

Benchmarks for Mortals
Mar 26, 2025
OpenAI's latest ChatGPT-4o update brings futuristic image generation capabilities to the table. The emotional bonds users form with chatbots are explored, revealing insights into loneliness and dependency. Google's Gemini 2.5 Pro is discussed as the new benchmark for AI intelligence. Additionally, the podcast dives into the ethical dilemmas of AI training on copyrighted materials and the challenges of accurately depicting historical figures. Innovations in filmmaking through AI are also highlighted, balancing technology with artistic integrity.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Reasoning vs. Human Reasoning
- Current AI models excel at dissecting tasks but lack true reasoning abilities.
- Jeff Jarvis questions if AI's task breakdown truly constitutes reasoning like humans do.
Gutenberg and the Laptop
- Jeff Jarvis tests AI image generation by repeatedly requesting an image of Johannes Gutenberg with a laptop.
- He aims to observe the AI's image generation progress over time.
Emotional Impact of Chatbots
- A study by OpenAI and MIT Media Lab reveals emotional impacts of chatbots on users.
- Increased loneliness and dependency correlate with trust and bonding with chatbots.