The OpenAI-Anthropic Leadership Conflict: How Personal Disagreements Shape the AI Industry
Long-standing tensions between Sam Altman (OpenAI), Greg Brockman (OpenAI), Dario Amodei (Anthropic), and Daniela Amodei (Anthropic) go beyond personal grievances. They are shaping AI safety strategies, defense contracts, and the public positioning of the companies. Anthropic positions itself as the 'safe alternative,' criticizing OpenAI for commercialization and political ties.
In 2026, tensions peaked: Anthropic ran a Super Bowl ad hinting at OpenAI's chatbot advertising. Dario Amodei accused OpenAI of dishonesty on Slack after their contract with the Pentagon for classified work. Meanwhile, Anthropic is suing the Trump administration over a ban on cooperation with the U.S. Department of Defense.
Roots of the Rift in San Francisco
In 2016, key figures gathered at a house on Delano Avenue. Dario and Daniela Amodei shared the space with Greg Brockman and Holden Karnofsky. Brockman, a Stripe alumnus, tried to recruit Dario to OpenAI.
The key dispute occurred in early 2016: Brockman insisted on publicly disclosing AI's potential to 300 million Americans. Dario and Karnofsky proposed informing the government first, fearing panic from optimistic forecasts. Brockman later called this disagreement fundamental to the companies' philosophies.
- Early Projects: Dario joined OpenAI in 2016, working on Universe—training AI agents for games and interfaces.
- 2017 Layoffs: Elon Musk initiated layoffs (10–20% of staff), which shocked the Amodeis.
- Ethical Initiatives: Dario proposed OpenAI as a coordination center for AI companies and governments.
Brockman saw this as an opportunity to sell AI to nuclear powers (UN Security Council members). Dario viewed the idea as a security threat.
Broken Agreements and the Fight for Projects
After Musk's departure in 2018, Altman took over. Dario agreed to stay on the condition that Brockman and Ilya Sutskever would not dominate. Altman promised but later gave them the power to fire him.
The conflict flared around GPT. Researcher Alec Radford laid the groundwork for the Generative Pre-Trained Transformer. Brockman wanted to participate, but Dario and Daniela blocked his access.
Daniela, who led the project with Radford, threatened to quit to keep Brockman out. In meetings, Dario listed reasons for exclusion: from work style to Radford's unwillingness to collaborate. Ultimately, Brockman was excluded from the language direction.
- GPT Releases: The Amodei team implemented GPT-2 and GPT-3.
- Recognition of Contribution: Dario felt undervalued; it irritated him that Brockman publicly discussed the charter, to which he contributed less.
- Corporate Friction: Brockman's request to review a slide in 2018 led to a dispute about roles.
Public Manifestations and Global Consequences
Disagreements spilled into public attacks. Dario compared Altman's lawsuit with Musk to Hitler and Stalin, called Brockman's donation to a pro-Trump fund evil, and likened OpenAI to tobacco companies.
At a summit in New Delhi (February 2026), Altman and Amodei avoided contact, only touching elbows under the gaze of Prime Minister Modi.
The companies are racing toward IPOs with valuations exceeding $300 billion each. The conflict impacts:
- AI Safety: Anthropic emphasizes a 'principled approach,' with departures from OpenAI in 2020 due to risks.
- Defense Contracts: OpenAI takes on classified Pentagon projects; Anthropic boycotts them.
- Branding: Anthropic—the 'healthy alternative' with advertising jabs.
- IPO and Market: Personal motives hinder industry consolidation.
Key Takeaways
- Personal grievances from 2016–2018 (layoffs, access to GPT) laid the foundation for the split.
- Anthropic uses criticism of OpenAI to differentiate on safety and ethics.
- The conflict involves the Pentagon, U.S. politics, and global summits.
- Both companies lead in AI; their trajectories determine technology accessibility.
- Disagreements evolved from internal disputes to public attacks and legal battles.
— Editorial Team
No comments yet.