Our Mistakes

As CEA has grown and changed as an organization, we recognize that our work has sometimes fallen short of the standards we have for ourselves. We believe we have made significant strides toward our current goals of professionalism, collaboration, and excellent implementation of our projects. But we also want to give context on some of CEA’s mistakes, both resolved and unresolved. Here are some reasons why:

  • To acknowledge ways our mistakes have affected others
  • To share information about problems you may have observed with our work
  • To help you assess whether you think our corrective measures are adequate

When you evaluate us as an organization, we recommend using this page, but also looking directly at what we've produced, rather than just taking our word for things.

This is not an exhaustive list of every problem with CEA’s work. In particular:

  • We may be missing mistakes that were made in CEA’s early years (for which current staff were not present).
  • It mostly covers mistakes we made that affected outside stakeholders (EA community members, group organizers, donors, etc.) rather than our own staff.
  • We don’t list all the ways our projects were inefficient or suboptimal.
  • The sections on Giving What We Can and EA Funds only cover the 2016-2020 period when they were run as part of CEA rather than as a separate organization.

Please contact us if you know of other items that should be listed here, or other ways we could improve.

Last updated: October 2022

Underlying problems

These are problems we believe contributed to many of our mistakes.

Running too many projects (2016 - 2020)

Since 2016, CEA has had a large number of projects for its size. For example, at the beginning of 2019, CEA had 20 staff and many projects that took significant staff time (EA Global conferences, EAGx conferences, the EA Forum, EA Funds, EA Grants, support for EA groups, Community Building Grants, community health support, Giving What We Can, and the EA Leaders Forum). We did not always identify which projects were not performing well and discontinue them. We think we should have taken on fewer new projects, set clearer expectations for them, and ended unsuccessful projects earlier.

Running this wide array of projects has sometimes resulted in a lack of organizational focus, poor execution, and a lack of follow-through. It also meant that we were staking a claim on projects that might otherwise have been taken on by other individuals or groups that could have done a better job than we were doing (for example, by funding good projects that we were slow to fund).

Some of the projects CEA launched during this period were discontinued, but detailed public evaluations of them were not always conducted due to limited staff capacity. While we think that this was the right tradeoff given our staff capacity, it’s not ideal, and it would have been good if we’d had more staff capacity for this.

Current status: While we don’t want to abruptly drop any projects that the community depends on, we evaluate existing projects on an ongoing basis to ensure both that they are delivering impact and that we are their appropriate owner.

In April 2020, EA Grants stopped considering new grants, and referred applicants to EA Funds. In 2020, we hired Luke Freeman to run Giving What We Can and Jonas Vollmer to run EA Funds. They made solid progress on those programs, and by December 2020 they were operating independently of our Executive Director's oversight and reporting directly to CEA's board, although CEA continues to provide operational support to these projects. We think that this allows us to be significantly more focused on our core work.

Executive turnover and limited board oversight (2016 - 2019)

Between 2016 and 2019, CEA had a series of five executive directors or CEOs. Some of this was related to changes in CEA itself (in 2016, moving from a minor umbrella organization encompassing projects like Giving What We Can and 80,000 Hours to a full-fledged organization with its own projects). But we then had several other executive transitions for various reasons. This meant that CEA was often in a period of transition and had frequent changes in vision for what types of projects to take on and what role to play in the effective altruism community.

The incoming leaders and the transitions themselves did not always have as much oversight or guidance from the board as we now think would have been helpful.

Current status: Max Dalton became interim executive director in January 2019. For the first time, CEA held an in-depth executive search with ongoing involvement from the board of trustees, and the trustees confirmed Max as executive director in October 2019. The board is now more actively involved than in the past. A designated trustee is responsible for writing semi-annual reviews of Max's performance (with input from other trustees), and meets with Max regularly to give feedback on CEA's work and strategy.

Specific problems

Representation of EA cause areas

We think CEA should have been clearer about which of our projects are meant to be broadly representative of the interests of the EA community, and which are aiming to be cause-specific or otherwise not representative of the whole community. At times, we’ve carried out projects that we presented as broadly EA that in fact overrepresented some views or cause areas that CEA favored. We should have either worked harder to make these projects genuinely representative, or have communicated that they were not representative.

EA Global content (2015 - 2016)

How we fell short: EA Global is meant to cover a broad range of topics of interest to the effective altruism community, but in 2015 and 2016 we did not provide strong content at EA Global from the area of animal advocacy. This stemmed from having organizers who were not themselves active in animal advocacy and who did not seek out adequate advice about speakers and topics in this area. This made some community members who focus on animal advocacy feel unvalued. We think that EA Global should provide good content on a variety of causes in EA, and we failed to do so in those cases.

Current status: In 2017, we hired a consultant from the animal advocacy field to advise us on better content in that area. Since 2018, we have used a panel of experts in different projects genuinely representative, or have communicated that they were not representative.cause areas to advise us on EA Global content, and we believe this has resulted in better representation of the different cause areas of interest to the EA community.

Effective Altruism Handbook (2018)

How we fell short: In 2018, we published the second edition of the Effective Altruism Handbook, which emphasized our longtermist view of cause prioritization, contained little information about why many EAs prioritize global health and animal advocacy, and focused on risks from AI to a much greater extent than any other cause. This caused some community members to feel that CEA was dismissive of the causes they valued.

We think that this project ostensibly represented EA thinking as a whole, but actually represented the views of some of CEA’s staff, which was a mistake. We think we should either have changed the content, or have presented the content in a way that made clear what it was meant to represent.

Current status: We consulted more broadly in the community, particularly amongst those who were critical of the 2nd edition. We also edited the second edition of the Handbook to make clearer that it was a project of CEA rather than community-sourced. Finally, we took this feedback into account when we developed the latest version of the handbook.

EffectiveAltruism.org (2017 - 2021)

How we fell short: From late 2017 until early 2021, the Reading List on the EffectiveAltruism.org homepage included content on a variety of longtermist cause areas but not global poverty (which was the community’s most popular cause at the time per the EA Survey). Similarly, the Resources page was disproportionately skewed toward longtermist content.

Current status: Since March 2021, EffectiveAltruism.org has included a wider variety of content. In November 2021 we hired a product manager with responsibility for EffectiveAltruism.org. Between February and July 2022 we completed a major overhaul of its design and content.

EA Leaders Forum (2016 - 2019)

How we fell short: Beginning in 2016, we held the EA Leaders Forum, initially intended as a space for core community members to coordinate and discuss strategy. The format and the number of attendees have changed over time, and as time passed we invited attendees disproportionately from organizations focused on global catastrophic risks or with a longtermist worldview. While the name is less significant than the structure of the event itself, we should not have continued calling it the “EA Leaders Forum” after it no longer involved a representative group of EA community leaders.

Current status: From 2020-2021, we used the name "Coordination Forum", and try to make clear that it is focused on building relationships and stimulating discussion, rather than making decisions for the community.

In-Depth EA Program (2021 - present)

How we fell short: The In-Depth EA Virtual Program is intended to be a deep dive into EA ideas and key uncertainties. From early 2021 until 2022, the program failed to represent the breadth of EA ideas. Topics included AI safety and biorisk, but gave less attention to global health and development and animal welfare.

Current status: In August 2022 we began to work on an update to the program to better represent a variety of cause areas. As of October 2022, new cause areas have been added, and will continue to be iterated and improved.

EA Global

Communication about events (2016 - 17)

How we fell short: As EA Global admissions criteria have changed over time, we have not always communicated these changes adequately to the community, leading to confusion and disappointment.

In 2016, we committed to a large event size and then had problems with our email system, leading to low ticket sales. We then heavily promoted the conference to the community in an attempt to fill the event. We used messages that we now think were dishonest (such as giving fake deadlines, and sending emails “from” other people when they had not vetted the content of the emails).

We now think we should not have publicly committed to a particular number of attendees in the first place, and should not have tried so hard to sell a given number of tickets, because some community members found our efforts to be overly pushy. This also created confusion in later years when admission standards were much more selective.

The following year, midway through admissions for EA Global London, we decided to make the 2017 London event more selective than initially planned. This meant that some applicants who had seen initial communications about the event were surprised not to be admitted, as they met the initial criteria but not the more selective criteria. In many cases, they also remembered the strong push for ticket sales in 2016 and were confused about the change.

In the years since, there has continued to be disagreement and confusion about the admissions process, some of it based on other mistakes we’ve made.

Current status: The staff involved in the misleading marketing effort no longer work at CEA.

We’ve tried to better explain the application process and criteria in our communications about the conference, on the application itself, in our emails responding to applications, in our FAQ, and in this Forum post.

Communication with EAGx organizers (2016 - 2019)

How we fell short: At times, our communication with EAGx organizers has been slow or absent, sometimes impeding their work. For example, in 2016 EAGxBerkeley organizers described unresponsiveness from our staff as a low point in their experience as event organizers.

Current status: We now have appropriate staff capacity for the scale of these events, enabling our staff to spend more time responding to organizers.

Admissions inbox monitoring (2021)

How we fell short: In the lead-up to EA Global: Reconnect (March 21-22, 2021), we set up an inbox for admissions-related communications. However, the team member who was responsible for the inbox failed to check it. The mistake was discovered after 36 days, a week before the event. While we responded to every message, many people waited a long time before hearing back, some of them having sent follow-up emails when they didn’t receive a timely response.

Current status: This inbox is now monitored by multiple team members.

Groups support

Failure to deliver an EA groups platform (2017 - 2018)

How we fell short: In 2017, we publicly committed to providing an online platform for EA groups which we never launched. We failed to communicate delays to the project and its cancellation in a timely fashion.

Current status: The EA groups platform was cancelled in 2018, though we think many of the planned features are present on the EA Forum’s groups platform.

Poor communication and missed commitments around Community Building Grants (2018 - 2021)

How we fell short: We initially overestimated how much support we'd be able to provide to groups through our Community Building Grants (CBG) program, and made commitments that we weren't able to keep.

Specific mistakes:

  • In late 2018 and 2019 we were overly optimistic about the speed with which we could launch the CBG program, which resulted in substantially less money being granted than we had anticipated, including in public communications about the program.
  • Many CBG recipients expected to receive more non-monetary support (e.g. coaching or professional development) than we were able to provide with our limited staff capacity. We think this is because we were too optimistic about our capacity in private communication with recipients, and also because of public statements like "we want to be able to spend more time on [...] providing support for existing grantees" (source).
  • In July 2020, we shared in a public post that we expected to open applications for our Community Building Grants program "around January 2021". We eventually decided to deprioritize this and push back the date. However, we didn't communicate any information about our timeline until March 2021. Several group leaders expressed their disappointment in our communication around this. While we believe we made the right decision in not reopening the program, we should have shared that decision with group leaders much earlier than we did.

Current status: By default we no longer share specific information about the timelines or budgets of our projects with others until we are reasonably confident that we can fulfil our commitments.

In May 2021, we published an update on our plans, shared this update directly with group organizers, and updated our page to make clear which types of application we are currently reviewing.

We've also selected a set of specific locations to serve through the CBG program, which we believe will allow us to provide better support to grantees without exceeding our staff capacity.

Finally, we've partnered with the Effective Altruism Infrastructure Fund, which will assess funding applications from groups not eligible for the CBG program. This gives any group the chance to apply for support, even if they aren't in one of the specific locations we currently serve through the program.

EA Grants

This post about EA Grants offers more information about the program’s problems and the steps we took to address them.

In early 2020, we closed the EA Grants program.

Poor grant administration (2017 - 2019)

How we fell short: Our record-keeping was poor and our grant transfers were slow. We did not maintain well-organized records of grant applications or decisions. Our staff sometimes verbally promised grants without fully documenting the decision in our system. The lack of a proper grant administration system, along with limited operational capacity, meant we were not always able to transfer payments promptly. As a result, some grantees waited an inordinately long time for payment. We are aware of cases where this contributed to difficult financial or career situations for recipients.

Steps we took to improve: In 2018, we hired a full-time grantmaker on EA Grants who communicated more clearly with applicants and recipients, and who cleaned up aspects of the project that had become disorganized. By mid-2019, CEA was able to disburse EA Grants within a month of receiving recipient bank details in most cases.

Slow communication with applicants, missed commitments (2017 - 2019)

How we fell short: When we initially opened the Grants program in 2017, we did not make most potential applicants aware until shortly before the application deadline. We were also slow in communicating with applicants about delays in processing their applications.

Several times during 2018, we said we would open a Grants round at a particular time and then did not do so. This led the community to have a mistaken impression of whether other funding mechanisms were needed.

Steps we took to improve: By default we no longer share specific information about the timelines or budgets of our projects with others until we are reasonably confident that we can fulfil our commitments.

Lack of post-grant assessment (2017 - 2019)

How we fell short: EA Grants did not conduct a public assessment of grant performance, despite publicly committing to doing so multiple times. This was despite our acknowledgement of the difficulties we faced evaluating grant applications. Ongoing assessment of grant rounds and individual grants would very likely have produced a more effective programme.

Steps we took to improve: Our recognition that other grantmakers had established better processes for grant evaluation informed by grant assessment contributed to our decision to close EA Grants in 2020.

Effective Altruism Funds

In 2020, we spun out EA Funds, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part).

Failure to provide regular updates (2016 - 2019)

How we fell short: We did not always provide regular updates to donors informing them about how their donations have been used.

Steps we’ve taken to improve: We began to send email updates to a fund’s donors each time that fund disbursed grants. In 2020, we spun out Giving What We Can, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part) board.

Outdated fund balances (2016 - 2019)

How we fell short: Fund balances were often outdated between 2016 and 2019. In October 2018, we implemented a new website feature which pulled data from our accounting system that indicated the current balance held for each Fund. While this was intended to be up-to-date and accurate, the delay between when payout reports were published and when these payouts were reflected in our accounting system sometimes meant that the balances were inaccurate. This specific problem has now been fixed, but we still expect that there will sometimes be accounting corrections and adjustments that cause discrepancies in the balances.

Steps we’ve taken to improve: We set up a system that entered payout information into our accounting system at the same time that we publish Fund payout reports, rather than doing an end-of-month reconciliation once all grant payments had been made, as we were doing before. In 2020, we spun out EA Funds, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part).

Inconsistent risk communication (2016 - 2019)

How we fell short: Until the end of 2019, we were often inconsistent in our communication about the level of risk involved with the EA Funds (whether a particular fund aims to recommend grants with a solid evidence base vs. speculative grants with a low certainty of success but a chance of a high upside). This has led to some donors being surprised or unhappy with some grant recommendations.

Steps we’ve taken to improve: We added risk profiles to each EA Fund, to give a broad sense of how much risk their management teams will likely take on (more info). In 2020, we spun out EA Funds, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part).

Slow grant disbursement and communication with grantees (2017 - 2018)

How we fell short: Our grant disbursement and communication was slow from 2017-2018. During the first 21 months after EA Funds began operating (2017-2018), some funds disbursed donations more rarely than was expected. Community members requested more transparency about how much money was in each fund and when it would likely be disbursed. In April 2018, we stated in a comment on the EA Forum that we expected to publish a post with more information in the next few weeks, but we did not actually publish more information until August 2018 (after a community member published an additional post to express their concerns).

Grant disbursement was slow in part because each fund initially had only one fund manager, with one person (Nick Beckstead) managing both the Long-Term Future Fund and the EA Infrastructure Fund (then called the EA Community Fund). Limited capacity by grantmakers meant some funds did not consider small, new projects for funding, which was originally one of the functions we hoped EA Funds would provide.

Steps we’ve taken to improve: In October 2018, we announced new management teams for most of the funds, increasing the number of people available to consider grants from those funds. At the same time, we moved to a three-times-yearly regular grantmaking schedule for three of the funds.

In 2020, we spun out EA Funds, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part).

Technical errors (2017 - 2018)

How we fell short: In 2017, technical mistakes led to errors which impacted donors and donations, including an error with our recurring payments system, which meant thousands of dollars of donations failed to be processed, and an error in our message queue system, which meant some donations were processed twice.

Steps we’ve taken to improve: We informed donors immediately these errors were identified, and implemented both technical and process corrections to prevent them recurring. The issue was fixed in May 2017.

How we fell short: In 2019, some grants recommended (but not made) by the managers of the Long-Term Future Fund were outside the implicit scope of the Fund. These grants were not ultimately made, but the recommendations were communicated to the public before CEA had completed our due diligence. This was confusing to donors. Additionally, we had some disagreements with the grant team about the scope of the fund, which slowed down the approval of some grants and caused delays for potential grantees.

Steps we’ve taken to improve: We introduced a more thorough and prompt internal review process for grants, with additional due diligence steps. Due diligence is now completed before grant recommendations are published, so that it’s clear to donors whether a grant will be made by the fund or not. We have also worked with Fund managers to clarify the scope of each Fund and the process we follow for assessing recommendations. In 2020, we spun out EA Funds, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part).

Inaccurate presentations of the project (2017 - 2018)

How we fell short: We presented the project in a more favourable light than was justified at the time. These included ex-post lowering the bar for success by this metric in the project’s first 18 months, as well as prematurely marketing the project in prominent locations such as effectivealtruism.org and Giving What We Can, prematurely announcing a workplace giving feature which was never launched, and downplaying critical feedback from the community.

Steps we’ve taken to improve: We addressed these presentation mistakes. In 2020, we spun out EA Funds, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part).

Giving What We Can

Mixed messages and unclear communication about the Pledge (2014 - 2017)

How we fell short: We encouraged student groups to run pledge drives which sometimes resulted in people taking the Pledge at a young age without being encouraged to seriously think it through as a long-term commitment. Some of our communications also presented the Pledge as something to be taken quickly rather than carefully considered.

Steps we took to improve: We updated GWWC's website and pledge campaign materials to emphasize that the Pledge is a career-long commitment to consider carefully rather than in haste. We recommend the temporary Try Giving pledge to people who do not yet have a clear picture of their financial future.

How we fell short: Our communications about Giving What We Can did not always make some points clear (for example, whether the Pledge is completely binding regardless of personal circumstances, or whether you can end your pledge.) This resulted in confusion in the community.

Steps we took to improve: In 2017, we added more information to our FAQ about this topic, as well as a post clarifying the nature of the Pledge. In 2020, we spun out Giving What We Can, which now operates independently of CEA and reports directly to the board of Effective Ventures (the legal entity of which we are a part).

Not following through on Giving What We Can (2016 - 2020)

How we fell short: We view it as a mistake that we put relatively little staff time toward Giving What We Can after merging it with CEA in 2016, which likely caused slower growth in the number of people taking the pledge and the amount of money donated than could have been achieved by a dedicated team. The main person responsible for Giving What We Can from 2017-2020, Julia Wise, already had full-time responsibilities working on community health. In addition, technical capacity was split between several major projects, so fixes to Giving What We Can tech problems sometimes took longer than they should have. Members received fewer communications than before, and website functionality was not what users expected. For example:

  • The Pledge signup form was broken for two weeks in 2017.
  • After an update to the platform where members recorded their donations, some functionality that had been present in the old version, such as the ability to record recurring donations, was not added to the new version for more than a year.
  • The “How rich am I?” calculator was based on increasingly outdated statistics until an update in 2019.

Current status:

  • The platform now provides all the functionality of the old donation recording platform, plus additional features like the ability to see progress both during a set time period (such as a year) and over the course of one's lifetime.
  • With help from Rob Wiblin and Phil Trammell, we updated the “How rich am I?” calculator in 2019 to use more recent data.
  • In 2020, we hired Luke Freeman to run Giving What We Can. He revamped the website, improved communications with members, and began to hold events for members to help the organization grow and develop to a greater extent. Giving What We Can now operates independently of CEA, with Luke as Executive Director.

Pareto Fellowship (2016)

How we fell short: The Pareto Fellowship was a program sponsored by CEA and run by two US-based CEA staff, designed to strengthen the skills and professional network of promising students or people early in their careers. At the time, CEA’s leadership was not providing much oversight of US-based staff and was not closely following most of the team’s work. We realized during and after the program that CEA’s leadership should have recognized some of these problems and intervened sooner.

  • Reports by some applicants indicate that the interview process was unprofessional and made them deeply uncomfortable.
  • Partly because of tensions between different parts of CEA at the time, the program was run in a somewhat different way from how it was originally presented to applicants. The location and a significant amount of the programming were provided by Leverage Research / Paradigm Academy. In general participants rated this content well on feedback forms, but some were not expecting this degree of involvement and felt uncomfortable with it.
  • A staff member leading the program appeared to plan a romantic relationship with a fellow during the program. Due to the power dynamics involved, we think this was unwise and could have been harmful, though as far as we know it did not cause harm in this case and did not break our formal policy. It also may have made some participants uncomfortable.

We had already decided not to continue the programme and the staff involved had left CEA by the time we committed to sharing a “detailed review of the Pareto Fellowship”. It was a mistake both to make this commitment publicly and to fail to deliver a review at that time. Given the time that has passed, we think it would no longer be worthwhile to invest substantial time and resources into a retrospective review.

Changes since that time:

  • We did not continue the program after 2016.
  • For a variety of reasons mostly unrelated to this program, neither the staff who directly ran the program nor the management staff who oversaw it still work at CEA.
  • Although we had a written conflict of interest policy at that time, we weren’t as clear about the practical application as we would be now. It’s hard for a formal policy to encompass every real-life situation that comes up, and we now have a stronger policy and stronger norms about flagging possible problems around power dynamics and conflicts of interest.
  • We now use a standard assessment process for hiring and other application rounds (such as grants), which is designed to be more comfortable for applicants.

Background on CEA’s relationship with Leverage Research: In the past, CEA collaborated on a few projects with Leverage Research and Leverage’s sister organization Paradigm Academy. Leverage began the EA conference series called the EA Summit, which they passed to us in 2015 and we expanded and turned into the EA Global conference series. They provided help and input on the first few years of EA Global, and provided work space that CEA used for EA Global and a team retreat. A few staff worked on projects for CEA and Leverage at the same time (though none of these staff are still with CEA). We were one of the fiscal sponsors of their 2018 EA Summit. CEA’s and Leverage’s joint involvement in the Pareto Fellowship is described above. There were more minor collaborations, such as them having a table at the careers fair at EA Global several times.

Our joint work with Leverage was a source of significant disagreement among CEA staff. As there was turnover in our staff and leadership, we decided not to work with them again, and haven’t worked with them since 2018. Our understanding is that Leverage no longer works on EA community building.

Communications for Effective Altruism

Effectivealtrusim.org (2017 - 2022)

How we fell short: CEA is responsible for effectivealtruism.org, which is one of the most prominent online resources introducing EA. However, for several years promoting the website, including through search engine optimization, was not a priority for us. Prior to 2022, the website was updated infrequently, giving an inaccurate impression of the community and its ideas as they changed over time.

Current status: In November 2021 we hired a product manager with responsibility for effectivealtruism.org. Between February and July 2022 we completed a major overhaul of its design and content, updatingthe introductory essay (What is effective altruism?).

External communications (2021 - 2022)

How we fell short: For several years, we’d been doing “reactive” communications work (responding to media requests, etc.) In mid-2021 we began to think that it was important to focus more on proactive external communications for effective altruism. We tried to initiate work on this, but for various reasons this work was delayed by around 9 months, which cost valuable time. We think we could and should have found ways to get this work going more quickly.

Current status: Our Community Health team has convened and led an EA network responding to ongoing media activity and identifying risks. We have worked with Forethought, Open Philanthropy and experienced communications professionals to develop a communications strategy for EA and longtermism. In July 2022 we hired a Head of Communications to lead efforts to accurately communicate EA ideas, including in the media.

Community Health

Confidentiality mistakes (2018 and 2021)

Julia Wise serves as a contact person for the EA community, including handling situations where she has promised confidentiality to people who contact her. In two cases that she knows of, she has accidentally broken confidentiality. She explains in more detail here.

Additionally, we should have added this information to Julia’s post about her role as a contact person earlier.

EA Librarian (2022)

How we fell short: We failed to adequately communicate the status of this project, including its discontinuation. We launched this service as an experiment in January 2022, with the goal of providing answers to frequently asked questions about EA. We also said that “We will aim to publish a thread every 2 weeks with questions and answers that we thought were particularly interesting or useful (to either the community or the question asker).”, but subsequently took over 3 weeks to respond to one question. We communicated privately with people who had submitted a question to say that we were behind.

We felt the experiment wasn’t successful and we discontinued it in May 2022, and updated the relevant tag on the EA Forum, and closed the form, but we failed to announce this beyond a footnote in a team update on the Forum.

Current status: The Librarian project remains inactive.

Another perspective

A thorough criticism of our work was published in September 2022 as part of the EA Criticism and Red Teaming Contest.