This Machine Kills 408. Synthetic Text Extruder Hype (ft. Emily Bender, Alex Hanna)
12 snips
Jun 4, 2025 
Emily Bender and Alex Hanna, authors of 'AI The Con,' delve into the myths surrounding AI technology. They simplify how systems like ChatGPT work, comparing them to ‘synthetic text extruding machines’ built from basic origins. The duo critiques the hype, especially regarding AI’s role in education, emphasizing the need for a thoughtful approach as it can complicate academic integrity. Additionally, they discuss the ethical implications of massive datasets and advocate for a community-oriented mindset in integrating AI into learning environments.
AI Snips
Chapters
Books
Transcript
Episode notes
LLMs as Advanced Language Models
- Large language models (LLMs) are advanced versions of old language models like T9 word texting systems.
- They predict the next word based on prior words, creating synthetic text rather than understanding or cognition.
AI Terms Feed Misconceptions
- AI terminology like "neural networks" misleads people to believe systems mimic human brains.
- This anthropomorphizing causes false assumptions about AI consciousness and intelligence.
LLM as Text Extruders
- "Synthetic text extruding machine" is a more precise term for LLMs than "stochastic parrots".
- It highlights AI's role in pattern-based text generation, emphasizing synthetic output rather than true understanding.
