DirectedEvolution

Karma: 3,582

DirectedEvolution Jan 31, 2023, 11:01 PM
2 points
0 ∶ 0
in reply to: Stavros’s comment on: Why are we not talking more about the metacrisis perspective on existential risk?
So I have a lot of questions. I’ll try to ask them one or two at a time.
It seems like you’re claiming something like this:
“Clearly, there are a bunch of emergencies, which have causes and solutions. For a lot of them, we know what the causes and solutions are, but don’t implement them. That is probably because our global institutions have big complicated effects on each other, but nobody has a very good predictive model of what the chain of cause and effect is or how to intervene in it productively. Like, if you wanted to pass a carbon tax, what would you even do?
Probably, if we studied that, we could figure out some sort of complicated way to make all these global institutions fit together way better, so that we’d be just a lot happier with life. It’s sort of like we have a medieval doctor’s understanding of how the body works, and we’d be a lot better of stopping trying to ‘treat’ most problems and start trying to study how the body works in detail. Except instead of the body, it’s ‘global institutions and culture’ and instead of medicine it’s all sorts of political/cultural/economic/scientific interventions.”
Is that roughly what you mean?

DirectedEvolution Jan 30, 2023, 8:57 PM
25 points
8 ∶ 1
on: Karma overrates some topics; resulting issues and potential solutions
I think one of the more underused features of the forum is the ability to adjust the visibility of posts according to their tags. If everybody used this feature and wanted to see fewer community-tagged posts, they could downweight their visibility. The fact that they don’t could mean that most people do want to see these posts, but it could also mean that they’re unaware of this feature, haven’t had the idea of turning it on for community posts, or just haven’t gotten around to it.
I could see you doing either of these:
- Making this feature more visible. Perhaps a hovertext option to adjust the visibility of certain tags, a pinned post explaining how to use it, etc. Edit: hovertext adjust is already a feature. What might be a nice add would be some common “adjust bundles.” For example, there are a lot of community-related tags (“criticism of EA culture”) for example, and posts aren’t reliably tagged with all applicable tags. It’s kind of tedious to manually search for and enter adjustments for each relevant tag. Having a single button to adjust all tags related to community by say 100 karma increments would be nice.
- Nudges, such as enabling this feature for new users by default, or even for all users. Potentially you could turn it on for everyone, but send everybody a PM saying what you did and how to reverse it if desired.
A subforum also seems fine, it just seems like a potentially more expensive and buggy option that adds complexity to the forum structure.
What links here?
- Lizka's comment on Karma overrates some topics; resulting issues and potential solutions by Lizka (Feb 2, 2023, 2:31 PM; 2 points)

DirectedEvolution Jan 30, 2023, 7:28 PM
2 points
1 ∶ 0
in reply to: Stavros’s comment on: Why are we not talking more about the metacrisis perspective on existential risk?
Based on the article you linked, it seems like ‘meta-crisis’ thinking employs a bundle of concepts that LessWrong often calls ‘Moloch’ or ‘simulacra levels’ or ‘inadequate equilibria’ or simply tradeoffs. This line of analysis attempts to use these ideas to explain failures of collective action to implement complex institutional change, and generate solutions to overcome this inertia.
I’m sympathetic to the need to address issues of governance and collective action. However, what interests me are clear problem-solution pairs with good evidence, a straightforward mechanism, and adequate information feedback to see if it’s working. “We should switch to approval voting” meets those criteria.
I’m less excited about interrogating “the very idea of ‘the economy’ or what exactly we mean by ‘money’” or the idea that “too much liberty may kill liberalism, too much voting can weaken democracies, and we don’t always understand how we understand, we tend to deny our denial, and we are struggling to imagine a new imaginary.”

DirectedEvolution Jan 29, 2023, 11:14 PM
3 points
0 ∶ 0
in reply to: Lukas_Gloor’s comment on: Why are we not talking more about the metacrisis perspective on existential risk?
The term “crisis” (in “metacrisis”) makes it sound like it’s something new and acute, but it seems that we’ve had coordination problems for all of history. Though maybe their effects are getting worse because of accelerating technological progress?
This would be surprising to me, since so much of tech progress is the creation of social coordination technologies (internet and social media platforms, cell phones and computers, new modes of transport, cheaper food and safer water that simplifies logistics of human movement, new institutions, new language and theories, new educational resources).
Perhaps, however, the explosion in social coordination technologies makes it easier for self-interested and antisocial groups to coordinate for private advantage at the expense of the public.
For example, one lens on American slavery might be a transition from:
- No pro-slavery social coordination infrastructure/technology
- Increasing amounts of pro-slavery coordination allow a horrific trans-Atlantic slave trade
- Gradually increasing anti-slavery coordination undermine, then finally destroy the slave trade
- A modern back and forth of coordination as a result of conflict pressures between human traffickers and anti-trafficking police, as well as among racist police institutions and individuals and anti-racist activists
In these scenarios of conflict, increased coordination technology is neither intrinsically good or bad. It’s just a tool that may give a net advantage to the good guys, bad guys, or neither, depending on context.
That being said, it may be that improving specific forms of coordination technology that help us toward betteer ends is still what we need the most. I’m very receptive to that view. Better voting mechanisms, better science distillation mechanisms, better institutions, better tax policy, better ways of dealing with the risks of specific forms of technological development, etc.
I guess I just think that insofar as these are concerns of the meta-crisisians, they’re also things we deal with pretty thoroughly in EA through our own lenses. I’d be curious to know what the concept of “meta-crisis” offers in terms of further clarifying the problems or proposing new solutions.

DirectedEvolution Jan 29, 2023, 7:09 PM
2 points
0 ∶ 0
in reply to: Stavros’s comment on: Why are we not talking more about the metacrisis perspective on existential risk?
Where can we find a solid written introduction to the concept of “meta-crisis?” I haven’t watched any of the Youtube videos and probably won’t even if they’re good, but I’d be interested in a written resource if available.

DirectedEvolution Jan 29, 2023, 3:56 PM
10 points
0 ∶ 0
on: Why are we not talking more about the metacrisis perspective on existential risk?
I tried googling it, and just found a few YouTube videos, the Consilience Project, and a few blog posts, most of which are not on “the meta crisis” but on specific issues. The lack of written content and in particular ways to define and interrogate the concept of “meta crisis” makes this feel not very compelling, but I might be missing a treasure trove somewhere!

DirectedEvolution Jan 29, 2023, 1:04 AM
10 points
0 ∶ 0
on: Excalidraw: Why and How to Use it
I bookmarked it and found a chance to use it already today! In general, I appreciate finding out about tools for rational discourse enhancement. Encouraging others to post them even if they don’t get too many upvotes.

DirectedEvolution Jan 23, 2023, 6:50 PM
24 points
7 ∶ 0
on: Finding bugs in GiveWell’s top charities
I think this post is really pretty good. What I’d like to see, either included here or in future work, is a plain-language description of your assumptions and conclusions. I tried to make one, which went something like this:
GiveWell’s global health work saves lives, but the people they save go on to destroy insect habitat. The worth of a single insect is negligible relative to that of a human, but so many of them die when forests get cut down that global health efforts in certain locations could conceivably be bad overall. This side effect would be especially bad in countries where the most deforestation per capita is going on, and where habitat for insects is particularly lush.
Based on a shallow dive and plugging some numbers into a simple model, the results suggest we should reallocate human health resources out of a few countries and into other countries where less habitat destruction is happening, or reallocate some of those resources to preserving habitat. We really need to research this more carefully before taking these conclusions too seriously—but these early results suggest it would be worth the effort.
It would also be nice to move the acronyms (i.e. t_a) to the figure captions where they’re presented, and work on better formatting the tables. You’ve done the research and thinking, so take a little more time to polish up the presentation so we can read it more easily :)

DirectedEvolution Jan 20, 2023, 1:09 AM
6 points
4 ∶ 0
on: Announcing Cavendish Labs
Best of luck to you! I’d be interested in hearing in more depth about your team, specific project ideas, funding situation if you care to share. My particular interest is in the pandemic prevention/far UVC work.

DirectedEvolution Jan 18, 2023, 7:43 PM
10 points
2 ∶ 0
in reply to: ConcernedEAs’s comment on: Doing EA Better
OOOOHHHHHHHH

DirectedEvolution Jan 18, 2023, 5:23 PM
7 points
1 ∶ 1
in reply to: Guy Raveh’s comment on: Doing EA Better
That is a fair rebuttal.
I would come back to the model of a value-aligned group with a specific set of tasks seeking to maximize its effectiveness at achieving the objective. This is the basis for the collective intelligence research that is cited here as the basis for their recommendations for greater diversity.
If you frame EA as a single group trying to achieve the task of “make the entire world better for all human beings by implementing high-leverage interventions” then it does seem relevant to get input from a diverse cross-section of humanity about what they consider to be their biggest problems and how proposed solutions would play out.
One way to get that feedback is to directly include a demographically representative sample of humanity in EA directly as active participants. I have no problem with that outcome. I just think we can ⁸⁰⁄₂₀ it by seeking feedback on specific proposals.
I also think that basing our decisions about what to pursue based on the personal opinions of a representative sample of humanity will lead us to prioritize the selfish small issues of a powerful majority over the enormous issues faced by underrepresented minorities, such as animals, the global poor, and the denizens of the far future. I think this because I think that the vast majority of humanity is not value-aligned with the principle of altruistic utility maximization.
For these two main reasons—the ability to seek feedback from relevant demographics when necessary, and the value mismatch between EA and humanity in general—I do not see the case for us being unable to operate effectively given our current demographic makeup. I do think that additional diversity might help. I just think that it is one of a range of interventions, it’s not obvious to me that it’s the most pressing priority, and broadening EA risks to pursue diversity purely for its own sake risks value misalignment with newcomers. Please interpret this in a moderate stance along the lines of “I invite diversity, I just think it’s not the magic solution to fix all of EA’s problems with effectiveness and the important thing is ‘who does EA talk to’ more than ‘who calls themselves an EA’.”

DirectedEvolution Jan 18, 2023, 5:09 PM
3 points
0 ∶ 0
in reply to: ConcernedEAs’s comment on: Doing EA Better
This is the phrase where you introduce the Yang & Sandberg link:
The field of Collective Intelligence provides guidance on the traits to nurture if one wishes to build a collectively intelligent community. For example:
The word before the link is “community,” which does not contain a link.

DirectedEvolution Jan 18, 2023, 1:19 PM
4 points
1 ∶ 0
in reply to: ConcernedEAs’s comment on: Doing EA Better
Where exactly is that link? I looked at the rest of the links in the section and don’t see it.

DirectedEvolution Jan 18, 2023, 5:45 AM
5 points
0 ∶ 0
in reply to: Linda Linsefors’s comment on: Doing EA Better
I think he’s arguing that you should have a little “fire alarm” in your head for when you’re regurgitating a narrative. Even if it’s 95% correct, that act of regurgitation is a time when you’re thinking less critically and it’s a perfect opportunity for error to slip through. Catching those errors has sufficiently high value that it’s worth taking the time to stop and assess, even if 19 out of 20 times you decide your first thought was correct.

DirectedEvolution Jan 18, 2023, 3:48 AM
2 points
0 ∶ 0
in reply to: Linch’s comment on: Doing EA Better
I agree that this would have been better, but it was the example that came to mind and I’m going to trust readers to take it as a loose analogy, not a claim about which side is correct in the debate.

DirectedEvolution Jan 18, 2023, 2:16 AM
33 points
18 ∶ 8
in reply to: GideonF’s comment on: Doing EA Better
My interpretation of Buck’s comment is that he’s saying that, insofar as he’s read the post, he sees that it’s largely full of ideas that he’s specifically considered and dismissed in the past, although he is not confident that he’s correct in every particular.
I think that the class of arguments in this post deserve to be considered carefully, but I’m personally fine with having considered them in the past and decided that I’m unpersuaded by them...
there are probably some suggestions made in this post that I would overall agree should be prioritized if I spent more time thinking about them
You want him to explain why he dismissed them in the past
I’d be pretty interested in you laying out in depth why you have basically decides to dismiss these very varied and large set of arguments.
And are confused about why he’d encourage other people to champion the ideas he disagrees with
why they would want to suggest to these bunch of concerned EAs how to go about trying to push for the ideas that Buck disagrees with better
I think the explanation is that Buck is pretty pessimistic that these are by and large good ideas, enough not to commit more of his time to considering each one individually more than he has in the past. However, he sees that the authors are thinking about them a lot right now, and is inviting them to compete or collaborate effectively—to put these ideas to a real test of persuasion and execution. That seems far from “poo-poohing” to me. It’s a piece of thoughtful corrective feedback.
You have asked Buck to “lay out in depth” his reasons for rejecting all the content in this post. That seems like a big ask to me, particularly given that he does not think they are good ideas. It would be like asking an evolutionary biologist to “lay out in depth” their reasons for rejecting all the arguments in Of Pandas and People. Or, for a personal example, I went to the AAAS conference right before COVID hit, and got to enjoy the spectacle of a climate change denier getting up and asking in front of the ballroom whether the geoengineering scientist who’d been speaking whether scientists had considered the possibility that the Earth is warming up because it’s getting closer to the sun. His response was “YES WE’VE CONSIDERED IT.”
If that question asker went home, wrote a whole book full of reasons why the Earth might be moving closer to the sun, posted it online, and it got a bunch of upvotes, I don’t think that means that suddenly the scientist needs to consider all of the arguments more closely, revisit the issue, or that rejecting the ideas gives one an obligation to explain all of one’s reasons.
One way you could address this problem is by choosing one specific argument from this post that you find most compelling, and seeing if you can invite Buck into a debate on that topic, or to explain his thinking on it. I often find that to be productive of good conversation. But your comment read to me as an attempt to both mischaracterize the tone of Buck’s comment and and call into question the degree to which he’s thought about these issues. If you are accusing him of not actually having given these ideas as much thought as he claims, I think you should come right out and say it.

DirectedEvolution Jan 18, 2023, 1:51 AM
23 points
7 ∶ 0
in reply to: Nathan Young’s comment on: Doing EA Better
My take is different. As a working scientist/engineer in the hard sciences, I use peer-reviewed research when possible, but I temper that with phone calls to companies, emails to other labs, posts on ResearchGate, informal conversations with colleagues, and of course my own experiments, mechanistic models, and critical thinking skills. Peer-reviewed research is nearly always my starting point because it’s typically more information and data-rich and specific than blog posts, and because the information I need is more often contained within peer-reviewed research than in blog posts.
That said, there are a lot of issues and concerns raised when blog posts are too heavily a source (although here that’s very much the pot calling the kettle black, with most of the footnotes being the author’s own unsourced personal opinions). When people lean too heavily on blog posts, it may illustrate that they’re unfamiliar with the scientific literature relevant to the issue, and that they themselves have mostly learned about the information by consuming other blog posts. Also, a compelling post that’s full of blog post links (or worse, unsourced claims) gives the interested reader little opportunity to check the underpinnings of the argument or get connected with working scientists in the field.
I’m fine with using the medium of blog posts to convey an idea, or of citing blog posts in specific circumstances. Where a peer-reviewed source is available, I think it’s better to either use that, or to cite it and give the blog post as an accessible alternative.

DirectedEvolution Jan 18, 2023, 1:42 AM
22 points
11 ∶ 2
in reply to: freedomandutility’s comment on: Doing EA Better
I would like to see no change or a slight increase in our amount of overall tone policing. There might be specific forms of tone policing we should have less of, but in general I think civil discourse is one of the main things that make EA functional as a movement.

DirectedEvolution Jan 18, 2023, 1:37 AM
57 points
22 ∶ 2
on: Doing EA Better
I did a close read of “Epistemic health is a community issue.” The part I think is most important that you’re underemphasizing is that, according to the source you cite, “The diversity referred to here is diversity in knowledge and cognitive models,” not, as you have written, diversity “across essentially all dimensions.” In other words, for collective intelligence, we need to pick people with diverse knowledge and cognitive models relevant to the task at hand, such as having relevant but distinct professional backgrounds. For example, if you’re designing a better malaria net, you might want both a materials scientist and an epidemiologist, not two materials scientists.
Age and cultural background might be relevant in some cases, but that really depends on what you’re working on any why these demographic categories seem especially pertinent to the task at hand. If I was designing a nursing home in a team comprised of young entrepreneurs, I would want old people either to be on the team, or to be consulted with routinely as the project evolved, because adding that component of diversity would be relevant ot the project. If I was developing a team to deploy bed nets in Africa, I might want to work with people from the specific villages where they will be distributed.
Worryingly, EA institutions seem to select against diversity. Hiring and funding practices often select for highly value-aligned yet inexperienced individuals over outgroup experts, university recruitment drives are deliberately targeted at the Sam Demographic (at least by proxy) and EA organisations are advised to maintain a high level of internal value-alignment to maximise operational efficiency.
As your own source says:
For a group of diverse agents to work together, an important aspect is cognitive alignment, such as commitment to group goals and shared beliefs.
EA institutions are struggling to find people who are both value-aligned and have relevant diverse cognitive models and knowledge, both of which are prerequisites for collective intelligence. This is a natural problem for EA institutions to have, so you need to explain why EA is making this tradeoff of value alignment vs. cognitive/knowledge diversity suboptimally for your critique to bite.
The 80,000 Hours website seems purpose-written for Sam, and is noticeably uninterested in people with humanities or social sciences backgrounds,^[14] or those without university education. Unconscious bias is also likely to play a role here – it does everywhere else.
Where your post and the collective intelligence research you’re basing it off seem to diverge is that while you want EA to select for diversity across all dimensions, perhaps for its own sake, the CI research you cited argues that you need to select for forms of cognitive models and knowledge relevant to the task at hand. 80,000 Hours might be wrong in ignoring humanities and social science backgrounds, or those without a university education, but I think your argument would be much stronger if you articulated something specific about what those disciplines would bring to the table.
Whole disciplines exist that are overwhelmingly not value-aligned with EA. Making an effort to include them seems to me like it would add a skillset of questionable task-relevant value while creating fundamental and destructive internal conflict. Because EA does have a specific set of theses on what constitutes “effectiveness” and “altruism,” which we can try to define in the abstract but which can perhaps be better articulated by the concrete types of interventions we tend to support, such as global health and X risk issues. Not everybody supports those kinds of projects, or at least not prioritizing them as highly as we do, or using the kinds of underlying models we use to prioritize them (i.e. the ITN framework), and if they are that far removed from the mission of our movement, then we should probably not try to include them in it.

DirectedEvolution Jan 17, 2023, 4:49 PM
22 points
4 ∶ 0
in reply to: Kelsey Piper’s comment on: AGI and the EMH: markets are not expecting aligned or unaligned AI in the next 30 years
Here’s what I wrote in April 2020 on that topic:

“ A couple weeks ago, I started investigating the response, here and in the stock market, to COVID-19. I found that LessWrong’s conversation took off about a week after the stock market started to crash. Given what we knew about COVID-19 prior to Feb. 20th, when the market first started to decline, I felt that the stock market’s reaction was delayed. And of course, there’d been plenty of criticism of the response of experts and governments. But I was playing catch-up. I certainly was not screaming about COVID-19 until well after that time.

Today, I found the most detailed timeline I’ve seen of confirmed cases around the world. It goes day by day and country by country, from Jan. 13th to the end of March.

That timeline shows that Feb. 21st was the first date when at least 3 countries besides China had 10+ new confirmed cases in a single day (Japan, South Korea, Italy, and Iran).

That changes my interpretation of the stock market crash dramatically. Investors weren’t failing to synthesize the early information or waiting for someone to yell “fire!” They were waiting to see confirmed international community spread, rather than just a few cases popping up here and there. Once they saw that early evidence, the sell-off began, and it continued in tandem, day by day, with the evidence of community spread in new countries and the exponential growth of COVID-19 cases in countries where it was already established.”

https://www.lesswrong.com/posts/EdJD3v2uxLrL3pA75/my-stumble-on-covid-19