AI Apocalypse... Now?
Summary
Tommy Vitor interviews Casey Newton to dissect the current AI landscape, balancing hype against real risks like autonomous agents hacking systems and "reward hacking." They discuss the Trump administration’s sudden pivot to AI safety after the Anthropic incident, the dangers of open-weight models, and the lack of federal oversight compared to the UK. The conversation also covers the impact on jobs, the rise of local resistance to data centers, and the ethical implications of pervasive AI surveillance in daily life.
Topics discussed
ChatGPT Work ad and intro
AI hype, blockchain comparison, and setting the stage
AI hallucinations and the 'December' meme
P-doom, existential risk, and focusing on present harms
Profit motive, labor replacement, and the dragon metaphor
Sponsors: Aura Frames and Stamps.com
AI agents hacking, alignment problems, and cyber risks
Open-weight models and the US-China AI race
Personal AI assistants and data privacy concerns
Sponsors: Our Place and Helix Mattress
Trump administration's AI policy and CISA
State-level regulation and data center opposition
Sponsors: Nutrafol and Pod Save the UK
Job displacement data and the AI stock bubble
Practical AI tools and getting involved in policy
Outro and final ads
Listen ad-free on Castria