concern about EA culture
  2. The community and often the criticised listens, takes it seriously, discusses it, empathises with the author
  3. Some reasonable counter arguments are made, like the fact that EA is a young movement (it's been over 10 years since the founding of 80k hours though), that there are "no leaders in EA" and "every EA community is different of course" (both of which are a bit questionable given the existence of CEA and the generally very clear lines of deference and importance) 
  4. Some solution that fits the culture is found (like adding this to cause prioritization, or launching new contests etc.)
  5. Nothing is really changed

One example could be the discussion after "democratizing risk" which might(?) to my non-insider eyes have influenced the red-team exercises and was widely debated....but only in a theoretical manner. Like this post https://forum.effectivealtruism.org/posts/sEpWkCvvJfoEbhnsd/the-ftx-crisis-highlights-a-deeper-cultural-problem-within and some of the comments note,  EA tends to have  aversion to putting institutional safeguards into place and prefers to run things on a trust basis. What all of the responses to crisis and continued existence of bad practices highlight is that there are no real attempts to do anything and instead contests are held, which are nice, but in the end are...contests. Games held at the discretion of the same culture, to cherry-pick whatever is convenient to appear acting, while not really changing anything. And to call out one person and organisation in particular, I found it quite rich that EY derided the red teaming contest and hinted at his own rich set of EA criticism while continually profiting of a system of intransparency, hype and lack of accountability to anyone with MIRI.

On the purpose of institutions

Which serves as a nice segue into the next topic that Abigail touches on in the video: what is the real purpose of institutions? She discusses Stafford Beers perspective of the purpose of an institution being what an institution does. MIRI, to continue the example, provides cushy job with little to no tangible output or interaction with the wider research community to its members - the output of the group would fail any institutional review.  Given this and the fact that to date there is not even a clear definition of the problem the group is meant to solve published in a way that has found acceptance or critique in the wider research community (outside of hype machines like OpenAI and Anthropic where OH HEY LOOK, SBF lead a 500e6$ series B round just a while ago), applying the principle would then hint that until we see further evidence, the point of MIRI is exactly this job providing. 

Now, what does EA do? And here I think, EA (in a wider sense) does a lot: if we count GiveWell, the communities actual giving preferences etc. it provides a lot of money flowing to worthwhile causes - like the LIONS club, the rotary club or similar humanitarian social associations. And I think that is a perfectly fine thing to be - the next elites social club, where you can network and find business partners while doing something for the community. Perfectly fine. EA in the narrower sense (i.e. the longtermist and more distinct work) is a bit different though. There, the main focus appears to be...growing the movement, educating and developing the culture and doing ever more cause prioritization...to the degree of losing touch with actually doing much that the broader world would consider impactful, and being under scrutiny even before the FTX topic broke. I served as a facilitator for an impact seminar, and a good chunk of my group mirrored the criticisms lanced in Emile Torres https://www.currentaffairs.org/2021/07/the-dangerous-ideas-of-longtermism-and-existential-risk  article. If EA/longtermism isn't  a utilitarian justification for continued wealth inequality,focus on growth now at the cost of seriously addressing climate change and general techno-optimism, then it's doing a poor job at communicating this. Saying EA isn't utilitarian as Will did in his twitter thread isn't worth much if the institutional actions and systems aren't in place to prove this through action. If the world keeps thinking you are a basic utilitarian who will, to paraphrase SBF "bet the earth for a 51% chance on a second one", then maybe there's something in your reasoning and actions that makes them think that.

Non-performative speech in responses so far

Which brings me to pointing out non-performative speech in three posts that triggered me into writing this:

First of all, I want to make clear: I'm not shitting on any of these people. I think it's good of will to say some personal stuff even while not having processed enough for an official statement. I think it's good that somebody spells out "fraud is bad actually" for those that need it, and evhub says all the right things. And I welcome people expressing emotions - including disgust. And I fully expect all three of the people I will directly speak towards to already be thinking about what I will say - I assume they are smart people, nice people, and I don't want to attack them.

What this section is about that these personal statements are not enough and a call to the  community to not stop at accepting or performing personal statements, denouncing fraud and feeling disgust. Instead, we should demand to actually  do something and do so ourselves.

 If we check 80khours "about us" section, we find the "our mistakes section, a favourite symbol of good EAs - but what I'm missing is mission statements, legal charter, some form of commitment to an accountability process. 

Likewise on effective ventures website: no charter linked, nothing about how we can trust them through institutional mechanisms.  I hope Rob feels disgusted enough to not just "trust less across the board" but to use his position to lobby for institutional safeguards and accountability

Same for Will and Evan: you hold influence in your respective institutions (especially Will). I hope you (and the community) aren't just satisfied with feeling personally betrayed and your communities and institutions threatened by it, I hope you will actually do things to change the system that enabled this.

Evan, while I can understand how nice it is not to work under deadline pressure and constant scrutiny the extremely opaque and doomsday-culty way that MIRI, OpenAI and it's recent imitators operate by are part of the problem.

As for Will, I hope that it won't take you months to realize that the guidelines, principles etc. you link are completely empty of mechanisms to enforce and be verified. That whether you think SBF acted withing EA principles or not, there was nothing to seriously challenge him: he was a rich dude doing "philantropic" things and for those with capital (be it social or otherwise), there are no control mechanisms being recommended by anything you cited in that thread, nor in any of the EA material I read in the years I've hung out with my local EA group (the call for a better governance norm post goes into more detail here).

I have been in sporting associations, student associations and open source  collectives which were much younger and smaller than EA but had more of these structures - formalized ways of distributing decision making power, of deciding who may or may not associate themselves with the organisation or movement, of electing their official and in-official representatives and leaders and  removing bad actors and of keeping track of provenance and availability of committed resources. This stuff isn't rocket science, and for a group that values introspection and carefulness as much as EA/longtermists, these things need to be developed.

Methods of accountability: doing the doing

Which brings us to the final section: how can we avoid non-performative speech? How can we avoid a big debate on this forum, or in the halls of Trajan House, that results in a firm letter  opposing doing bad in the name of EA and urging, URGING members of the community to please not do bad things?

Well, I'm biased and a single person, but if you want my opinion (and you made it through the post so I guess you either do or are collecting arguments for criticism [that I welcome]), the most important things are

  • try to diffuse the soft powers that come with the current "old boys" structure of EA: 
    • put money in escrow accounts and be transparent about funding source and reliability,
    • in fact, make all money flows public as a rule, 
    • where applicable, publish official charters which have some legal heft to them, 
    • make the structures of EA more transparent and remove the conflicts of interest that comes from everyone being advisor or board member of one-anothers NGO,
  • beyond that, dismantle the old boys structure:
    • develop methods of formal ascension through the ranks of EA (with short cuts, you can make this flexible! but it needs to be formalised so there is a tack record)
    • democratise said ranks! I see very few meeting minutes, official protocols, votes etc. published for EA organisations. These things have a purpose
  • open up to the public, invite "hostile" journalists to do critical,investigative journalism and reports on you, put out official stances on things that polarize and make it possible to slot EA onto the spectrum of things - and that can make it possible to say that "SBF wasn't acting on EA principles" when, as of right now, the media responses I've read are not buying that.


End note

I want to emphasize: I've been hanging out with EAs for most of my adult life now. Most of you all are lovely people. Evan, Will, Rob, despite being singled out in this post are lovely people trying their hardest. But the one thing that drives me insane is the tendency to pontificate and philosophize about accountability, when most of the charities and associations are not run by their communities, and make no effort to encourage to be run by their communities. Even some Protestants and Catholics have figured out how to do democratic self governance by now by voting in their dioceses and equivalent communities and some are even starting to put accountability and prevention mechanisms into their system. I like the good parts of EA too much and respect the intelligence of the people in this movement too  to accept "it's a young movement" as an excuse to do worse than these century old, conservative institutions.

"Doing Good Better" - well, then let's do Good, better.


More posts like this