Effective altruism changed how a generation thinks about giving. Its central idea is hard to argue with: if you are going to help, help as much as you can, and use evidence to find out where your help goes furthest.
Humanity's Checklist starts from the same impulse. Where it parts ways is on what to measure, and therefore on where the money goes.
What effective altruism gets right
Effective altruism, usually shortened to EA, grew out of a few ideas that are now mainstream. Peter Singer's 1972 essay "Famine, Affluence, and Morality" argued that distance does not reduce our obligation to help. GiveWell, founded in 2007, began comparing charities by how much good they do per dollar. Giving What We Can, founded in 2009, asked people to pledge a tenth of their income to the most effective causes.
The movement brought real discipline to a sector that often runs on sentiment. It popularized a simple test for choosing causes: how big is the problem, how neglected is it, and how tractable is it? It directed large sums toward some of the most cost-effective health programs in the world, such as distributing insecticide-treated bed nets against malaria. It made it normal to ask what a donation actually achieves.
All of that is worth keeping.
Where the checklist differs
The difference starts with the unit of analysis. EA's signature tools evaluate interventions: a specific program, delivered to specific people, with an outcome that can be counted. Humanity's Checklist evaluates sources: the shared systems that keep producing many problems at once, such as the story a culture tells itself, how people coordinate, and what earns status.
That leads to a different metric. Cost per outcome is precise and comparable, but it favors work whose results are measurable soon. Systemic change is slow, indirect and hard to attribute, so it scores poorly on the very metric used to rank causes. There is a name for this trap: the streetlight effect, looking for your keys under the lamp because that is where the light is.
It is fair to say EA is not one thing. Parts of the movement fund policy work, institutional reform, AI safety and research on long-term risks, so the old charge that it ignores systems is partly out of date. But its best-known tools still reward what can be counted, and most people who meet EA meet it through charity rankings.
Side by side
| Effective altruism | Humanity's Checklist | |
|---|---|---|
| Core question | How do I do the most good with this dollar? | Where do the problems come from, and what flips the source? |
| Unit of analysis | Interventions and organizations | Sources: shared systems beneath many problems |
| Main metric | Cost per outcome, such as cost per life saved | Leverage: how many problems one change reaches |
| Evidence style | Randomized trials and cost-effectiveness models | Systems reasoning, leading indicators, open revision |
| Typical funding | Proven health programs, research, risk reduction | Six beachhead projects aimed at culture and coordination |
| Time horizon | Measurable near term, plus very long-term risk work | Compounding change over years, through feedback loops |
| Biggest risk | Missing what cannot be measured | Betting on change that is hard to prove |
The systemic critique
The sharpest criticism of EA has always come from people who care about systems. In a 2015 review of William MacAskill's book Doing Good Better for the London Review of Books, the philosopher Amia Srinivasan argued that the movement works inside existing systems rather than asking how to change them. Others have pointed out that a donor can save lives with bed nets every year, forever, without touching the reasons malaria persists in some places and not others.
Humanity's Checklist takes that criticism seriously but tries to answer it constructively. Instead of saying "change the system" in general, it names six specific systems, explains why each one sits upstream of many problems, and proposes one buildable project for each.
Where they agree
The two approaches agree on more than they disagree. Both reject giving that only makes the giver feel good. Both want results. Both are open to evidence and willing to change their minds. And EA's own test of scale, neglectedness and tractability can be applied to the checklist: the story humanity tells itself is about as large and as neglected as a cause can be, even if its tractability is uncertain.
The checklist simply adds a fourth question to the three: does this change a source, and how many problems sit downstream of it?
Why we say the only effective altruism is systemic
If a problem keeps regenerating, the real cost per outcome has to be counted against the flow of new cases, not a single case. A dollar that slows the generator can beat a dollar that treats one case, even when the first is far harder to measure. In the long run, the most effective altruism is the kind that reduces the need for altruism.
That is the bet behind the checklist's central question: what if $5 million, spent on the sources, was all it took?
What this means if you give
- Keep giving to relief if you already do; people need help today.
- Set aside a slice for sources, and ask of any project: which source does it change?
- Ask how many problems sit downstream of that source, and whether the change compounds.
- Prefer projects that are open, so others can build on what they start.