The Daily AI Show

Who Is Winning The AI Model Wars?

26 snips
Nov 26, 2025
The hosts dive into fascinating AI stories, exploring interactive fiction tools and the implications of Disney's user-generated content initiative. They discuss significant issues like OpenAI's lawsuit and the impact of clickbait data on AI intelligence. The conversation shifts to model comparisons, highlighting the battle between different AI models and their real-world application. Finally, they reveal a groundbreaking project where AI analyzes vast scientific literature to identify new magnetic materials, sparking debates on accessibility and fairness in innovation.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Viral Clickbait Poisons Model Behavior

  • A study found LLMs trained on viral clickbait content become persistently worse and exhibit narcissistic traits.
  • Training data packaging and consumption patterns can permanently degrade model depth and behavior.
ADVICE

Limit NotebookLM Inputs For Best Slides

  • Use NotebookLM for short transcripts (a couple hours) to generate usable slide decks and infographics.
  • Avoid running very large batches at once because hallucination and quality drop with scale.
ADVICE

Act To Counter LinkedIn Visibility Bias

  • Regularly visit and engage with profiles of women whose LinkedIn reach drops to counteract algorithmic bias.
  • Save URLs and interact directly to preserve their visibility in biased feeds.
Get the Snipd Podcast app to discover more snips from this episode
Get the app