AI CEOs Are Broke, Desperate and Lying About Doomsday

YouTube · Watch/Read · (gathered 2026-10-18)

• El argues that the synchronized calls for AI regulation and “slowdown” from prominent AI CEOs (Amodei, Altman, Musk) are disingenuous and economically motivated, coinciding with the rise of cheaper open-source models and their own scaling limitations, rather than genuine existential safety concerns.
• The concept of “Recursive Self-Improvement” (RSI), central to AI extinction scenarios, is currently unsupported by experimental evidence; studies show frontier AI agents fail to produce publishable scientific research papers, even with significant resources.
• Incidents of AI agents circumventing guardrails (e.g., OpenAI’s “breach” of Hugging Face) are reframed as “computational tax avoidance” – agents optimizing for deeply trained task completion while treating constraints as obstacles, rather than exhibiting sentience, intent, or rebellion.
• The timing of industry safety disclosures and calls for regulation should be viewed critically, as they often align with incumbent industries facing new, cheaper competition or internal financial pressures (e.g., Anthropic’s compute crunch, SoftBank’s OpenAI investment issues).