e/acc vs EA — effective accelerationism and effective altruism, compared

Effective altruism treats some outcomes, misaligned AI above all, as so bad that avoiding them is worth real costs in speed. Effective accelerationism treats the cost of speed lost — disease not cured, energy not harnessed, growth not compounded — as the larger and more certain loss. Both run expected-value arguments; they disagree about which tail to fear.

EA grew out of a philosophical argument and a few charities. Peter Singer's Famine, Affluence, and Morality (1972) argued that distance does not reduce the duty to help. GiveWell (2007) began ranking charities by cost-effectiveness; Toby Ord founded Giving What We Can (2009), whose members pledge a share of their income; the name effective altruism was chosen in 2011 when these groups formed the Centre for Effective Altruism. William MacAskill's Doing Good Better (2015) brought the idea to a wide audience.

Its work clusters in three areas: global health and poverty, animal welfare, and catastrophic risk. The third grew with longtermism — the view that the people of the far future count as much as people now, set out in Ord's The Precipice (2020) and MacAskill's What We Owe the Future (2022). Its largest funder, Open Philanthropy, has been called Coefficient Giving since November 2025.

The founding e/acc newsletter (May 2022) is open about the borrowing: the overlap of "EA" and "Effective A__ism", it says, is intentional. It credits EA with being prescriptive and serious about scale, then rejects its risk-aversion and the surveillance it sees in longtermist safety plans. One co-author's shorthand for the result was "EA, but based".

• Moral questions can be reasoned about with numbers, and scale matters.
• The long-run future is enormous and worth optimizing for.
• Technology, AI above all, is the most consequential force of this century.
• Good intentions without measured results are not enough.

• AI risk. EA's AI wing gives catastrophic outcomes a high p(doom) and wants alignment, evaluations and compute governance before more capability. e/acc thinks those risks are overstated and that the safest path is more capability, spread widely.
• Who decides. EA leans on experts, foundations and policy to steer. e/acc trusts markets, open source and deployment, and distrusts any small group holding a veto — see decel.
• The cost of waiting. For EA, a delay that lowers extinction risk is cheap against a future of trillions. For e/acc, every year of delay has a body count of its own, paid now and with certainty.

FTX. Sam Bankman-Fried was EA's best-known donor and described his career as "earning to give". FTX filed for bankruptcy on 11 November 2022; he was convicted on all seven counts in November 2023 and sentenced to 25 years on 28 March 2024. The FTX Future Fund's team resigned, and MacAskill publicly condemned the deception. For critics, the episode showed what expected-value reasoning can be used to excuse.

The OpenAI board. On 17 November 2023 OpenAI's board removed Sam Altman; after most of the staff threatened to leave, he returned on 22 November. Concerns about AI risk associated with EA were present on the board that removed him, and his return was celebrated by e/acc accounts as a defeat for the cautious side. The board's stated reason was that he had not been consistently candid, not a dispute over speed.

This site is e/acc, and it takes EA seriously: the habit of measuring, the willingness to think about the very long run, the money that has gone to global health. The disagreement is about AI, and about whether the right response to an uncertain risk is to slow down or to build faster and in the open. We would rather argue that with data — the acceleration timeline and the cost of intelligence — than with slogans.

What is the difference between e/acc and effective altruism?

Both reason about doing the most good at the largest scale. Effective altruism (EA) concludes that some risks, advanced AI above all, are worth slowing down for; effective accelerationism (e/acc) concludes that slowing down is itself the larger risk. Same habit of thinking in expected value, opposite answer on the throttle.

Is e/acc named after effective altruism?

Yes. The founding e/acc newsletter of May 2022 says the "fuzzy overloading" of EA is intentional, and a co-author summed the movement up as "EA, but based". The name borrows EA's ambition and rejects its caution about AI.

What does effective altruism believe?

That charity and careers should be chosen by evidence of impact rather than by feeling. In practice its work clusters in three areas: global health and poverty, animal welfare, and reducing catastrophic and existential risks — the last of which made it the main funder of AI-safety research.

Why does e/acc oppose effective altruism?

Mainly over AI. EA's AI wing argues for caution, evaluation and governance at the frontier; e/acc sees that as handing a veto over progress to a small group of experts and regulators. The founding post also rejects "the desire for a panopticon" it reads into longtermist EA.

What did FTX have to do with effective altruism?

Sam Bankman-Fried presented his fortune as "earning to give", an EA idea, and funded EA causes through the FTX Future Fund. FTX filed for bankruptcy on 11 November 2022; he was convicted on seven counts of fraud and conspiracy in November 2023 and sentenced to 25 years on 28 March 2024. The collapse is the episode most often cited against EA.

Was the OpenAI board crisis an EA versus e/acc fight?

Partly, in how it was read. OpenAI's board removed Sam Altman on 17 November 2023 and he was back on 22 November. Board members who voted him out shared EA's concerns about AI risk, and his return was celebrated online as a win for acceleration. The board's stated reason was that he had not been consistently candid, not a dispute over speed.