Consider This from NPR They warned about AI before it was cool. They're still worried
24 snips
Sep 25, 2025 Martin Costi, an NPR reporter, dives into the creeping worries surrounding advanced AI and its potential dangers. He highlights how sci-fi fears shape real-world concerns, showcasing the excitement and anxiety in San Francisco's AI scene. Developers envision utopian benefits but also acknowledge significant risks of catastrophe. Some researchers advocate for a pause in AI development, urging stronger safety measures despite critiques that doom scenarios might be exaggerated. The discussion raises questions about policy, economic incentives, and the future of humanity in an AI-driven world.
AI Snips
Chapters
Books
Transcript
Episode notes
Dating App Bot Shut Down
- Jonathan Liu described creating Cupidly, an AI agent that swiped on Hinge for users and then shut it down after bans.
- His story illustrates startups pushing practical AI into messy real-world constraints.
Alignment Is The Central Puzzle
- Building superhuman AI raises an alignment problem: how to keep it aligned with human values as it surpasses us.
- Nate Soares and others worry self-improving AI could outpace human understanding and control.
From Research To Alarm
- MIRI shifted from technical alignment work to urgent public warnings because recent ML progress feels too fast.
- They argue accelerating capabilities increase the chance of a dangerous feedback loop of self-improvement.
