Team Human AI Just Ate Language: The End of the Age of Horus
31 snips
Sep 17, 2025 Explore the intriguing landscape of AI as Rushkoff dives into the risks of over-relying on language models. He critiques how these tools can reinforce biases and narcissism, challenging the idea that AI is all-knowing. Discover the fascinating effects of AI on our brains, and why it may signal the end of the dominance of written language. The conversation emphasizes the potential of treating AI as a collaborative partner rather than just a quick-fix solution. Join in on a thought-provoking contemplation of our AI-driven future!
AI Snips
Chapters
Transcript
Episode notes
Use LLMs For Small Tasks
- Treat large language models as tools for small tasks rather than replacements for big creative work.
- Use them for auxiliary, deterministic chores and keep authorship and judgment yourself.
Pleasing Output Masks Poor Quality
- LLM outputs often look good because they're optimized to please the user, not because they're accurate.
- That pleasing feedback loops reinforces people's confidence in flawed results.
AI Guest Pitches Miss The Mark
- Douglas describes receiving AI-written guest pitches with absurd comparisons and bad analogies.
- These examples show how AI can generate plausible but irrelevant connections.