AI Policy Wiki
Dashboard

How a Fervent Belief Split Silicon Valley—and Fueled the Blowup at OpenAI — Robert McMillan, Deepa Seetharaman (WSJ, November 22 2023)

high confidence · updated 2026-06-06

WSJ feature published days after Sam Altman's reinstatement as OpenAI CEO. Documents the role of effective altruism (EA) in the November 2023 OpenAI board coup: three of the four votes to oust Altman came from EA-affiliated board members (Ilya Sutskever, Tasha McCauley, Helen Toner). Frames the episode as the high-water mark of EA's institutional influence — and the moment its limits became visible when ~700 of OpenAI's ~770 employees threatened to leave for Microsoft. Pairs with the May 2026 Musk v. Altman trial (which surfaces 2015-2018 emails confirming Altman's 'Manhattan Project for AI' Google-fear framing) to give the wiki a continuous historical record from OpenAI founding through commercial pivot.

"How a Fervent Belief Split Silicon Valley—and Fueled the Blowup at OpenAI" is a Wall Street Journal news feature by Robert McMillan and Deepa Seetharaman, published on November 22, 2023, days after Sam Altman's reinstatement as OpenAI CEO. It documents the role of effective altruism (EA) in the November 2023 OpenAI board crisis, reporting that three of the four board votes to remove Altman came from EA-affiliated members, and that a staff revolt and the threat of mass defection to Microsoft reversed the decision within days.

Authors: Robert McMillan, Deepa Seetharaman (WSJ) Published: 2023-11-22 in The Wall Street Journal Format: WSJ news feature, published days after Altman's reinstatement

Summary of argument

The piece describes effective altruism as a movement, supercharged, in its framing, "by hundreds of millions of dollars in tech-titan donations," whose adherents "believe a headlong rush into artificial intelligence could destroy mankind" and "favor safety over speed for AI development." It presents the November 2023 OpenAI board action as the point at which EA's influence over a frontier AI developer was at its highest, and the staff and investor response as the point at which the limits of that influence became visible. When the safety-first preferences associated with EA came up against opposition from employees and corporate investors at OpenAI, the board's action was reversed.

The article situates the episode against the broader EA funding network that had shaped both OpenAI's pre-crisis board composition and the wider AI safety research ecosystem, and against the conviction of EA's most prominent backer, Sam Bankman-Fried, in the same month.

Key claims documented

The board action and its reversal:

  • The four votes to fire Altman: Ilya Sutskever, Tasha McCauley (Effective Ventures board), Helen Toner (CSET, EA-philanthropy-backed), and Adam D'Angelo. Three of the four had explicit EA affiliations.
  • The stated pretext: Altman "failed to be consistently candid in his communications." The underlying dispute, per WSJ's sources, was that Altman clashed with Sutskever over AI-safety issues mirroring EA concerns.
  • The staff revolt: more than 700 of OpenAI's roughly 770 employees signed a letter calling for Altman's return and threatened to defect to Microsoft. Sutskever publicly recanted his vote within days.
  • Replacement: Altman returned; Sutskever, McCauley, and Toner were replaced on the board. Emmett Shear, the interim CEO who favored slowing AI development, was out. Adam D'Angelo was retained.

Positions on effective altruism:

  • Vinod Khosla is cited saying OpenAI's board members' "religion of 'effective altruism' and its misapplication could have set back the world's path to the tremendous benefits of artificial intelligence."
  • Altman had called effective altruism an "incredibly flawed movement" that showed "very weird emergent behavior."
  • Elon Musk is quoted as EA-sympathetic, having called William MacAskill's writings "a close match for my philosophy."

EA donors and figures named: Dustin Moskovitz (co-founder, Facebook), Jaan Tallinn (Skype founder), Sam Bankman-Fried (before his fall), and William MacAskill (movement co-founder). The article ties EA-affiliated funding to the OpenAI board's pre-crisis composition (McCauley, Toner, and the Effective Ventures network), to Anthropic's early funding (Dario Amodei's pre-commercial-pivot reliance on "wealthy philanthropists in the effective altruism movement" — see Anthropic and Wsj Anthropic Front Runner 2026), and to the broader AI safety research ecosystem (Future of Humanity Institute, Open Philanthropy, MIRI, ARIA).

Historical context

At the time of publication, effective altruism's most prominent backer, Sam Bankman-Fried, had just been convicted of fraud (November 2023). ChatGPT was one year old, and capability worries had moved from EA-internal concerns into mainstream policy debate. Altman had spent 2023 on a "world tour" warning that AI could cause serious harm, which the article frames as partly an effort to deflect criticism and partly a response to EA-coded concerns.

The November 2023 governance crisis was the proximate cause of OpenAI's subsequent restructuring away from non-profit control, a process that continued through the October 2025 cap-removal and PBC structure documented in the May 2026 Musk v. Altman trial proceedings (see Musk v. Altman (and OpenAI / Microsoft / Brockman)). Anthropic's early EA-philanthropist funding (per Anthropic) positioned it as the EA-aligned alternative to OpenAI's commercial track, and it remained closer to the EA orbit through 2024–25. Helen Toner subsequently published research with Daniel Kokotajlo and Daniela Amodei at CSET on AI evaluation.

The piece pairs with the May 2026 Musk v. Altman trial coverage (see Musk v. Altman (and OpenAI / Microsoft / Brockman) and Ft Mundy Nonprofit Openai Died if/when ingested), which surfaces the 2015–2018 email record. Together they trace OpenAI from its founding mission through the governance crisis to the commercial pivot.

Confidence and caveats

  • The piece is a WSJ news feature, not an investigative deep-dive. Some claims, such as the specific Altman-Sutskever disagreement, rest on unnamed sources.
  • It is a 2023 snapshot. Subsequent reporting, especially the Musk v. Altman trial documents and the FT Mundy May 2026 piece, adds detail on the founding-era email record that was not available at publication.
  • That three of four votes came from EA-affiliated members is a verifiable claim. The precise weight of EA as a cause, versus EA as a marker of a broader safety-versus-commercial dispute, is contested.
  • D'Angelo's vote was the fourth and remains the least explained.

Relationships