The Diff Are We Already Building a Piecemeal AI Data Royalty Model?
21 snips
Mar 16, 2026 They examine how language models absorb online writing and what that means for creators' rights. They explore historical examples where industries adapted to free distribution and how platforms are shifting toward creator rewards. They discuss AI adoption in medicine, data as a strategic asset, and how firms may use small payments and cooperation to avoid legal fights.
AI Snips
Chapters
Transcript
Episode notes
LLMs Capture The Zeitgeist Through Published Text
- Large language models encode the zeitgeist by training on the bulk of online text, effectively immortalizing widely published ideas.
- Byrne Hobart argues that creators' sentences become part of model weights, improving AI's ability to mimic their writing over time.
Zero Marginal Cost Doesn't Always Destroy Incumbents
- Market responses to zero-marginal-cost challengers often preserve original creators' economics by shifting product positioning or licensing.
- Examples include music streaming licensing, cable retransmission fees, and ATMs expanding rather than killing branch banking.
Grammarly Shuttered Expert Review After Creator Pushback
- Grammarly launched and then shut down Expert Review after creators and fans pushed back on using writers' styles without consent.
- Byrne Hobart notes creators' fan loyalty made Grammarly's multi-billion-dollar brand less persuasive than individual writers.