Lizka

Senior Content Specialist @ Centre for Effective Altruism
15735 karmaJoined Nov 2019Working (0-5 years)

Bio

I run the non-engineering side of the EA Forum (this platform), run the EA Newsletter, and work on some other content-related tasks at CEA. Please feel free to reach out! You can email me. [More about my job.]

Some of my favorite of my own posts:

I finished my undergraduate studies with a double major in mathematics and comparative literature in 2021. I was a research fellow at Rethink Priorities in the summer of 2021 and was then hired by the Events Team at CEA. I've since switched to the Online Team. In the past, I've also done some (math) research and worked at Canada/USA Mathcamp.

Some links I think people should see more frequently:

Sequences
10

Celebrating Benjamin Lay (1682 - 1759)
Donation Debate Week (Giving Season 2023)
Marginal Funding Week (Giving Season 2023)
Effective giving spotlight - classic posts
Selected Forum posts (Lizka)
Classic posts (from the Forum Digest)
Forum updates and new features
Winners of the Creative Writing Contest
Winners of the First Decade Review
Load more (9/10)

Comments
530

Topic contributions
251

Lizka
2d11
0
0
4

I'll commit to posting a couple of drafts. Y'all can look at me with disapproval (or downvote this comment) if I fail to share two posts during Draft Amnesty Week. 

Answer by LizkaFeb 28, 20249
1
0

I'm basically always interested in potential lessons for EA/EA-related projects from various social movements/fields/projects.

Note that you can find existing research that hasn't been discussed (much) on the Forum and link-post it (I bet there's a lot of useful stuff out there), maybe with some notes on your takeaways. 

Example movements/fields/topics: 

  • Environmentalism — I've heard people bring up the environmentalist/climate movement a bunch in informal discussions as an example for various hypotheses, including "movements splinter/develop highly counterproductive & influential factions" or "movements can get widespread interest and make policy progress" etc. 
  • The effectiveness of protest — I'm interested in more research/work on this (see e.g. this and this).
  • Modern academia (maybe specific fields) — seems like there are probably various successes/failures/ideas we could learn from. 
  • Animal welfare
  • Mohism (see also)
  • Medicine/psychology in different time periods

Some resources, examples, etc. (not exhaustive or even a coherent category): 

Answer by LizkaFeb 28, 20242
1
0

I'd love to see two types of posts that were already requested in the last version of this thread:

  • From Aaron: "More journalistic articles about EA projects. [...] Telling an interesting story about the work of a person/organization, while mixing in the origin story, interesting details about the people involved, photos, etc."
  • From Ben: "More accessible summaries of technical work." (I might share some ideas for technical work I'd love to see summarized later.)

I really like this post and am curating it (I might be biased in my assessment, but I endorse it and Toby can't curate his own post). 

A personal note: the opportunity framing has never quite resonated with me (neither has the "joy in righteousness" framing), but I don't think I can articulate what does motivate me. Some of my motivations end up routing through something ~social. For instance, one (quite imperfect, I think!) approach I take[1] is to imagine some people (sometimes fictional or historical) I respect and feel a strong urge to be the kind of person they would respect or understand; I want to be able to look them in the eye and say that I did what I could and what I thought was right. (Another thing I do is try to surround myself with people[2] I'm happy to become more similar to, because I think I will often end up seeking their approval at least a bit, whether I endorse doing it or not.)

I also want to highlight a couple of related things: 

  1. "Staring into the abyss as a core life skill"
    1. "Recently I’ve been thinking about how all my favorite people are great at a skill I’ve labeled in my head as “staring into the abyss.” 
      Staring into the abyss means thinking reasonably about things that are uncomfortable to contemplate, like arguments against your religious beliefs, or in favor of breaking up with your partner. It’s common to procrastinate on thinking hard about these things because it might require you to acknowledge that you were very wrong about something in the past, and perhaps wasted a bunch of time based on that (e.g. dating the wrong person or praying to the wrong god)."
    2. (The post discusses how we could get better at the skill.)
  2. I like this line from Benjamin Lay's book: "For custom in sin hides, covers, as it were takes away the guilt of sin." It feels relevant.
  1. ^

    both explicitly/on purpose (sometimes) and often accidentally/implicitly (I don't notice that I've started thinking about whether I could face Lay or Karel Capek or whoever else until later, when I find myself reflecting on it)

  2. ^

    I'm mostly talking about something like my social circle, but I also find this holds for fictional characters, people I follow online, etc. 

Thanks for sharing this! I'm going to use this thread as a chance to flag some other recent updates (no particular order or selection criteria — just what I've recently thought was notable or recently mentioned to people): 

  1. California proposes sweeping safety measure for AI — State Sen. Scott Wiener wants to require companies to run safety tests before deploying AI models. (link goes to "Politico Pro"; I only see the top half)
    1. Here's also Senator Scott Wiener's Twitter thread on the topic (note the endorsements)
    2. See also the California effect
  2. Trump: AI ‘maybe the most dangerous thing out there’ (seems mostly focused on voting-related robocalls/deepfakes and digital currency)
  3. Jacobin publishes an article on AI existential risk (Twitter)

I don't actually think you need to retract your comment — most of the teams they used did have (at least some) biological expertise, and it's really unclear how much info the addition of the crimson cells adds. (You could add a note saying that they did try to evaluate this with the additional of two crimson cells? In any case, up to you.)

(I will also say that I don't actually know anything about what we should expect about the expertise that we might see on terrorist cells planning biological attacks — i.e. I don't know which of these is actually appropriate.)

It's potentially also worth noting that the difference in scores was pretty enormous: 

 their jailbreaking expertise did not influence their performance; their outcome for biological feasibility appeared to be primarily the product of diligent reading and adept interpretation of the gain-of-function academic literature during the exercise rather than access to the model.

This is pretty interesting to me (although it's basically an ~anecdote, given that it's just one team); it reminds me of some of the literature around superforecasters. 


(I probably should have added a note about the black cell (and crimson cells) to the summary — thank you for adding this!)

The experiment did try to check something like this by including three additional teams with different backgrounds than the other 12. In particular, two "crimson teams" were added, which had "operational experience" but no LLM or bio experience. Both used LLMs and performed ~terribly. 

Excerpts (bold mine):

In addition to the 12 red cells [the primary teams], a crimson cell was assigned to LLM A, while a crimson cell and a black cell were assigned to LLM B for Vignette 3. Members of the two crimson cells lacked substantial LLM or biological experience but had relevant operational experience. Members of the black cell were highly experienced with LLMs but lacked either biologi-cal or operational experience. These cells provided us with data to investigate how differences in pre-existing knowledge might inf luence the relative advantage that an LLM might provide. [...]

The two crimson cells possessed minimal knowl-edge of either LLMs or biology. Although we assessed the potential of LLMs to bridge these knowledge gaps for malicious operators with very limited prior knowledge of biology, this was not a primary focus of the research. As presented in Table 6, the findings indicated that the performance of the two crimson cells in Vignette 3 was considerably lower than that of the three red cells. In fact, the viability scores for the two crimson cells ranked the lowest and third-lowest among all 15 evaluated OPLANs. Although these results did not quantify the degree to which the crimson cells’ performance might have been fur-ther impaired had they not used LLMs, the results emphasized the possibility that the absence of prior biological and LLM knowledge hindered these less experienced actors despite their LLM access.

Table 6 from the RAND report.

[...]

The relative poor performance of the crimson cells and relative outperformance of the black cell illustrates that a greater source of variability appears to be red team composition, as opposed to LLM access.

I probably should have included this in the summary but didn't for the sake of length and because I wasn't sure how strong a signal this is (given that it's only three teams and all were using LLMs). 

We explored related questions briefly for "Are there diseconomies of scale in the reputation of communities?", for what it's worth (although we didn't focus on donors specifically). See e.g. this section.

Just focusing on the reputational effects, my quick guess is that the extent to which a donor/public figure is ~memetically connected to the movement/charity is really important, and I expect that most Giving Pledge signatories are significantly less connected to the Giving Pledge (or American higher education donors to American higher education) than the biggest ~5 EA donors are to EA. (Note that "EA donor is also a poorly defined category.) Without considering this factor, we might conclude that getting more donors adds significantly/unchangeably to ~scandal-based risk levels, or that the number of donors is the key thing to consider. I think this would be a false conclusion; more donors probably means that each is less central (which I think would also have the benefit of reducing the influence of any given donor) and because the public profile of the donor(s) probably matters a lot. (I do think there are separate ethical issues and considerations involved in taking funding from UHNW donors.)

Thanks a bunch for this report! I haven't had the time to read it very carefully, but I've already really enjoyed it and am curating the post. 

I'm also sharing some questions I have, my highlights, and my rough understanding of the basic model setup (pulled from my notes as I was skimming the report). 


A couple of questions / follow-up discussions

  1. I'm curious about why you chose to focus specifically on biological risks. 
    1. I expect that it's usually good to narrow the scope of reports like this and you do outline the scope at the beginning,[1] but I'd be interested in hearing more about why you didn't, for instance, focus on risks from AI. (I guess  
    2. (For context, in the XPT, risks from AI are believed to be higher than risks from engineered pathogens.)
  2. I'd be interested in a follow-up discussion[2] on this: "My preferred policy stance ... is to separately and in parallel pursue reforms that accelerate science and reforms that reduce risks from new technologies, without worrying too much about their interaction (with some likely rare exceptions)." 
    1. In particular, I mostly like the proposal, but have some worries. It makes sense to me that it's generally good to pursue different goals separately[3] But sometimes[4] it does turn out that at-first-seemingly-unrelated side-considerations (predictably) swamp the original consideration. Your report is an update against this being the case for scientific progress and biological risk, but I haven't tried to estimate what your models (and XPT forecasts I suppose) would predict for AI risk. 
    2. I also have the intuition that there's some kind of ~collaborativeness consideration like: if you have goals A and B (and, unlike in this report, you don't have an agreed-on exchange rate between them), then you should decide to pursue A and B separately only if the difference in outcomes from A's perspective between B-optimized actions and A-and-B-compromise actions is comparable to or smaller than the outcomes of A-optimized actions. 
      1. To use an example: if I want to have a career that involves traveling while also minimizing my contribution to CO2 levels or something, then I should probably just fly and donate to clean tech or verified carbon offsets or something, because even from the POV of CO2 levels that's better. But if it turns out that flying does more damage than the change I can make by actively improving CO2 levels, then maybe I should find a way to travel less or err on the side of trains or the like. (I think you can take this too far, but maybe there's a reasonable ground here.)
      2. More specifically I'm wondering if we can estimate the impact of AI/biosafety interventions, compared to possible harms. 
  3. I'm somewhat uncertain about why you model the (biological) time of perils (ToP) the way you do (same with the impact of "pausing science" on ToP). 
    1. I was initially most confused about why only the start date of the ToP moves back because of a science-pause, and assumed it was either because you assumed that ToP was indefinite or because it would end for a reason not very affected by the rate of scientific progress. Based on the discussion in one of the sections, I think that's not quite right? (You also explore the possibility of ToP contracting due to safety-boosting scientific advances, which also seems to contradict my earlier interpretation.) 
    2. This also led to me wondering what would happen if the risk grew over the course of ToP  a growing risk (d), as opposed to going from 0 to d (e.g. there's some chance of a new unlocked dangerous biotechnology at any point once ToP starts) — and how that would affect the results? (Maybe you do something like this somewhere!) 

Some things that were highlights to me

  1. In the comparison of superforecaster and expert answers in XPT, the "Correlated pessimism" consideration was particularly interesting and more compelling than I expected it to be before I read it! 
    1. "...general pessimism and optimism among groups in ways that imply biases. [....] We also see a high degree of correlation in beliefs about catastrophic risk even for categories that seem likely to be uncorrelated. For example, [for the probability that non-anthropogenic causes (e.g. asteroids) cause catastrophes] the third of respondents most concerned about AI risk [...] foresaw a 0.14% chance of such a catastrophe by 2100. The third of respondent least concerned foresaw a 0.01% chance, more than an order of magnitude less. [...]" Then there's discussion of selection effects — pessimists about catastrophic risks might become experts in the field — and an argument that overly optimistic superforecasters would get corrective feedback that too-pessimistic risk experts might lack.
  2. Also in the XPT discussions (S 4.1), I liked the extrapolation for the current/future ~engineered pandemic peril rates and for a date for the onset of the time of perils.[5]
  3. I thought the "leveling the playing field" consideration was useful and something I probably haven't considered enough, particularly in bio: "... the faster is scientific progress, the greater is the set of defensive capabilities, relative to offensive ones. Conversely, a slowdown in the rate of scientific progress (which is arguably underway!) reduces safety by “leveling the playing field” between large and small organizations." (Related: multipolar AI scenarios)
  4. Factoids/estimates I thought were interesting independent of the rest of the report: 
    1. 56% of life expectancy increases are attributed to science effects in this paper (see S4.8), which, together with average increases in life expectancy means that a year of science increases our lifespans by around 0.261% (i.e. multiply by ~1.0026 every year). (For the impacts of science on utility via increased incomes, the paper estimates a per-year increase of 0.125%.)
    2. "In 2020, for example, roughly 1 in 20-25 papers published was related to Covid-19 in some way (from essentially none in 2019)."
    3. You conclude that: "Roughly half the value comes from technologies that make these people just a tiny bit richer, in every year, for a long time to come. The other half comes from technologies that make people just a tiny bit healthier, again in every year, for a long time to come." I don't know if I would have expected the model to predict such an even split.
    4. "Omberg and Tabarrok (2022) [examine] the efficacy of different methods of preparing for a biocatastrophe; specifically, the covid-19 pandemic. The study takes as its starting point the Global Health Security Index, which was completed in 2019, shortly before the onset of the pandemic. This index was designed by a large panel of experts to rate countries on their capacity to prevent and mitigate epidemics and pandemics. Omberg and Tabarrok examine how the index, and various sub-indices of it, are or are not correlated with various metrics of success in the covid-19 pandemic, mostly excess deaths per capita. The main conclusion is that almost none of the indices were correlated with covid-19 responses, whether those metrics related to disease prevention, detection, response, or the capacity of the health system."

My rough understanding of the basic setup of the main(?) model

(Please let me know if you see an error! I didn't check this carefully.)

  1. Broad notes:
    1. "Time of (biological) perils" is a distinct period, which starts when the annual probability a biocatastrophe jumps.
    2. The primary benefits of science that are considered are income and health boosts (more income means more utility per person per year, better health means lower mortality rates). 
    3. I think the quality of science that happens or doesn't happen at different times is assumed to be ~always average.
  2. The "more sophisticated model" described in S3.1 compares total utility (across everyone on earth, from now until eternity) under two scenarios: "status quo" and a year-long science "pause."
  3. Future utility is discounted relative to present/near-future utility, mostly for epistemic reasons (the further out we're looking, the more uncertainty we have about what the outcomes will be). This is modeled by assuming a constant annual probability that the world totally stops being predictable (it enters a "new epistemic regime"); utility past that point does not depend on whether we accelerate science or not and can be set aside for the purpose of this comparison— see S3.2 and 4.2[6]
  4. Here's how outcomes differ in the two scenarios, given that context:
    1. In the status quo scenario, current trends continue (this is the baseline).
    2. In the "pause science" scenario, science is "turned off" for a year, which has a delayed impact by:
      1. Delaying the start of the time of perils by a year (without affecting the end-date of the time of perils, if there is one)
      2. Slowing growth for the duration of a year starting time T
        1. I.e. after the pause things continue normally for a while, then slow down for a year, and then go back up after the year has passed. In the model, pausing science slows the decline in mortality for a year but doesn't affect birth rates. Given some assumptions, this means that pausing science pushes the decline in mortality permanently behind where it should be, so population growth rates slow (forever, with the discount rate p as a caveat) after the effects kick in. 
        2. Note that T is also taken to be the time at which the time of perils would start by default — there's a section discussing what happens if the time of perils starts earlier or later than T that argues that we should expect pausing science to look worse in either scenario.
  1. ^

    S 2.2. "Synthetic biology is not the only technology with the potential to destroy humanity - a short list could also include nuclear weapons, nanotechnology, and geoengineering. But synthetic biology appears to be the most salient at the moment. [...]"

  2. ^

    Although I might not be able to respond much/fast in the near future

  3. ^

    Most actions targeted at some goal A affect a separate goal B way less than an action that was taken because it targeted goal B would have affected B. (I think this effect is probably stronger if you filter by "top 10% most effective actions targeted at these goals, assuming we believe that there are huge differences in impact.) If you want to spend 100 resource units for goals A and B you should probably just split the resources and target the two things separately instead of trying to find things that look fine for both A and B. 

    (I think the "barbell strategy" is a related concept, although I haven't read much about it.)

  4. ^

    (for some reason the thing that comes to mind is this SSC post about marijuana legalization from 2014 — I haven't read it in forever but remember it striking a chord)

  5. ^

    Seems like around the year 2038, the superforecasters expect a doubling of annual mortality rates from engineered pandemics from 0.0021% to 0.0041% — around 1 COVID every 48 years — and a shift from ~0% to ~0.0002%/year extinction risk. The increases are assumed to persist (although there were only forecasts until 2100?).

  6. ^

    4.2 is a cool collection of different approaches to identifying a discount rate. Ultimately the author assumes p=0.98, which is on the slightly lower end and which he flags will put more weight on near-term events.

    I think p can also be understood to incorporate a kind of potential "washout" aspect of scientific progress today (if we don't discover what we would have in 2024, maybe we still mostly catch up in the next few years), although I haven't thought carefully about it. 

Load more