What posts are you thinking about writing?

Fascinating, I've never heard of this before, thanks! If anyone's curious, I had Deep Research [take a stab at writing this] (https://chatgpt.com/share/67ac150e-ac90-800a-9f49-f02489dee8d0) which I found pretty interesting (but have totally not fact checked for accuracy)

I would love to see this. Not a take I've seen before (that I remember). 

This sounds like it could be interesting, though I'd also consider if some of the points are fundamentally to do with RCTs. E.g., "80% statistical power meaning 20% chance of missing real effects" - nothing inherently says an RCT should only be powered at 80% or that the approach should even be one of null hypothesis significance testing.

I would welcome a blog post about RCTs, and if you decide to write one, I hope you consider the perspective below.

As far as I can tell ~0% of nonprofits are interested in rigorously studying their programs in any way, RCTs or otherwise, and I can't help but suspect that this is largely because mostly when we do run RCTs we find that these cherished programs have ~no effect. It's not at all surprising to me that most charities that conduct RCTs feel pressured to do so by donors; but on the other hand basically all charity activities ultimately flow from donor preferences, because donors are the ones with most of the power.

Living Goods is one interesting example, where they ran an RCT because a donor demanded it, got an unexpected (positive) result, and basically pivoted the whole charity based on that. I view that as a success story.

I am certainly not claiming that RCTs are appropriate for all kinds of programs, or some kind of silver bullet. It's more like, if you ask charities "would you like more or less accountability for results", the answer is almost always going to be, "less, thanks".

Would also love this. I think a useful contrast will be A/B testing in big tech firms. My amateur understanding is big tech firms can and should run hundreds of “RCTs” because:

• No need to acquire subjects.
• Minimal disruption to business since you only need to siphon off a minuscule portion of your user base.
• Tech experiments can finish in days while field experiments need at least a few weeks and sometimes years.
• If we assume treatments are heavy tailed, then a big tech firm running hundreds of A/B tests is more likely to learn of a weird trick that grows the business when compared to a NGO who may only get one shot. 

Personally I'd be so keen on seeing that - it's part of the pitch that I make to authors. 

I would very much like to see something like this. Being able to communicate EA ideas to people that are roughly aligned in terms of many altruistic values is useful.

Very much in favor of posts clarifying that cause neutrality doesn't require value neutrality or deference to others' values.

This would be great to read, I walked away from at least one application process because I couldn't produce a decent WFM. I hope you write it!

Nice. I'd find this super interesting!

Might be a nice candidate for a bullet-point outline draft amnesty post (like this one)? There's no rule that you can't republish it as a full post later on, and perhaps you could get some feedback/ ideas from the comments on a draft amnesty post...

Sounds great, and always happy to give feedback :)

I do want to write something along the lines of "Alignment is a Political Philosophy Problem"

My takes on AI, and the problem of x-risk, have been in flux over the last 1.5 years, but they do seem to be more and focused on the idea of power and politics, as opposed to finding a mythical 'correct' utility function for a hypothesised superintelligence. Making TAI/AGI/ASI go well therefore falls in the reference class of 'principal agent problem'/'public choice theory'/'social contract theory' rather than 'timeless decision theory/coherent extrapolated volition'. The latter 2 are poor answers to an incorrect framing of the question.

Writing that influenced my on this journey:

• Tan Zhi Xuan's whole work, especially Beyond Preferences in AI Alignment
• Joe Carlsmith's Otherness and control in the age of AGI sequence
• Matthew Barnett's various posts on AI recently, especially viewing it as an 'institutional design' problem
• Nora Belrose's various posts on scepticism of the case for AI Safety, and even on Safety policy proposals conditional on the first case being right.^[1]
• The recent Gradual Disempowerment post is something along the lines I'm thinking of too

I also think this view helps explain the huge range of backlash that AI Safety received over SB1047 and after the awfully botched OpenAI board coup. They were both attempted exercises in political power, and the pushback often came criticising this instead of looking on the 'object level' of risk arguments. I increasingly think that this is not an 'irrational' response but perfectly thing, and "AI Safety" needs to pursue more co-operative strategies that credibly signal legitimacy.

1. ^{^}

   I think the downvotes these got are, in retrospect, a poor sign for epistemic health

I don't think anyone wants or needs another "Why I'm leaving EA" post but I suppose if people really wanted to hear it I could write it up. I'm not sure I have anything new or super insightful to share on the topic.

I have some initial data on the popularity and public/elite perception of EA that I wanted to write into a full post, something along the lines of What is EA's reputation, 2.5 years after FTX? I might combine my old idea of a Forum data analytics update into this one to save time.

My initial data/investigation into this question ended being a lot more negative than other surveys of EA. The main takeaways are:

• Declining use of the Forum, both in total and amongst influential EAs
• EA has a very poor reputation in the public intellectual sphere, especially on Twitter
• Many previously highly engaged/talented users quietly leaving the movement
• An increasing philosophical pushback to the tenets of EA, especially from the New/Alt/Tech Right, instead of the more common 'the ideas are right, but the movement is wrong in practice'^[1]
• An increasing rift between Rationalism/LW and EA
• Lack of a compelling 'fightback' from EA leaders or institutions

Doing this research did contribute to me being a lot more gloomy about the state of EA, but I think I do want to write this one up to make the knowledge more public, and allow people to poke flaws in it if possible.

1. ^{^}

   To me this signals more values-based conflict, which makes it harder to find pareto-improving ways to co-operate with other groups

Cosmological Fine-Tuning Considered:

The title’s kind of self-explanatory – over time I’ve noticed the cosmological fine-tuning argument for the existence of god become something like the most favored argument, and learning more about it over time has made me consider it more formidable than I used to think as well.

I’m ultimately not convinced, but I do consider it an update, and it makes for a good excuse for me to talk more about my views on things like anthropic arguments, outcome pumps, the metaphysics of multiverses, and interesting philosophical considerations more specific to this debate – I might particularly interact with statements by Phillip Goff on this subject.

Unfortunately if this sounds like a handful, it is, and I got bogged down early in writing it during the anthropics section. This might be a good time to get more feedback from people with more metaphysics/epistemology under their belts than me, and maybe finally to get a solid idea of the difference between self-indicating and self-selecting anthropic assumptions and which anthropic arguments rely on each. I don’t have much of this to post, so I might either do this as an outline, the small portion I do have, or some combination.

Topic from last round:

Okay, so, this is kind of a catch all. Out of the possible post ideas I commented last year, I never posted or wrote “Against National Special Obligation”, “The Case for Pluralist Evaluation”, or “Existentialist Currents in Pawn Hearts”. So, this is just the comment for “one of those”.

Observations on Alcoholism Appendix G:

This would be another addition to my Sequence on Alcoholism – I’ve been thinking in particular of writing a post listing out ideas about coping strategies/things to visualize to help with sobriety. I mention several in earlier appendices in the sequence – things like leaning into your laziness or naming and yelling at your addiction – but I don’t have a neat collection of advice like this, which seems like one of the more useful things I could put together on this subject.

Mid-Realist Ethics:

I occasionally bring up my meta-ethical views in blog posts, but I keep saying I’ll write a more dedicated post on the topic and never really do. A high level summary includes stuff like: “ethics” as I mean it has a ton of features that “real” stuff has, but it lacks the crucial bit which is actually being a real thing. The ways around this tend to fall into one of two major traps – either making a specific unlikely empirical prediction about the view, or labeling a specific procedure “ethics” in a way that has no satisfying difference from just stating your normative ethics view - and a couple thought experiments make me unpersuaded that I’m really interested in a realist view anyway. I discuss these things a bit in this comments thread.

I would also plan to talk about the role different kinds of intuitions play in both my ethical reasoning and in my “unethical” reasoning, something I keep mentioning but not developing in blog posts, especially these two. I don’t really have anything written for this, so I might just collect snippets from these sources and supplement with bullet-point type additions if I go with this idea.

Moral problems for environmental restoration:

A post idea I’ve been playing with recently is converting part of my practicum write-up into a blog post about the ethics of environmental restoration projects. My practicum was with the “Billion Oyster Project”, which seeks to use oyster repopulation for geoengineering/ecosystem restoration, and I spent a big chunk of my write-up worrying about the environmental ethics of this, and I’ve been thinking this worrying could be turned into a decent blogpost.

I’ll discuss welfare biology briefly, but lots of it will survey non-consequentialist possibilities, like “does non-aggression animal ethics bar us from restoration?”, “if we care about the ecosystem as a moral patient, what does it take for restoration to be creating a new patient versus aiding an existing one?”, or “does creating a new ecosystem burden us with new special obligations, and are they obligations we can actually fulfill?”

I already have a substantial amount written for this, just because I have that section of my practicum write-up already, but it’s currently a bit rough and I might modify it to be more general than just the billion oyster case, or even to expand it in the direction of discussing the ethics of terraforming other planets. This is the one I am most leaning towards posting just because I am most likely to have a substantial amount of writing done for it on time.

(See here for a draft I whipped up for this, and feel free to comment!) Hayden Wilkinson’s “In defence of fanaticism” argues that you should always take the lower-probability odds of a higher-value reward over the inverse in decision theory, or face serious problems. I think accepting his argument introduces new problems that aren’t described in the paper:

1. It is implied that each round of Dyson’s Wager (e.g. for each person in the population being presented with the wager) has no subsequent effect on the probability distribution for future rounds, which is unrealistic and doesn’t. I illustrate this with a “small worlds” example.
2. Fanaticism is only considered under positive theories of value and therefore ignores the offsetting principle, which assumes both the existence of and an exchange rate (or commensurability) between independent goods and bads. I’d like to address this in a future draft with multiple reframings of Dyson’s Wager under minimalist theories of value.

(See here for a draft I whipped for this, and feel free to comment!) An Earth-originating artificial superintelligence (ASI) may reason that the galaxy is busy in expectation, and that it could therefore eventually encounter an alien-originating ASI. ASIs from different homeworlds may find it valuable on first contact to verify whether they can each reliably enter into and uphold agreements, by presenting credible evidence of their own pro-social behaviour with other intelligences. If at least one of these ASIs has never met another, the only such agreement it could plausibly have entered into is with its progenitor species – maybe that's us.