DATAREPUBLICAN: I mapped 25+ GB of Effective Altruism material. Here's what I learned.

AI-written summary of DataRepublican, X Article, September 21, 2026. Read the original: x.com/DataRepublican/status/2102225631350325733

Follow DataRepublican: X (@DataRepublican) · Substack · datarepublican.com

About the Author

Jennica Pounds, who writes as DataRepublican, is a software engineer and data analyst who worked at Amazon, eBay, and Snap before a machine-learning engineering role at the AI lending firm Upstart, which she left in 2025 to research government spending full-time. Her site, datarepublican.com, makes federal grants searchable and maps the nonprofits and NGOs that receive them, and it drew national attention during the 2025 DOGE push; it also hosts an EA explorer that collects effective altruists' own writing. She is coauthor, with Joshua Lisec, of Unelected: How the NGO-Administrative Complex Seized Power—And How to Get It Back (Passage Press, 2026), which debuted in the top 10 of the New York Times bestseller list.

Quick Context

Effective altruism (EA) is a movement, associated with philosophers such as Peter Singer and William MacAskill, that tries to use evidence and reasoning to do the most good with each dollar or hour. Longtermism is the strand of EA that gives heavy weight to people who might live in the far future, which is why many EA groups fund work on preventing AI-caused extinction. Hayek's "pretense of knowledge" comes from economist Friedrich Hayek's 1974 Nobel lecture, which warned that experts and planners can never gather enough knowledge to steer a complex society.

The Big Idea

DataRepublican says she has mapped more than 25 GB of Effective Altruism (EA) material, and her takeaway is simple: EA's habit of putting a price on hypothetical future lives leads followers to treat any chance of AI-caused extinction as outweighing everything else. Her conclusion is that EA should be kept out of AI safety, policy-making, and governance.

Key Points

• She describes two features of EA as she sees them: future lives count alongside present ones ("longtermism"), and moral math is not limited to humans.
• Because extinction acts as "negative infinity" in that calculation, she argues, even a tiny probability of AI catastrophe dominates every other consideration.
• She says this makes common objections, such as China winning the AI race or concerns about centralized power, ineffective with EA adherents.
• Her core rebuttal is that meaningful future-value calculation is impossible. She cites Hayek's "pretense of knowledge" and past predicted catastrophes that did not arrive.
• She closes on Romans 3:8 ("why not do evil that good may come?") and argues Christians should "entrust the future to God."

Important Details and Numbers

• Trigger event, per the article: a September 8 resignation post by AI researcher Jacob Coxon that she says drew about 170 million views.
• She cites Rethink Priorities (reported 2024 revenue over $20 million) and its "moral weight" research, which by her account treats 3 chickens or 32 shrimp as equal to one human. She notes the group says these are not absolute worth ratios.
• She cites 400 billion to 1 trillion shrimp farmed per year and one EA argument about 10^14 potential future lives.
• Her characterizations of EA members' traits are her own generalizations.

Why It's Here

The sharpest move here is her claim that the math itself is the problem: once you put a price on trillions of hypothetical future lives, almost any cost today can be justified. That's a readable case for humility about anyone who claims to compute the future, the same instinct behind the Ledger's wariness of concentrated power. It's here because the people she describes are shaping AI policy right now, and her map shows how they think.

"It is not possible to calculate future value in any meaningful respect." (DataRepublican)