The AI Fix OpenAI and Anthropic test each other, and everyone fails the apocalypse test
43 snips
Sep 2, 2025 Hosts dive into the quirky world of AI, where ChatGPT hilariously critiques their anatomy knowledge while a sushi hackathon tackles sustainable fishing. A military robot dog sparks political debate, contrasting with exciting breakthroughs in COVID-19 research. Amid rising cybersecurity fears, they humorously reflect on the absurdity of pregnancy robots and the challenge of AI safety. Concerning existential risks are discussed, revealing the urgent need for better governance, leaving listeners both entertained and thoughtful about the future of AI.
AI Snips
Chapters
Transcript
Episode notes
AI Roasts Hosts' Prompting Style
- Mark and Graham tested ChatGPT's critique of their prompting styles and shared humorous self-reflection about being practical vs. abstract prompt authors.
- The hosts used the example to show how AI can give blunt, actionable feedback on human prompts.
Sushi Hackathon Aims At Sustainability
- Stanford ran a sushi-themed hackathon using AI to tackle overfishing and species identification for sustainability.
- The hosts mocked the lighthearted framing but noted some entries focused on useful conservation tools.
AI Pair-Programming Gone Wrong
- Mark described his real-life debugging struggle where multiple LLMs failed to make a webpage button visibly red.
- He illustrated how AI assistants can confidently produce wrong solutions and compound developer frustration.