Do the most good you can with what you have. Effective altruism (EA) applies evidence and reason to the question of how to help others: compare interventions by impact (lives saved, welfare gained per dollar), prioritise the best options, and follow the evidence even when it points away from intuitive or emotionally salient causes. The movement emerged in the 2000s–2010s, drawing on utilitarian ethics, development economics, and prioritisation research. The core ideas are cause neutrality (choose the cause that does the most good, not the one that feels best), marginal impact (your next dollar or hour goes where it helps most), and openness to revision (update as evidence changes).
EA doesn't prescribe a single cause. It prescribes a method: estimate impact, compare, prioritise. In practice, many EAs have focused on global health (e.g. bed nets, deworming, cash transfers), animal welfare (factory farming, wild animal suffering), and long-term risk (AI safety, pandemic preparedness) — not because those are morally special but because rough cost-effectiveness estimates suggest they may offer very high impact per unit of resource. The discipline is to ask: where does my next unit of time or money do the most good? The answer may be counterintuitive (e.g. earning to give, or working on a cause that feels abstract).
Critics argue that EA over-relies on measurable outcomes, underweights non-quantifiable values, and can justify odd conclusions (e.g. that one might do more good in finance than in direct service). The response is that measurement is imperfect but better than intuition alone, and that cause neutrality is a default to be tempered by evidence and values, not abandoned. The strategic use is not to adopt every EA conclusion but to adopt the habit: compare impact, prioritise, and update.
Section 2
How to See It
Effective altruism reveals itself when someone asks "what does the most good?" or "where does my next dollar have the highest impact?" and then seeks evidence to answer. Look for the pattern: cause neutrality (willingness to switch causes), marginal thinking (next unit of resource), and use of cost-effectiveness or similar frameworks. When giving or career decisions are framed in terms of expected impact per dollar or hour, EA reasoning is at work.
Business
You're seeing Effective Altruism when a founder or executive donates a portion of income and chooses charities by impact (e.g. GiveWell top charities, or cause-specific cost-effectiveness) rather than by personal connection or brand. The same logic can apply to corporate giving: where does the next dollar of CSR or foundation spending do the most good? EA pushes toward comparison and evidence.
Technology
You're seeing Effective Altruism when technical talent chooses to work on AI safety or biosecurity because rough models suggest these areas may have very high expected impact, even though the work is abstract and the impact is uncertain. "Earning to give" (high-income job, donate heavily) is another EA pattern: the marginal dollar to top charities may do more good than a lower-paid direct-service role. The frame is impact per unit of effort or capital.
Investing
You're seeing Effective Altruism when an investor allocates part of returns to high-impact causes or uses impact-weighted metrics to compare investments. Impact investing can be EA-adjacent when it explicitly compares impact across options. The discipline is the same: where does the next unit of resource do the most good? The answer may favour proven interventions over fashionable ones.
Markets
You're seeing Effective Altruism when policy or philanthropy is evaluated by cost-effectiveness — lives saved per dollar, DALYs averted — and resources are directed accordingly. The approach can favour underfunded, evidence-backed interventions over salient but low-impact ones. The market doesn't do this by itself; EA is a lens for allocating charitable or public resources.
Section 3
How to Use It
Decision filter
"When allocating time or money to help others, ask: where does the next unit do the most good? Compare options by impact (even roughly). Be willing to switch causes if evidence points elsewhere. Update as you learn. Don't assume the most salient or intuitive cause is the best."
As a founder
You have leverage: capital, platform, talent. EA asks where that leverage does the most good. That might be your core business (if it solves a big problem), your philanthropy (if you give), or your advocacy. The discipline is to compare: what's the impact of the next dollar or hour in each use? You don't have to adopt EA conclusions — but the habit of comparing impact and following evidence can improve decisions. Be aware of scope neglect: we underweight large, distant, or statistical harms. Correct for it when you can.
As an investor
Impact can be part of the thesis. EA doesn't say profit doesn't matter — it says that if you care about doing good, you should allocate to where impact is highest. That might mean impact investing in underfunded areas, or conventional investing plus giving to high-impact charities. The marginal thinking applies: where does the next dollar of investment or donation do the most good? Compare; don't assume.
As a decision-maker
When you give time or money, compare options. Use cost-effectiveness data (e.g. GiveWell, WHO-CHOICE) where it exists. Be cause-neutral in principle: if evidence suggested a different cause had higher impact, would you switch? The discipline is not to give by habit or salience alone. Update as evidence changes. Accept uncertainty — rough comparison is still better than no comparison.
Common misapplication: Assuming that only measurable impact counts. EA emphasises evidence and comparison, but many important outcomes are hard to measure. The move is to use measurement where we have it and to be explicit about values and uncertainty where we don't. EA is a method, not a claim that the unmeasurable doesn't matter.
Second misapplication: Letting the perfect be the enemy of the good. EA can seem demanding (earn to give, work on AI risk). The point is to do more good, not to meet a purity standard. Marginal improvement — better allocation of the next unit — is the target. Don't use EA as a reason to do nothing because you can't do everything.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
Bill GatesCo-founder, Microsoft; Co-chair, Bill & Melinda Gates Foundation
The Gates Foundation's focus on global health (vaccines, malaria, diarrheal disease) is EA-adjacent: they prioritise by impact, use cost-effectiveness evidence, and are willing to shift as evidence changes. Gates has argued for "best buys" in development — interventions with the strongest evidence and highest impact per dollar. The discipline is cause-neutral, evidence-based allocation at scale.
Sam AltmanCEO, OpenAI; former Y Combinator president
Altman has supported effective altruism and long-term risk work (e.g. AI safety). His reasoning is consequentialist: if AI could pose existential risk, devoting resources to making it safer could have very high expected impact. The EA move is to take that possibility seriously and to allocate some margin to it, while acknowledging uncertainty. Cause neutrality includes causes that feel abstract or long-term.
Section 6
Visual Explanation
Effective Altruism — Allocate the next unit of resource (time, money) where it does the most good. Compare causes by impact; be cause-neutral; update with evidence.
Section 7
Connected Models
Effective altruism sits at the intersection of ethics, prioritisation, and evidence. The models below either supply the ethical frame (consequentialism, utilitarianism), correct for bias (scope neglect), or provide tools for comparison (cost-benefit, marginal analysis).
Reinforces
Consequentialism
EA is applied consequentialism: we judge actions by their consequences (impact). The right allocation is the one that produces the best outcome. Consequentialism supplies the ethical frame; EA applies it to giving and career.
Reinforces
[Utilitarianism](/mental-models/utilitarianism)
EA is often utilitarian: maximise total welfare (or lives saved, or DALYs averted). Utilitarianism specifies the good; EA asks where we can do the most of it per unit of resource.
Reinforces
[Scope Neglect](/mental-models/scope-neglect)
We underweight large, distant, or statistical harms. EA corrects by explicitly comparing scale: how many lives, how much welfare? Scope neglect is a bias EA tries to overcome by making scale salient.
Leads-to
Marginal [Cost](/mental-models/cost)/Benefit
EA thinks at the margin: where does the next dollar or hour do the most good? Marginal cost/benefit is the same frame — we compare incremental impact across uses. EA applies it to altruistic allocation.
Reinforces
Section 8
One Key Quote
"If you can save a life by donating money, and you don't do it, then you're responsible for the death. … We're all in a position to save lives. The question is whether we will."
— Peter Singer, The Life You Can Save
Singer's formulation pushes the reader to treat preventable harm as a reason to act and to compare options: if you can save more lives with the same money by giving to a more effective charity, you should. The quote captures EA's consequentialist core and the move from "giving is good" to "giving where it does the most good."
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Compare impact. When you give time or money, don't default to the most salient cause. Ask: where does the next unit do the most good? Use cost-effectiveness data (GiveWell, WHO, cause-specific research) where it exists. Rough comparison beats no comparison.
Be cause-neutral in principle. You don't have to be indifferent — but be willing to switch if evidence points elsewhere. The cause that feels best is often not the cause that does the most good. Cause neutrality is a discipline, not a claim that all causes are equal.
Think at the margin. The question is not "is this cause good?" but "where does my next dollar or hour have the highest impact?" That can point to underfunded, evidence-backed interventions and away from crowded or low-impact ones. Marginal thinking is the core of EA prioritisation.
Update. EA is evidence-based. As evidence changes — new studies, new interventions, new risks — update your allocation. Don't lock in a cause and ignore new data. Humility about uncertainty is part of the method.
Section 10
Test Yourself
Is this mental model at work here?
Scenario 1
A donor chooses to give to a charity that distributes bed nets for malaria prevention rather than to a local arts organisation, after reading cost-effectiveness estimates.
Scenario 2
Someone says they'd like to help but don't know where to give, so they don't give.
Singer's case for giving more and giving effectively. The book popularised the idea that we can do a lot of good with modest donations if we choose high-impact interventions. Foundational for EA.
GiveWell researches and recommends charities by cost-effectiveness. The site operationalises EA: compare impact, recommend where the next dollar does the most good. Updated annually.
MacAskill's guide to effective altruism: cause neutrality, marginal impact, career choice, and evidence. Accessible introduction to the framework and its applications.
[Trade-offs](/mental-models/trade-offs)
EA is about trade-offs: more impact here means less there. Trade-offs make the comparison explicit. EA says: choose the trade-off that maximises impact. The discipline is to compare rather than assume.
Tension
Expected [Utility](/mental-models/utility) Theory
Expected utility compares options by expected value. EA compares charitable and career options by expected impact — same structure. Uncertainty is central to both: we act on expectations, not certainties. EA applies expected utility to doing good.
