Amodei torches OpenAI in leaked memo

2026-03-05 · Source: The Rundown AI · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Novice, medium

Summary

Anthropic CEO Dario Amodei criticized OpenAI's Pentagon AI deal in a leaked 1,600-word internal memo, labeling it "80% safety theater" and accusing Sam Altman of "gaslighting." This follows the Pentagon's designation of Anthropic as a "supply chain risk" and OpenAI's subsequent agreement to a similar deal. Amodei's memo escalates the rivalry between the two AI companies, referencing past incidents like a handshake refusal and Super Bowl ads. Separately, OpenAI is reportedly developing an internal code repository to replace Microsoft's GitHub due to repeated outages, potentially creating further tension with its largest investor. Additionally, OpenAI, Stanford, and the University of Tartu introduced a new framework to measure AI's impact on student learning, with initial trials showing microeconomics students scored 15% higher using ChatGPT's study mode.

Key takeaway

For CTOs and VPs of Engineering evaluating AI partnerships and infrastructure, this content highlights the critical need to scrutinize "safety theater" in defense contracts and consider the long-term implications of vendor lock-in. Your teams should assess the strategic risks of relying on a single provider for core development tools, especially when that provider is also a competitor. Explore internal development of critical infrastructure to maintain control and reduce dependency, as OpenAI is doing with its GitHub alternative.

Key insights

A leaked memo reveals escalating tensions between Anthropic and OpenAI over defense contracts and strategic independence.

Principles

• AI safety claims can be perceived as performative.
• Strategic independence drives internal tool development.

Method

A four-step prompt-based method uses Claude in Excel to transform raw CSV data into a clean, visualized dashboard with summary tables and highlighted metrics.

In practice

• Use Claude in Excel for rapid CSV-to-dashboard conversion.
• Develop custom internal tools to reduce SaaS dependency.
• Implement AI-powered study modes to potentially boost learning outcomes.

Topics

• Anthropic-OpenAI Rivalry
• AI Governance
• Developer Tools
• AI in Education
• LLM Applications

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, Data Scientist, AI Product Manager

Related on AIssential

• Open Source Ecosystems — Open protocols are vulnerable to "complement capture" by private actors, shifting competitive advantage to developer experience.
• Sovereign AI? Anthropic shutdown reveals Canada’s weakness — The Anthropic shutdown reveals Canada's AI sovereignty gap, stemming from a failure to commercialize research leadership into domestic control.
• OpenAI's Momentum is Spiraling Down ▼ — OpenAI is losing market share, talent, and investor confidence to rivals like Anthropic and Google.
• Why OpenAI and Anthropic Are Becoming Consultants — Organizational transformation is critical for successful AI deployment, as individual AI capabilities often outpace enterprise readiness.
• The 3 Betrayals Of Anthropic — AI model providers are shifting from neutral utilities to strategic actors controlling access, compute, and competition.
• AI Weekly Issue #481: Musk wants Altman fired, Anthropic passes OpenAI, Meta goes closed — Enterprise demand is driving AI revenue, leading to strategic shifts in open-source commitments and increased geopolitical tensions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Rundown AI.