# GiveWell

> Source: https://aiwiki.ai/wiki/givewell
> Updated: 2026-06-24
> Categories: AI Safety, Research Organizations
> From AI Wiki (https://aiwiki.ai), a free encyclopedia of artificial intelligence. Quote with attribution.

**GiveWell** is an American nonprofit charity evaluator, founded in 2007 by [Holden Karnofsky](/wiki/holden_karnofsky) and [Elie Hassenfeld](/wiki/elie_hassenfeld), that uses cost-effectiveness analysis to recommend a short list of evidence-backed global health and poverty charities, and that has directed more than $2.6 billion to its recommended organizations as of September 2025.[1][5] Its two founders were former hedge fund analysts at [Bridgewater Associates](/wiki/bridgewater_associates) who wanted to know which charities did the most measurable good per dollar and could not find an answer. The organization is legally registered as The Clear Fund, a 501(c)(3) public charity, and directed about $397 million in metrics year 2024 alone.[5][7]

GiveWell matters to the AI safety story for one specific reason: it incubated [Open Philanthropy](/wiki/open_philanthropy), which spun out as a separate organization in June 2017 and went on to become one of the largest funders of AI safety research, including a $30 million grant to [OpenAI](/wiki/openai) in 2017 and major early backing for [Anthropic](/wiki/anthropic), [Redwood Research](/wiki/redwood_research), the [Alignment Research Center](/wiki/alignment_research_center), and [MIRI](/wiki/miri).[12][18] The two organizations are now legally separate and have different missions, but they share an intellectual lineage and overlapping personnel. Daniela Amodei, president and co-founder of Anthropic, currently sits on GiveWell's board.[4]

## Quick facts

| Item | Detail |
|---|---|
| Founded | 2007 (incorporated as The Clear Fund) |
| Founders | Holden Karnofsky, Elie Hassenfeld |
| Type | 501(c)(3) public charity |
| Headquarters | Oakland, California (incorporated in New York; relocated to San Francisco in 2013) |
| Chief Executive Officer | Elie Hassenfeld |
| Board Chair | Timothy Ogden |
| Cumulative directed (through Sep 2025) | More than $2.6 billion[5] |
| 2024 metrics year directed | $397 million[5] |
| Current top charities (2025) | Against Malaria Foundation, Malaria Consortium, Helen Keller International, New Incentives |
| Spun out | Open Philanthropy (June 2017) |

## When was GiveWell founded, and by whom?

Karnofsky and Hassenfeld met in 2003 at Bridgewater Associates, the Connecticut hedge fund founded by Ray Dalio. In 2006, they and six friends in finance started looking into how to give away their own money effectively. They were not satisfied with what existing charity rating sites were measuring (mostly overhead ratios and financial transparency, not whether programs actually worked) and assumed there had to be deeper analysis somewhere. They could not find any.[1]

In mid-2007, with about $300,000 raised from coworkers and themselves, Karnofsky and Hassenfeld quit Bridgewater and incorporated The Clear Fund.[1][2] The plan was to research charities full time, publish the work openly, and direct the Clear Fund's pooled donations to whichever charity looked best. The organization was based in New York City for its first several years before relocating to San Francisco in 2013, partly to be closer to Silicon Valley supporters of effective altruism. The current headquarters is in Oakland.

Late 2007 brought the organization's first scandal. Karnofsky was caught using sockpuppet accounts on internet forums (most prominently Ask MetaFilter) to post questions about where to donate under one pseudonym ("geremiah") and then answer them with recommendations for GiveWell under another ("Holden0"). MetaFilter members exposed the pattern within days. The board investigated, demoted Karnofsky from executive director to program officer, and required him to undertake professional development paid for by a $5,000 deduction from his salary. Karnofsky published a detailed apology titled "I had a lapse in judgment, did a horrible thing, and I apologize" on the GiveWell blog on December 31, 2007.[14][15] He was later reinstated as co-executive director. The episode is still cited by critics, but GiveWell's response (an open investigation, public board minutes, the salary penalty) became a kind of template for how the organization handled mistakes afterward.

The [William and Flora Hewlett Foundation](https://hewlett.org) was an early funder via its Nonprofit Marketplace Initiative. In late 2008, GiveWell received a one-year $100,000 grant under that initiative, which the organization later credited as decisive in carrying it through the post-scandal period. Hewlett continued to support GiveWell until March 2014, when it announced it was ending the Nonprofit Marketplace Initiative after research suggested only about 3 percent of donors selected charities based on performance metrics.[21] By the early 2010s, GiveWell had become closely identified with the [effective altruism](/wiki/effective_altruism) movement that was emerging out of Oxford philosophy circles around [William MacAskill](/wiki/william_macaskill) and Toby Ord, and out of [Peter Singer](/wiki/peter_singer)'s utilitarian ethics. The two communities were separate in origin but converged on similar questions: which causes do the most good, and how would you actually know?

## How does GiveWell evaluate charities?

GiveWell's evaluation method is built around cost-effectiveness analysis. For each charity it considers, the research team tries to estimate, in dollar terms, how much good a marginal donation produces, usually in a unit like "deaths averted" or "years of doubled consumption." The estimates are public, the spreadsheets are public, and the underlying judgment calls (moral weights, discount rates, assumptions about counterfactual funding) are written out in narrative form.

A few principles run through the work:

- **Programs first, charities second.** GiveWell starts by identifying interventions with strong evidence behind them, ideally from [randomized controlled trials](/wiki/randomized_controlled_trial), and then looks for organizations delivering those interventions well at scale.
- **Cost per life saved.** For interventions like long-lasting insecticide-treated bed nets and seasonal malaria chemoprevention, GiveWell publishes cost-per-death-averted figures. GiveWell's models for grants to its top charities have estimated an average cost between roughly $3,000 and $5,500 per life saved over the past several years, depending on the program and a long list of moral and empirical assumptions; for example, it estimated an average of about $4,000 per life saved for seasonal malaria chemoprevention grants recommended in 2022 to 2024.[23] GiveWell notes that "many people are surprised at how expensive it can be to save a life, even for the most cost-effective programs," because not all recipients would have died without the intervention.[23]
- **Room for more funding.** A charity can be excellent and still be a bad place to send the next dollar if it is already fully funded for everything it can absorb. GiveWell tries to estimate how much each top charity can productively use.
- **Transparency about mistakes.** GiveWell maintains a "Mistakes" page on its website that catalogs errors in published analyses, going back to 2008.[13] Hassenfeld has said the practice is borrowed in spirit from Bridgewater's culture of writing down disagreements.

The approach has its critics. Philosopher Leif Wenar wrote a 2011 essay ("Poverty Is No Pond") arguing that GiveWell systematically understated potential harms from recommended charities and underplayed how hard it is to know what is actually happening on the ground in distant aid programs.[16] Other critics have argued that GiveWell's preference for legible, measurable interventions biases it against systemic causes (governance reform, advocacy, structural change) that are harder to quantify but might matter more. GiveWell has generally engaged with these criticisms in writing rather than dismissing them, which is part of why effective altruists tend to like the organization in the first place.

## What are GiveWell's top charities?

GiveWell publishes a short list of "top charities" each year. The list is meant to be a recommendation for individual donors, not a comprehensive ranking of everything good in the world. As of 2025, the top charities list contains four organizations, all working on global health interventions in sub-Saharan Africa or other low-income regions.[10]

### Current top charities (2025)

| Charity | Intervention | Approximate cost per death averted |
|---|---|---|
| Against Malaria Foundation | Long-lasting insecticide-treated bed nets | Roughly $5,500[23] |
| Malaria Consortium | Seasonal malaria chemoprevention for children | Roughly $4,000[23] |
| Helen Keller International | Vitamin A supplementation programs | Roughly $3,500 |
| New Incentives | Conditional cash transfers for routine childhood vaccinations in Nigeria | Roughly $5,000 |

The figures are GiveWell's own and depend on a long chain of assumptions about counterfactuals, the contribution of other donors, and moral weights between adults and children. They are best understood as comparison points within GiveWell's framework rather than precise truths.

### Historical top charities

The list has shifted over time as evidence has accumulated and as some charities have grown beyond their funding gaps. A non-exhaustive history:

| Period | Notable additions or removals |
|---|---|
| 2009 to 2011 | VillageReach (vaccine logistics in Mozambique) was the #1 recommendation |
| 2011 | [Against Malaria Foundation](/wiki/against_malaria_foundation) and Schistosomiasis Control Initiative (SCI) added |
| 2012 | [GiveDirectly](/wiki/givedirectly) added; the organization gives unconditional cash transfers to people in extreme poverty in Kenya, Uganda, and elsewhere |
| 2013 | Deworm the World Initiative (run by Evidence Action) added |
| 2014 | AMF temporarily removed for not absorbing previous funds quickly enough, then later restored |
| 2017 | Helen Keller International (vitamin A) and Malaria Consortium (SMC) added |
| 2020 | New Incentives added |
| 2022 to 2025 | GiveDirectly removed from the top charities list (still recommended via the All Grants Fund); SCI, END Fund and Sightsavers' deworming work moved to standout or All Grants categories rather than top |

GiveWell also maintains an "All Grants Fund" for higher-risk, higher-uncertainty grants (often through smaller charities or new programs) and a "Top Charities Fund" for donors who want a simple split across the four top recommendations. The Top Charities Fund replaced the earlier "Maximum Impact Fund" branding in 2022.

## How much money has GiveWell directed?

"Money moved" is GiveWell's term for the amount of donations directed to its recommended charities each year, including grants from [Open Philanthropy](/wiki/open_philanthropy) and individual donor gifts that flow through GiveWell. The figures below come from GiveWell's annual metrics reports.[5][6][7]

| Year | Money moved (approximate) | Notes |
|---|---|---|
| 2007 | $1.2 million | First year |
| 2010 | $1.5 million | Pre-effective altruism mainstreaming |
| 2014 | $28 million | Open Philanthropy partnership well underway |
| 2017 | $117 million | Year Open Philanthropy formally separated |
| 2019 | $161 million | |
| 2020 | $244 million | |
| 2021 | $518 million | First year above half a billion |
| 2022 | $595 million | All-time high in raw fundraising |
| 2023 | $355 million | Lower Open Philanthropy contribution this year |
| 2024 | $415 million raised; $397 million directed to programs | 55 grants to 34 organizations across 22 countries |

In metrics year 2024 (February 2024 to January 2025), the four top charities received about $303 million between them: roughly $150 million to Against Malaria Foundation, $87 million to Malaria Consortium, $54 million to Helen Keller International, and $12 million to New Incentives. GiveWell estimated that 2024 grants would help around 34 million people not otherwise reached and avert about 74,000 deaths (a figure that covers only the 17 mortality-focused grants made that year).[5][7]

Cumulatively, as of September 2025, GiveWell reports having directed more than $2.6 billion to recommended organizations.[5] GiveWell estimates that donations to its four current top charities between 2009 and 2024 will avert more than 340,000 deaths.[5] Those figures depend entirely on GiveWell's own cost-effectiveness models and should be read accordingly.

## How is GiveWell related to effective altruism?

GiveWell is one of the institutions most often associated with the effective altruism community, alongside [80,000 Hours](/wiki/80000_hours), the [Centre for Effective Altruism](/wiki/centre_effective_altruism), and Open Philanthropy. The relationship is complicated. GiveWell predates the effective altruism label by a few years and was not founded as an EA project, but as effective altruism took off around 2011 to 2014 the two communities became deeply intertwined. Most large EA-aligned individual donors give to GiveWell-recommended charities, and GiveWell hires and is hired from EA-adjacent labor markets.

GiveWell's leadership has consistently positioned the organization as a charity evaluator, not a movement organization. Hassenfeld has emphasized that GiveWell's recommendations are about evidence and cost-effectiveness in global health and poverty, not about EA's broader cause-prioritization claims. In practice this means GiveWell does not make grants in animal welfare, [AI safety](/wiki/ai_safety), or [existential risk](/wiki/existential_risk) reduction, even though those are major cause areas in the EA movement. Open Philanthropy was created in part to do that other work without forcing GiveWell to take positions on it.[3]

### The FTX collapse and GiveWell

The November 2022 collapse of FTX, the cryptocurrency exchange founded by [Sam Bankman-Fried](/wiki/sam_bankman_fried), brought the FTX Future Fund and its EA-aligned grantmaking into bankruptcy. The Future Fund had directed more than $190 million to longtermist and AI-related causes since launching in February 2022, and its grant program was sometimes treated as a parallel to Open Philanthropy in the AI safety field. After the collapse, the Effective Ventures group (which housed the [Future Fund](/wiki/ftx_future_fund) and several adjacent EA organizations) settled with the FTX bankruptcy estate for roughly $26.8 million in 2024, equal to the full sum it had received from FTX entities in 2022. GiveWell itself did not receive significant FTX funding and was not a direct party to the clawback proceedings, but the affair damaged the EA brand to which GiveWell was popularly attached and prompted internal discussion at GiveWell about how to manage reputational risk from movement-adjacent funders.

## How did Open Philanthropy spin out of GiveWell?

The story of [Open Philanthropy](/wiki/open_philanthropy) is almost more important to the [AI Wiki](/wiki/ai_wiki) than the story of GiveWell itself, because Open Philanthropy is where the AI safety money came from.

In 2011, [Cari Tuna](/wiki/cari_tuna) and [Dustin Moskovitz](/wiki/dustin_moskovitz) (Moskovitz is a co-founder of Facebook and Asana, and Tuna is a former Wall Street Journal reporter) approached GiveWell. They had recently founded [Good Ventures](/wiki/good_ventures), a private foundation for their philanthropy, and were thinking about how to give away what would eventually become roughly $14 to $15 billion in Facebook and Asana wealth. They had been inspired by Peter Singer's *The Life You Can Save* and were among the youngest signatories of the Giving Pledge. They wanted GiveWell's help.

Good Ventures gave heavily to GiveWell's existing top charities, but the partners also wanted to explore causes outside global health where the evidence base was thinner and the bets were larger. That joint exploration was initially called "GiveWell Labs," was rebranded to the **Open Philanthropy Project** in 2014, and operated as a collaborative project between GiveWell and Good Ventures.[3][19] Holden Karnofsky led the work.

On June 1, 2017, Open Philanthropy formally separated from GiveWell and became an independent limited liability company.[18] GiveWell sold assets and transferred staff to Open Philanthropy Project LLC, the new entity. The reasons were both legal (Good Ventures was funding very large grants in policy and other areas where the GiveWell brand did not fit cleanly) and structural (the two missions were diverging: Open Philanthropy targeted a few major philanthropists rather than many individual donors, and used an "important, neglected, tractable" framework rather than the evidence-and-scalability frame GiveWell continued to apply). The two organizations continued sharing office space for a transition period.

In late 2019 the spinoff dropped "Project" from its name and became simply Open Philanthropy. In November 2025 it rebranded again, this time as Coefficient Giving, to reflect its growing role advising multiple donors beyond Good Ventures.[8][20] By that point Open Philanthropy had directed more than $4 billion in cumulative grants and had committed more than $100 million in 2024 from donors other than Good Ventures, with non-Good-Ventures grantmaking more than doubling in 2025.[8]

GiveWell and Open Philanthropy / Coefficient Giving remain entangled. Cari Tuna is on GiveWell's board. Hassenfeld is on Coefficient Giving's board of managers. Coefficient Giving and Good Ventures together provide operational support to GiveWell, capped at 20 percent of GiveWell's operating expenses.[3] Daniela Amodei, president of Anthropic and sister of Anthropic CEO Dario Amodei, also sits on GiveWell's board.[4]

### How does GiveWell connect to AI safety funding?

Open Philanthropy's AI safety grantmaking is the main reason GiveWell appears in AI history at all. As of June 2025, Open Philanthropy / Coefficient Giving had directed more than $4 billion in cumulative grants across all causes, of which roughly $336 million had gone toward reducing risks from advanced AI since 2017, making it the largest single private funder of AI safety research by a wide margin.[8] The most widely cited grants include:

- **OpenAI, March 2017.** Open Philanthropy granted $30 million to OpenAI ($10 million per year for three years) as general support, at a time when OpenAI was a research nonprofit.[12] As part of the deal, Holden Karnofsky joined OpenAI's board. Open Philanthropy described the grant's primary goal as "to make Open Philanthropy and its staff likely to be more influential, and more involved, in OpenAI's work over the coming years, in order to improve the likelihood that this work is beneficial."[12] The grant has been controversial in retrospect, both because OpenAI later restructured to a capped-profit company and because some critics argued the funding helped accelerate frontier AI capabilities work rather than safety work.
- **Anthropic.** Open Philanthropy and entities aligned with the EA donor base were among the early backers of Anthropic, founded in 2021 by Dario Amodei, Daniela Amodei, and other former OpenAI researchers. Public reporting and EA Forum threads indicate Open Philanthropy and individuals connected to Moskovitz made substantial early investments. Specific dollar amounts have been the subject of community debate; readers should treat figures circulating online with care.
- **Redwood Research, ARC, MIRI, and others.** Open Philanthropy has been a major supporter of AI alignment and safety organizations including Redwood Research, the Alignment Research Center (founded by Paul Christiano), the Machine Intelligence Research Institute, the Center for Human-Compatible AI at Berkeley, the Center for Security and Emerging Technology at Georgetown, and a range of academic alignment groups.
- **Field building.** Open Philanthropy has funded AI safety fellowships, scholarships, and undergraduate-to-research pipelines through programs run with 80,000 Hours, MATS, and others.

The practical consequence is that a large fraction of the early funding for what became the AI safety field flowed, ultimately, from a single donor couple (Tuna and Moskovitz) through a single advisor (Open Philanthropy) that itself grew out of GiveWell. The community's later concern about funding concentration in EA and AI safety is largely about exactly this lineage.

## Controversies and criticisms

A few recurring criticisms of GiveWell are worth noting:

- **Sockpuppet incident, 2007.** Already discussed. Used by skeptics to argue GiveWell is institutionally untrustworthy; used by defenders to argue GiveWell is one of the few nonprofits that handles its own scandals openly.
- **Narrow methodology.** Wenar's 2011 critique and successors have argued that GiveWell's emphasis on quantifiable interventions undervalues structural causes (governance reform, education systems, advocacy) where measurement is harder.[16]
- **Cost-per-life-saved estimates.** GiveWell publishes specific dollar figures for life-saving impact that are widely repeated. Critics, including some sympathetic ones, point out that these figures contain large hidden assumptions about discount rates, moral weights, and counterfactual funding, and that the precision they convey can be misleading.
- **EA entanglement.** As effective altruism faced reputational damage following the 2022 collapse of FTX (Sam Bankman-Fried's exchange) and the related criminal case, GiveWell faced questions about whether it had taken FTX-related funding and how it thought about the EA community's reputational risks. GiveWell's response was to publish detailed accounting of any FTX-linked donations and clawback exposure.
- **Concentration of funding.** As discussed above, GiveWell's growth has been heavily underwritten by Open Philanthropy / Good Ventures, which itself depends on the Moskovitz / Tuna fortune. This creates real questions about the resilience of the field if that funding source were to change direction.
- **Longtermist disagreement.** Within the broader EA community, GiveWell's continued focus on near-term measurable global health interventions has drawn pushback from longtermists who argue that the highest-expected-value bets concern existential risks, AI safety, and biosecurity rather than malaria nets. GiveWell has responded that it specializes in cause areas where evidence is strong and that other organizations (including Open Philanthropy) are better placed to handle longtermist work.

## Current operations and leadership

GiveWell is headquartered in Oakland, California. The organization had roughly 70 to 90 staff as of the mid-2020s, depending on how contractors are counted, and operates with a board of directors and a research team that publishes most of its work openly on the GiveWell website.

### Leadership (as of 2025)

| Name | Role |
|---|---|
| Elie Hassenfeld | Co-founder and Chief Executive Officer |
| Teryn Mattox | Chief Research and Program Officer |
| Anna Weldon | Chief of Staff |
| Steven Profaizer | Vice President of Communications |
| Lisa McCandless | Vice President of Philanthropy |
| Sai Jahann | General Counsel |

### Board of directors (as of 2025)

| Name | Notes |
|---|---|
| Timothy Ogden (Chair) | Managing Director, Financial Access Initiative |
| Elie Hassenfeld | Co-founder and CEO of GiveWell |
| Cari Tuna | Co-founder and Chair of Good Ventures; Chair of Coefficient Giving |
| Daniela Amodei | President and co-founder of Anthropic |
| Norma Altshuler | Coefficient Giving global aid policy lead |
| Albert Chu | Operating Partner, Hanover Investors |
| James McClave | Former trader, Jane Street |

Karnofsky stepped down from day-to-day leadership of GiveWell in 2017 to focus on Open Philanthropy. He served as CEO and later co-CEO of Open Philanthropy until 2023, when he stepped back from that role and shifted into a Director of AI Strategy position.[9] In April 2024 he became a Visiting Scholar at the Carnegie Endowment for International Peace, working on international security risks from advanced AI. He joined [Anthropic](/wiki/anthropic) in January 2025 as a member of technical staff, working on the company's Responsible Scaling Policy and other safety planning ahead of highly advanced AI systems, and reporting to chief science officer Jared Kaplan.[9][22][24] An Anthropic spokesperson said he "is working on the responsible scaling policy and other aspects around safety planning ahead of highly advanced AI systems."[24] The move drew attention in part because Karnofsky is married to Anthropic president Daniela Amodei.[24]

## Trust and transparency

If one feature distinguishes GiveWell from typical nonprofits, it is the published paper trail. The GiveWell blog has been running since 2007 and contains long-form analyses, conversation notes with grantees, internal debates, and the "Mistakes" page.[13] Board meeting minutes are public. Budgets are published. Cost-effectiveness models are downloadable spreadsheets that any reader can copy and modify with different assumptions.

This culture is partly a reaction to the 2007 sockpuppet incident and partly a deliberate import from Bridgewater's documented-disagreement style. Critics argue the culture sometimes shades into self-flagellation or into a pretense of total transparency that actually elides the most consequential judgment calls. Supporters argue it is the closest thing the nonprofit sector has to a falsifiable research program. The tension is unresolved and probably constitutive of the organization.

## Why is GiveWell on the AI Wiki?

For readers coming to this article from articles on AI safety or [Anthropic](/wiki/anthropic): GiveWell itself does not fund AI work, and its top charity recommendations have nothing to do with AI. The reason GiveWell is in this wiki is the founding chain. GiveWell convinced Tuna and Moskovitz to give effectively in 2011, that partnership produced GiveWell Labs and later Open Philanthropy, Open Philanthropy provided much of the early institutional money for AI safety research from roughly 2015 onward, and several of the people involved (Karnofsky, the Amodei siblings, Paul Christiano) moved on to lead or seed major AI safety organizations including Anthropic, ARC, and Redwood Research. The AI safety field as it existed in the late 2010s and early 2020s was, to a significant extent, an outgrowth of the same people, money, and intellectual habits that produced GiveWell.

## See also

- [Open Philanthropy](/wiki/open_philanthropy)
- [Effective altruism](/wiki/effective_altruism)
- [Anthropic](/wiki/anthropic)
- [Holden Karnofsky](/wiki/holden_karnofsky)
- [Dustin Moskovitz](/wiki/dustin_moskovitz)
- [Good Ventures](/wiki/good_ventures)
- [GiveDirectly](/wiki/givedirectly)
- [80,000 Hours](/wiki/80000_hours)
- [AI safety](/wiki/ai_safety)

## References

1. "Our Story." GiveWell. https://www.givewell.org/about/story
2. "GiveWell." Wikipedia. https://en.wikipedia.org/wiki/GiveWell
3. "The Relationship Between GiveWell and Coefficient Giving." GiveWell. https://www.givewell.org/about/gw-op-relationship
4. "Our People." GiveWell. https://www.givewell.org/about/people
5. "GiveWell's 2024 Metrics and Impact." The GiveWell Blog, August 13, 2025. https://blog.givewell.org/2025/08/13/givewells-2024-metrics-and-impact/
6. "GiveWell Metrics Report, 2023 Annual Review." GiveWell. https://files.givewell.org/files/GiveWell_Metrics_Report_2023.pdf
7. "GiveWell Metrics Report, 2024." GiveWell. https://files.givewell.org/files/metrics/GiveWell_Metrics_Report_2024.pdf
8. "Coefficient Giving." Wikipedia. https://en.wikipedia.org/wiki/Coefficient_Giving
9. "Holden Karnofsky." Wikipedia. https://en.wikipedia.org/wiki/Holden_Karnofsky
10. "Top Charities." GiveWell. https://www.givewell.org/charities/top-charities
11. "Top Charities, November 2012 Archived Version." GiveWell. https://www.givewell.org/charities/top-charities/November-2012
12. "OpenAI: General Support." Open Philanthropy, March 2017. https://www.openphilanthropy.org/grants/openai-general-support/
13. "Mistakes." GiveWell. https://www.givewell.org/about/official-records/mistakes
14. "I had a lapse in judgment, did a horrible thing, and I apologize." The GiveWell Blog, December 31, 2007. https://blog.givewell.org/2007/12/31/i-had-a-lapse-in-judgment-did-a-horrible-thing-and-i-apologize/
15. "Full account of GiveWell staff's inappropriate online promotion in December 2007." GiveWell. https://www.givewell.org/about/official-records/board-meeting-3/FAQ-on-inappropriate-marketing/full-account
16. Wenar, Leif. "Poverty Is No Pond: Challenges for the Affluent." In Giving Well: The Ethics of Philanthropy, Oxford University Press, 2011.
17. "How Dependent is the Effective Altruism Movement on Dustin Moskovitz and Cari Tuna?" EA Forum. https://forum.effectivealtruism.org/posts/4BJSXH9ho4eYNT73P/how-dependent-is-the-effective-altruism-movement-on-dustin
18. "Separating GiveWell and the Open Philanthropy Project." The GiveWell Blog, June 12, 2017. https://blog.givewell.org/2017/06/12/separating-givewell-open-philanthropy-project/
19. "Our History." Coefficient Giving. https://coefficientgiving.org/our-history/
20. "Open Philanthropy Is Now Coefficient Giving." Coefficient Giving, November 2025. https://coefficientgiving.org/research/open-philanthropy-is-now-coefficient-giving/
21. "Thoughts on the end of Hewlett's Nonprofit Marketplace Initiative." The GiveWell Blog, August 5, 2014. https://blog.givewell.org/2014/08/05/thoughts-on-the-end-of-hewletts-nonprofit-marketplace-initiative/
22. "Leaving Open Philanthropy, going to Anthropic." LessWrong. https://www.lesswrong.com/posts/3ucdmfGKMGPcibmF6/leaving-open-philanthropy-going-to-anthropic
23. "How Much Does It Cost to Save a Life?" GiveWell. https://www.givewell.org/how-much-does-it-cost-to-save-a-life
24. "Anthropic hired president Daniela Amodei's husband to work on the company's AI safety strategy." Fortune, February 13, 2025. https://fortune.com/2025/02/13/anthropic-hired-president-daniela-amodei-husband-ai-safety-responsible-scaling/