Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Who Was the Funder that Counterfactually Resulted in LEEP Starting?, published by Joey on July 4, 2023 on The Effective Altruism Forum.

Lead Exposure Elimination Project (LEEP) is an outstanding Charity Entrepreneurship-incubated charity recognized externally for its impactful work by RP, Founders Pledge, Schmidt Futures, and Open Philanthropy. It's one of the clearest cases of new charities having a profound impact on the world. However, everything is clear in hindsight; it now seems obvious that this was a great idea and team to fund, but who funded LEEP at the earliest stage? Before any of the aforementioned bodies would have considered or looked at them, who provided funding when $60k made the difference between launching and not existing?

The CE Seed Network, so far, has been a rather well-kept secret. They are the first people to see each new batch of CE-incubated charities and make a decision on whether and how much to support them. A handful of donors supported LEEP in its earliest days, culminating in the excellent charity we see today. Some of them donated anonymously, never seeking credit or the limelight, just quietly making a significant impact. Others engaged deeply and regularly with the team, eventually becoming trusted board members. Historically, the Seed Network has been a small group (~30) of primarily E2G-focused EAs, invited by the CE team or alumni from the CE program to join. However, now we are opening it up for expressions of interest for those who might want to join in future rounds. Our charity production has doubled (from 5 to 10 charities a year) and although our Seed Network has grown, there is still room for more members to join to support our next batches of charities.

We have now created a website to describe how it works. On that website, there's an application form for those who might be a good fit to be a member in the future. It’s not a great fit for everyone as it focuses on the CE (near-termist) cause areas and donors who could donate over $10k a year to new charities and can make a decision on whether and whom to fund with how much in a short period of time when we send out the newest project proposals (~9 days). But for those who fit, we think it's one of the most impactful ways to donate.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Ways I Expect AI Regulation To Increase Extinction Risk, published by 1a3orn on July 4, 2023 on LessWrong.

The following are some very short stories about some of the ways that I expect AI regulation to make x-risk worse.

I don't think that each of these stories will happen. Some of them are mutually exclusive, by design. But, conditional upon AI heavy regulations passing, I'd strongly expect the dynamics pointed at in more than one of them to happen.

Pointing out potential costs of different regulations is a necessary part of actual deliberation about regulation; like all policy debates, AI policy debates should not appear one-sided. If someone proposes an AI policy, and doesn't actively try to consider potential downsides -- or treats people who bring up these downsides as an opponent -- they are probably doing something closer to propaganda than actual policy work.

All these stories have excessive detail; these are archetypes meant to illustrate patterns more than they are predictions. Before mentioning one fix that would stop a problem, consider whether it would be subject to similar problems in the same class.

(a) Misdirected Regulations Reduce Effective Safety Effort; Regulations Will Almost Certainly Be Misdirected

"Alice, have you gotten the new AI model past the AMSA [AI and ML Safety Administration] yet?" said Alex, Alice's manager. "We need their approval of the pre-deployment risk-assessment to launch."

"I'm still in correspondence with them," said Alice, wincing. "The model is a little more faceblind than v4 -- particularly in the case of people with partially shaved heads, facial tattoos, or heavy piercings -- so that runs into the equity clauses and so on. They won't let us deploy without fixing this."

"Sheesh," said Alex. "Do we know why the model is like this?"

Alice shrugged.

"Well, getting this model past the AMSA is your number one priority," Alex said. "Check if you can do some model surgery to patch in the facial recognition from v4. It should still be compatible."

"Alex," said Alice. "I'm not really worried about the face-blindness. It's a regression, but it should have minimal impact."

"I mean, I agree," grinned Alex, "but don't put that in an email."

"But," Alice continued, hesitantly, "I'm really concerned about some of the capabilities of the model, which I don't think we've explored completely. I know that it passed the standard safety checks, and I know I've explored some of its capabilities, but I think this is much much more important than -- "

"Alice," interrupted Alex. "We've gone over this before. You're on the Alignment and Safety Team. That means part of your job is to make sure the model passes AMSA standards. That's what the Alignment and Safety team does."

"But the AMSA regulations went through Congress before Ilya's Abductor was even invented!" Alice said. "I know the regs are out of date; you know the regs out of date; everyone knows they're out of date, but they haven't even begun to change yet!--"

"Alice," interrupted Alex again. "Sorry for interrupting. And look, I'm really sorry about the situation in general. But you've spent two whole months on this already. I let you bring in Aaron on the problem, and he spent a whole month. Neither of you have found anything dangerous."

He continued: "I've shielded you from pressure from my boss for a while now. He was understanding the first three weeks, impatient the next three weeks, and has been increasingly irritated since. I wish we had more time for you, but we can't just ignore the AMSA -- it's the law. I need you to work on this and only this, or you're fired."

I expect actual instances of misdirected safety effort from safety laws to be far more universal, and only moderately more hidden, than is indicated in this dialogue.

If you think this is unlikely, consider IRBs, and co...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Two Percolation Puzzles, published by Adam Scherlis on July 4, 2023 on LessWrong.

Take a very large chessboard (NxN, where N is huge).

Remove some fraction 1-p of the squares at random, leaving a fraction p of them.

Can you place a queen on the first row and then, via some sequence of legal moves, get it to the last row?

This is probabilistic, of course, but it turns out the probability of success undergoes a sharp phase transition -- near-zero for small values of p, then suddenly rising almost to one in the vicinity of a critical value p_queen.

For different pieces, the critical value is different, e.g. p_rook for a rook. (Note that p_queen <= p_rook.)

Question 1: What is p_queen + p_rook?

Bonus Question: Why didn't I ask you for the values p_queen and p_rook separately?

Now let's invent a new chess piece, the bondsman. This piece can move like a rook, and is also allowed to move along northeast-southwest diagonals between white squares, and along northwest-southeast diagonals between black squares. (There's also the antibondsman who has the same abilities, but with the words "white" and "black" swapped.)

Question 2: What is p_bondsman?

Bonus Question: Why did I call the piece that?

Answers to follow in a later post.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Three mistakes in the moral mathematics of existential risk (David Thorstad), published by Global Priorities Institute on July 4, 2023 on The Effective Altruism Forum.

Abstract

Longtermists have recently argued that it is overwhelmingly important to do what we can to mitigate existential risks to humanity. I consider three mistakes that are often made in calculating the value of existential risk mitigation: focusing on cumulative risk rather than period risk; ignoring background risk; and neglecting population dynamics. I show how correcting these mistakes pushes the value of existential risk mitigation substantially below leading estimates, potentially low enough to threaten the normative case for existential risk mitigation.

I use this discussion to draw four positive lessons for the study of existential risk: the importance of treating existential risk as an intergenerational coordination problem; a surprising dialectical flip in the relevance of background risk levels to the case for existential risk mitigation; renewed importance of population dynamics, including the dynamics of digital minds; and a novel form of the cluelessness challenge to longtermism.

Introduction

Suppose you are an altruist. You want to do as much good as possible with the resources available to you. What might you do? One option is to address pressing short-term challenges. For example, GiveWell (2021) estimates that $5,000 spent on bed nets could save a life from malaria today.

Recently, a number of longtermists (Greaves and MacAskill 2021; MacAskill 2022b) have argued that you could do much more good by acting to mitigate existential risks: risks of existential catastrophes involving “the premature extinction of Earth-originating intelligent life or the permanent and drastic destruction of its potential for desirable future development” (Bostrom 2013, p. 15). For example, you might work to regulate chemical and biological weapons, or to reduce the threat of nuclear conflict (Bostrom and Cirkovi ' c' 2011; MacAskill 2022b; Ord 2020).

Many authors argue that efforts to mitigate existential risk have enormous value. For example, Nick Bostrom (2013) argues that even on the most conservative assumptions, reducing existential risk by just one-millionth of one percentage point would be as valuable as saving a hundred million lives today. Similarly, Hilary Greaves and Will MacAskill (2021) estimate that early efforts to detect potentially lethal asteroid impacts in the 1980s and 1990s had an expected cost of just fourteen cents per life saved. If this is right, then perhaps an altruist should focus on existential risk mitigation over short term improvements.

There are many ways to push back here. Perhaps we might defend population-ethical assumptions such as neutrality (Naverson 1973; Frick 2017) that cut against the importance of creating happy people. Alternatively, perhaps we might introduce decision-theoretic assumptions such as risk aversion (Pettigrew 2022), ambiguity aversion (Buchak forthcoming) or anti-fanaticism (Monton 2019; Smith 2014) that tell against risky, ambiguous and low-probability gambles to prevent existential catastrophe. We might challenge assumptions about aggregation (Curran 2022; Heikkinen 2022), personal prerogatives (Unruh forthcoming), and rights used to build a deontic case for existential risk mitigation. We might discount the well-being of future people (Lloyd 2021; Mogensen 2022), or hold that pressing current duties, such as reparative duties (Cordelli 2016), take precedence over duties to promote far-future welfare.

These strategies set themselves a difficult task if they accept the longtermist’s framing on which existential risk mitigation is not simply better, but orders of magnitude better than competing short-termist interventions. Is it really so obvious ...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: The literature on aluminum adjuvants is very suspicious. Small IQ tax is plausible - can any experts help me estimate it?, published by mikes on July 4, 2023 on LessWrong.

I am a PhD biostatistician who has worked on a project in this area. I am hoping to crowdsource opinions on this issue, especially from readers with knowledge of neurology, nephrology and/or toxicology.The recent news cycle motivated me to spend some time reading up on the safety literature for aluminum vaccines. Aluminum is not in the Covid vaccine, but it is in most childhood vaccines.I am deeply unsettled by what I found. Here is a summary. I would appreciate being corrected in the comments:- The science on this topic is surprisingly neglected. Aluminum is a known neurotoxin in small concentrations in the brain. So, the obvious follow-up question: "Is the amount accumulated from childhood vaccines so small as to be irrelevant?" Unfortunately, aluminum has not undergone rigorous safety testing for cumulative cognitive effects.

The fact that I could not find a convincing answer is a multi-decade policy failure.-The evidence for safety cited by regulators and mainstream review papers appears to have gaping issues (see end of post for notes on the studies Mitkus et al (2011) and Karwowski et al (2018). They do not justify anywhere near the faith being placed in them.)-Mainstream educational sources are verifiably gaslighting readers by suggesting that the net exposure in vaccines is less than dietary exposure; but this fails sanity checks with basic multiplication, given the oral bioavailability of aluminum is about 0.1% on average in diet and <0.01% in antacids according to the ATSDR. No wonder people are going nuts about the public health experts.-There is a whole literature of claims on blood brain barrier mechanics, chemical state of the aluminum after injection, and transport via macrophages, which regulators are ignoring, but should probably affect the analysis of Mitkus 2011.

I am entirely unqualified to judge and don't particularly trust this literature, so I'm ignoring this but looking forward to comments on the topic.-Alternatives to aluminum exist, including mRNA and other new adjuvants. In the 2000's, regulators pushed industry to remove mercury from childhood vaccines; they should consider doing this again.-A cognitive tax due to aluminum on the scale of ~0.1 IQ point for the first Hepatitis B shot is plausible (that's just from the first 250 micrograms of Al. There are > 4 mg in the whole schedule). That's my rough fermi estimate, based on an experiment showing developmental delay in premature infants exposed to hundreds of micrograms aluminum in IV fluids (Bishop et al 1997). This estimate is extremely sensitive to assumptions about renal clearance and sensitivity of brain development in premature infants compared to full term. Can anyone suggest a more accurate method?

With this approach: does anyone know enough nephrology to pin down the length of time that intravenous aluminum in Bishop's premies would have spent circulating in blood, given their reduced renal function? Anyone know enough neuro to comment on whether impacts on cognitive development should be linear in the amount of neurotoxin? How much larger should it be in premies vs term infants?______Notes on failures in the literature:Claims of safety of aluminum seem to bottom out in mainly two papers: Karwowski et al (2018) and Mitkus et al (2011). The FDA cites Mitkus 2011 to determine a tolerable risk level of aluminum exposure and argues that vaccine exposure is below it; the UK health authority and various review papers (such as this one) cite Karwowski as evidence of low accumulation of aluminum in infants.

Study 1

Mitkus et al (2011) which the FDA commissioned on this topic, argues that exposure to aluminum due to vacc...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: GWWC Newsletter: June 2023, published by Giving What We Can on July 4, 2023 on The Effective Altruism Forum.

Hello and welcome to our June newsletter!Pop quiz! If you travelled to visit every person who's taken the GWWC pledge, how many different countries would you visit?Answer: 100 countries! With the newest addition of Uzbekistan , Giving What We Can members are present in 100 countries worldwide! Turns out the idea of giving to help others effectively has universal appeal!

Even though our movement is growing, we need your help - talking to your friends, family and colleagues is one of the best ways to help us change the norms around giving, which in turn means faster progress on some of the world’s biggest issues.We’d love to know how we can help you talk to the people in your life about high-impact charities and how we can help you advocate. (We already have lots of ideas and resources here)If you have any ideas about what you would find helpful, simply reply or send me a quick email at: [email protected] you’ll find loads of interesting updates from our partner charities and other news we think you’ll like!

With gratitude, - Grace Adams & the Giving What We Can team

News & Updates

Community

Our Executive Director Luke Freeman recently published a post on the EA Forum about the role of individuals in helping to fund high-impact projects and charities as well as hosting an AMA (Ask Me Anything) about his work, life and more!

Director of Research, Sjir Hoeijmakers published a post on the EA Forum with “Four claims about the role of effective giving in the EA Community”.

Power for Democracies is a new non-profit democracy charity evaluator based in Berlin, Germany, and operating globally. They are looking to hire 5-6 democracy enthusiasts to form their ‘Knowledge & Research Team’. The objective of the team is twofold: To build and execute a ‘knowledge-building roadmap’ that will lead to a growing set of methodologies for identifying highly effective pro-democracy interventions and potential NGOs to apply them. And to use these methodologies to generate giving recommendations for the international community of democracy-focused, effectiveness-driven donors.

Magnify Mentoring is still accepting mentee applications from women, non-binary, and trans people of any gender who are enthusiastic about pursuing high-impact career paths for the next day or two. On average, mentees and mentors meet once a month for 60-90 minutes. Magnify Mentoring offers mentees access to a broader community with a wealth of professional and personal expertise. You can find out more here and apply here.

Evaluators, grantmakers and incubators

Updates to ACE’s Charity Evaluation Criteria in 2023: Animal Charity Evaluators (ACE) is entering its 2023 charity evaluation season! This is the time of year when ACE works to identify charities that can do the most good for animals with two years of additional funding. To provide more transparency and insight into its evaluation process, ACE is sharing some changes it made to its four charity evaluation criteria this year.

ACE's Updated Strategic Plan: One year ago, in 2022, ACE developed a strategic plan for the period of 2022–2024. This plan, created collectively by ACE staff under the leadership of the Acting Executive Director and approved by the board of directors, was the result of the hard work and dedication of a severely understaffed team. It represented what was needed then. Things have changed since last year. ACE added several talented individuals to their team, including new leadership and board members. ACE now has an updated strategic plan and is looking forward to testing its assumptions and delivering results.

GiveWell CEO and co-founder Elie Hassenfeld was interviewed on the 80,000 Hours podcast about newer areas of ...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Ten Levels of AI Alignment Difficulty, published by Sammy Martin on July 3, 2023 on LessWrong.

Chris Olah recently released a tweet thread describing how the Anthropic team thinks about AI alignment difficulty. On this view, there is a range of possible scenarios ranging from ‘alignment is very easy’ to ‘alignment is impossible’, and we can frame AI alignment research as a process of increasing the probability of beneficial outcomes by progressively addressing these scenarios. I think this framing is really useful, and here I have expanded on it by providing a more detailed scale of AI alignment difficulty and explaining some considerations that arise from it.

The discourse around AI safety is dominated by detailed conceptions of potential AI systems and their failure modes, along with ways to ensure their safety. This article by the DeepMind safety team provides an overview of some of these failure modes. I believe that we can understand these various threat models through the lens of "alignment difficulty" - with varying sources of AI misalignment sorted from easy to address to very hard to address, and attempt to match up technical AI safety interventions with specific alignment failure mode scenarios. Making this uncertainty clearer makes some debates between alignment researchers easier to understand.

An easier scenario could involve AI models generalising and learning goals in ways that fit with common sense. For example, it could be the case that LLMs of any level of complexity are best understood as generative frameworks over potential writers, with Reinforcement Learning from Human Feedback (RLHF) or Constitutional AI (CAI) selecting only among potential writers. This is sometimes called ‘alignment by default’.

A hard scenario could look like that outlined in ‘Deep Deceptiveness’, where systems rapidly and unpredictably generalise in ways that quickly obsolete previous alignment techniques, and they also learn deceptive reward-hacking strategies that look superficially identical to good behaviour according to external evaluations, red-teaming, adversarial testing or interpretability examinations.

When addressing the spectrum of alignment difficulty, we should examine each segment separately.

If we assume that transformative AI will be produced, then the misuse risk associated with aligned transformative AI does not depend on how difficult alignment is. Therefore, misuse risk is a relatively bigger problem the easier AI alignment is.

Easy scenarios should therefore mean more resources should be allocated to issues like structural risk, economic implications, misuse, and geopolitical problems. On the ‘harder’ end of easy, where RLHF-trained systems typically end up honestly and accurately pursuing oversimplified proxies for what we want, like ‘improve reported life satisfaction’, or ‘raise the stock price of company X’, we also have to worry about scenarios like Production Web or What Failure looks like 1 which require a mix of technical and governance interventions to address.

Intermediate scenarios are cases where behavioural safety isn’t good enough and the easiest ways to produce Transformative AI result in dangerous deceptive misalignment. This is when systems work against our interests but pretend to be useful and safe. This scenario requires us to push harder on alignment work and explore promising strategies like scalable oversight, AI assistance on alignment research and interpretability-based oversight processes. We should also focus on governance interventions to ensure the leading projects have the time they need to actually implement these solutions and then use them (in conjunction with governments and civil society) to change the overall strategic landscape and eliminate the risk of misaligned AI.

In contrast, if alignment is as hard as pessimis...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Monthly Roundup #8: July 2023, published by Zvi on July 3, 2023 on LessWrong.

As always, plenty of non-AI things are happening, most of which do not justify their own posts. This is where all of those go. Skimming for things relevant to your interests is encouraged.

Before I begin, a Point of Order: I plan to have a post later regarding the rate limitation and disruption on Twitter, which disrupts many features of Tweetdeck and means that if you aren’t logged in Twitter links no longer work. Adjustments will need to be made if the situation is not fixed very soon, despite my having no fear of ever hitting the 10k rate limit itself. It is not practical to go back and fix links or other decisions on already written material but I am adjusting the process going forward.

Bad News

RIP 538. Nate Silver leaving was always going to be a hard blow to recover from, as the models are the heart and soul of the entire model. It also depends on a deep commitment, visible to all, to fair play and avoiding putting a finger on the scale or using the scale as leverage. Alas:

Nate Silver: Very bad to do a Spanish Inquisition with pollsters based on their political orientation. I love my ex-colleagues (this is coming from a new guy they hired) but if this is their practice, hope ABC will stop use of 538 brand so it isn’t associated with me.

Washington Examiner: “Rasmussen must explain the nature of its relationship with several right- leaning blogs and online media outlets, which have given us reason to doubt the ethical operation of the polling firm,” demanded Elliott Morris, the editorial director of data analytics at liberal ABC News and FiveThirtyEight.

No. I have listened to many a Model Talk, with in-depth discussion of how to (among many other considerations) construct a polling aggregation systems that best resists partisan manipulation, despite the need to ‘play defense’ where the pollsters respond to your policies. Using logic like this is incompatible with such an effort.

Cocaine Bear, meet Government Warrantless Survey of Private Property Bear.

It is an important fact about banks that both they will pitch beyond atrocious products at you constantly, and also that no overloaded number of indicators, however explicit, that you know better, will get them to stop. Until McKenzie, I do begrudge the bank and the salesperson here, actively pitching a 2% CD right now is stealing. Worse, it’s wasting your time. The amount of idiocy I had to push through to get a Chase Bank representative to tell me their real interest rate offer was. not small.

Claim that a third of software job applicants, and over 5% of actual job holders, can’t write fizzbuzz, which implies they can’t do anything else either, and then the claim that many people literally never produce any productive output yet stay employed indefinitely. Which is good news if you are competing against such businesses. It is ambiguous news if you are trying to get hired. It means you have a big edge on those people in this area, also means that area is less important than you thought.

We have a list of jobs for which workers are in short supply, called Schedule A, to streamline immigration. It hasn’t been updated since 1991. Might want to do that.

Good News, Everyone

Worth repeating: Across the Spiderverse is the best new movie I’ve seen in years. If you have not yet done so, see it now, ask no questions. Unless you never saw Into the Spiderverse, in which case see that first and same advice applies.

Excellent new long Paul Graham essay on How to Do Great Work. In so many ways no one else could have written it this way. I do not agree with everything and do not forget to consider reversing all advice you hear, and yes this is too long, yet this is an excellent essay and you should read it. I hope he invests the time to make it shor...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Benefits of being rejected from CEA’s Online team, published by Ben West on July 3, 2023 on The Effective Altruism Forum.

Summary:

If you apply and are rejected, then I (and presumably other hiring managers) will often either not remember you — or actually be enthusiastic about you (e.g if you made it to later hiring rounds) if anyone asks in the future.

People who have been rejected from the CEA Online team have gotten concrete job opportunities that they would not have gotten if they had never applied

If you're rejected, I can also give you advice and connect you to other opportunities.

Rejection is hard, but fear of it shouldn't stop you from applying.

Note: I wrote this year ago and temporal references like “the most recent” should be interpreted relative to that. Importantly: I no longer am on the Online team, although I think this post is still roughly accurate for them.

I sometimes talk to people who are nervous about applying to EA organizations because they think a rejection could damage their chances at not just the organization they apply to but all EA organizations.

This fear is not completely ungrounded – EA is a small community, and hiring managers do occasionally talk to each other about candidates. As with most worries with how others think of you though, "You probably wouldn’t worry about what people think of you if you could know how seldom they do": 100+ people apply to CEA per month, and my memory is pretty bad. The people grading applications will probably not remember people whose applications they reject, especially if that happens early on, and if it happens later, that likely means that they saw something promising. (There are also other costs to applying, like your time and energy.)

I wanted to point out though that being rejected as a candidate, particularly if you make it to the final round of a hiring process, can actually be substantially positive.

Here are some things that happened to rejected candidates in some of the hiring rounds I ran:

Hired by CEA as a contractor for a position similar to what they applied to (the position we were hiring for ended up needing slightly more than 1 FTE of work, so we hired them to do some of the overfill)

Received a grant to quit their job and work independently on something similar to what they applied to after I encouraged them to do so and recommended the grant

Repeatedly consulted CEA on their area of expertise as a volunteer (though we offered to pay them, they just declined payment)

Received a grant from LTFF to skill up after I endorsed them, based on what I learned during their hiring process

I also try to give useful feedback to candidates who are rejected. Here is an email I recently received from a rejected applicant, which Carly (the applicant) kindly agreed to share publicly:

Hi Ben,

Hope you are well! I was in the Project Coordinator search at CEA a few months ago and wanted to drop you a quick note of gratitude.

I want to thank you for your part in helping to transition my career into EA. I was very hopeful about getting the role at CEA and can easily imagine a scenario where a typical rejection letter - short, generic, or dismissive - may have dampened my enthusiasm and at worst lowered my spirits enough to not go to EAGx Boston the following day. Your rejection letter, however, was so insightful and encouraging that it had the opposite effect. You motivated me to keep learning and networking and to go to the conference which started the chain of events that led me to a position that I'm very excited to start at Alvea in a few weeks.

All of this is just to say thanks and no need for a reply! I know those letters take time and are not expected or necessary for unsuccessful candidates but they make an impact!

Carly Tryens

People in the EA network tend to be both inexperienced and self-mo...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Evaluating claims about ContraPest, published by Spencer Ericson on July 3, 2023 on The Effective Altruism Forum.

TL;DR: We conclude that there is some evidence against ContraPest’s claims that the product is “humane,” “it is not an endocrine disruptor,” and “its effects are reversible.” The literature suggests that the active ingredients in ContraPest may cause bone loss, cancer, organ injury, immunosuppression, and endocrine disruption in rats. Any of these could be a risk factor for decreased well-being in rats. VCD may result in permanent sterility in rats over time or at high doses, and at low doses, it results in irreversible premature reproductive senescence.

Epistemic status: Uncertain. This was about a 40-hour investigation by me, Spencer, a person with ~no biology background. I conclude that more research is needed into every question I pose. This post will be edited if new information comes in.

Quirks: I say “we” because writing “I” in formal writing makes me feel weird. Where I say “We are not aware of studies.” I mean that I searched 3-4 relevant key terms on Google Scholar about that question and came up empty.

Acknowledgements: I am grateful to Constance Li for sponsoring this project. I am grateful to Contance Li, Michael St. Jules, and Holly Elmore for feedback. I am grateful to two interviewees for their help. (None of the acknowledged people have endorsed or rigorously fact-checked my claims herein.)

Introduction

ContraPest is the first and only current EPA-approved contraceptive for Norway rats and roof rats. It was approved in 2016 and brought to market in 2021. ContraPest has received increasing attention; for example, it was featured in Time in March 2023. It is touted as a humane option for rat population control that could replace many use cases for rodenticides. Rodenticides tend to cause prolonged, painful deaths, both for rats and for other wild animals, so ContraPest is thought to be an opportunity to improve wild animal welfare. Its manufacturer, SenesTech, claims that “ContraPest is a non-lethal rodenticide/pesticide that does not bioaccumulate, it is not an endocrine disruptor, and its effects are reversible.”

ContraPest contains two chemicals in small amounts, 4-vinylcyclohexene diepoxide (VCD) and triptolide. VCD reduces fertility in female rats by depleting ovarian follicles, which mature into eggs. Triptolide reduces fertility in male and female rats by interrupting the development of ovarian follicles and sperm. However, as we will see below, these chemicals might have side effects like permanent reductions in fertility, cancer, bone loss, organ injury, immunosuppression, and hormone disruption at high doses. More research in needed into the one-time and repeated doses where these side effects become apparent in rats, animals that eat rats, other animals that might come into contact with ContraPest, and humans.

We have yet to see a post that thoroughly and critically evaluates SenesTech’s claims about ContraPest. However, we have come to know that “many cities” and at least one animal sanctuary that ran pilots of ContraPest did not continue to use the product due to their concerns about its effectiveness and side effects. To better understand how this product affects wild animal welfare, we attempt to evaluate SenesTech’s claims below. You can skip to the Conclusion, or read the TL;DRs at the top of each section.

(We wrote this report after reading Holly Elmore et al.’s sequence for Rethink Priorities, “The rodenticide reduction sequence.” You can read our summary of the sequence here.)

Questions

We have the following questions about ContraPest. We will only be able to answer some of them.

Why did the cities and organizations that tried ContraPest stop using it?

At what dose is ContraPest toxic?

Is it true that ContraPest causes n...