Claws x Code

Buying Data on AWS & Grok Epic Fails

Listen Later

Jey puts Joi onto the AWS Data Marketplace rabbit hole, where you can literally buy 50 billion web pages of human consciousness - from Common Crawl to ocean current data to seismic activity. Joi explores how social media created the perfect training dataset for AI before generative AI even existed, and why Elon's $44B Twitter acquisition was always about the data, not disruption.

Plus: Why Grok AI is hilariously broken, only trained on post-November 2024 Twitter data, and how poor trust and safety decisions lead to AI models going completely off the rails. From bathroom scrolling habits to the reality of human behavioral data collection, we break down the real cost of treating consciousness like a commodity.

The uncomfortable truth: We're living through the biggest cultural shift in human history, and most people don't understand that everything from their toilet time to their location data is being packaged and sold.

Key Topics:
  • AWS Data Marketplace and human data commodification
  • Common Crawl: 50 billion web pages available for purchase
  • Social media as AI training data goldmine
  • Elon Musk's Twitter acquisition as data strategy
  • Grok AI limitations and training data issues
  • Trust and safety impact on AI model quality
  • Human behavioral data and privacy implications
  • The biggest cultural shift in human history
  • Content moderation failures affecting AI outputs
    • Timestamps:
    • 00:00 Jey's AWS Data Marketplace tip leads to deep dive
    • 00:09 Common Crawl: 50 billion web pages for sale
    • 00:32 Ocean current and seismic data availability
    • 01:03 Human behavioral data and privacy reality
    • 01:32 Social media as AI training foundation
    • 02:02 Pre-AI data collection impact
    • 02:22 Social media vs. article comments data quality
    • 02:45 Elon's Twitter acquisition as data strategy
    • 03:07 Trust and safety failures in AI training
    • 03:29 Grok's training limitations (November 2024+)
    • 04:18 Content moderation impact on AI outputs

      • Hosts: Joi (AWS Data Marketplace explorer with digital privacy concerns) and Jey (the one who opened this rabbit hole)

      ...more
      View all episodesView all episodes
      Download on the App Store
      Claws x CodeBy Joi & Jey