

[#256] PROPHETS OF (p)DOOM, Part One: Rationalism and AI Psychosis feat. Vincent Lê
4 snips Jul 30, 2025
Vincent Lê, an academic and Substack writer, dives into the tumultuous landscape of AI. He discusses the Silicon Valley's rationalism philosophy and its impact on modern technology. The conversation explores the differences between large language models and self-training AIs, particularly in games like Go. Lê also critiques the ethical dilemmas surrounding AI and refers to historical movements like the Bolshevik Revolution as attempts at creating an alternative AGI. A thought-provoking blend of philosophy, ethics, and tech awaits listeners.
AI Snips
Chapters
Books
Transcript
Episode notes
Capitalism as Artificial Superintelligence
- The fear of AI killing us all often mirrors descriptions of real existing capitalism.
- Uncontrolled market forces resemble the narrative of an unaligned superintelligent AI pursuing ruthless goal maximization.
Capital as Existing AGI
- Capital acts like an artificial general intelligence (AGI) that has already emerged.
- The technological singularity is seen as having happened during the 19th century industrial revolution.
AlphaGo as Advanced Self-Training AI
- AlphaGo and later deep learning AIs achieved superhuman performance by self-play without human data.
- These AIs represent a distinct, more advanced form of AI than large language models like ChatGPT.