Jan 31, 2025 · 1h 3m · more-or-less
#84: China's DeepSeek Shakes the Tech World · More or Less Podcast
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the More or Less podcast, hosts Dave Morin, Sam Lessin, and Jessica Lessin analyze Chinese AI lab DeepSeek's trillion-dollar disruption of global markets, examining the erosion of proprietary software moats, the rise of narrative-driven financial memes, and the evolving dynamics of modern information consumption.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 100% of the talking time here. How this is scored →
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sam forcefully rejects the conventional tech premise that proprietary algorithms provide durable protection, calling software moats a fool's errand that history repeatedly disproves.
Hardest push from the hosts ▶ 52:16 Jessica Challenges Sam on Meta's Capex LogicJessica refuses to buy into Sam's thesis that Meta cannot lose, pressing him on how spending tens of billions of dollars on unneeded compute infrastructure can be spun as a win-win.
Biggest teaching moment ▶ 48:30 Sam Analyzes the Collapse of the Hyperscaler MoatSam breaks down how hyperscalers relied strictly on enormous piles of capital as their sole defensive moat, which DeepSeek effectively undermined by democratizing low-cost training.
The host holds their own ▶ 6:28 Jessica Articulates Media Subscription EconomicsDrawing on her industry knowledge as head of The Information, Jessica articulates why digital news subscriptions succeed by offering trusted access when diffs matter rather than unbundled micropayments.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome and the Rapid-Fire News Metabolism | 6 | 5 | 2 | 2 | Jessica leads the discussion on the rapid news cycle while Sam unpacks his theory of news consumption shifting from loading local state to pulling live diffs. Jessica demonstrates sharp domain expertise by connecting this dynamic to subscription economics and why micropayments fail. | |
| Quitting Social Media and the Parasocial Epidemic | 5 | 6 | 3 | 5 | Dave explains his personal experiment quitting social media and his theory of parasocial relationships crowding out Dunbar's 150 real relationships. Jessica pushes back firmly when Sam mischaracterizes Dave as a non-news consumer rather than someone specifically avoiding social feeds. | |
| DeepSeek's Disruption and the Rise of Meme Warfare | 5 | 6 | 4 | 2 | The panel dives into DeepSeek's sudden market impact, with Sam mocking incumbent AI firms for crying foul over model distillation after scraping the entire web themselves. Dave highlights how American markets erased a trillion dollars in value based on unverified claims from a Chinese hedge fund. | |
| The Spectrum of Memes: From Messiah to AGI | 6 | 6 | 5 | 6 | Sam lays out a broad theoretical spectrum where value is driven entirely by coordination memes, from the Messiah to AGI. Jessica repeatedly challenges his definitions, questioning whether his framework reductively treats all social consensus as mere memes. | |
| Vine's Crypto Revival and Asymmetric Retail Bets | 5 | 7 | 4 | 4 | Dave highlights the $300M Vine token launch on Solana to illustrate the resurgence of meme coins. Sam explains the socioeconomic rationale behind retail participation, arguing that building wealth through traditional wage income is mathematically impossible, making asymmetric bets on memes rational. | |
| Crypto Classification, Coinbase's Dilemma, and Digital Communities | 5 | 6 | 6 | 4 | Dave references David Sacks' proposed crypto regulatory taxonomy dividing assets into commodities, securities, and collectibles. Sam contrarily rejects categorizing meme coins as collectibles, arguing they represent high-velocity digital communities rather than stagnant physical memorabilia. | |
| DeepSeek's Impact: Small Models and Meta's Strategy | 6 | 7 | 5 | 6 | Dave and Sam break down DeepSeek's efficiency breakthroughs and the threat they pose to hyperscaler moats. When Sam argues Meta is in an unlosable position regardless of capex waste, Jessica pushes back sharply on how burning billions on unnecessary compute could possibly benefit shareholders. | |
| DeepSeek's Server Capacity and Financial Questions | 5 | 8 | 6 | 3 | Dave cites Section 5.2 of DeepSeek's paper to praise its mathematical unification of reinforcement learning. In response to proprietary software advocates arguing open source models are too easily copied, Sam forcefully dismisses the idea that algorithms or code can ever serve as a defensible moat. | |
| Pop Culture Watchlists and DeepSeek's Board Game Hallucination | 3 | 4 | 2 | 2 | The conversation winds down into pop culture recommendations and testing LLMs. Sam humorously describes how DeepSeek confidently hallucinated an imaginary sequel to Ricochet Robots called Ricochet Rockets and refused to back down when questioned. |