The AI Cold War, Odyssey Rips, Tyler Cowen Joins | Danny Yeung, Connor Love, Kahlil Lalji, Tarek Mansour, Tony Zhao

July 20, 2026

AI Summary

5 min read

The hosts open the episode with a striking admission: they have no idea what to make of the latest Chinese open-source model drop. The AI cold war narrative is back, and this time it feels different. Moonshot’s Kimi K3 has landed with benchmarks that put it on par with the best frontier models from early 2026, and the weights are about to go public. The timeline is in full chaos mode—some say it’s the end of American AI dominance, others say it’s just another Red Hat moment. The truth, as usual, is somewhere in the middle.

The Kimi K3 Shock and the Open Source Reckoning

The biggest story of the episode is the arrival of Kimi K3, a 2.8 trillion parameter model from Chinese lab Moonshot that appears to invalidate the thesis that open-source models were falling permanently behind closed frontier labs. The model is gated behind an app for now, but the weights are expected to drop imminently. The hosts walk through the two competing narratives. The above-board version: Moonshot is a talented team that marshaled enough compute to release a great model, and open-sourcing it is a smart business play—grab attention, attract talent, and drive demand for hosted API services, the same playbook Red Hat used to become a $30 billion company. The skeptical version: the compute was smuggled in against chip controls, the data was exfiltrated from frontier lab APIs through wrapper companies

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

Never miss an episode of TBPN

Get every new episode summarized in your inbox — free, ~5 minutes to read.

No spam. Unsubscribe anytime.

What you'll learn

1 (00:28) **Kimi K3 Drop and the AI Cold War** - The hosts recap the weekend's biggest story: Chinese open-source model Kimi K3 from Moonshot AI, which has sparked debate about a U.S.-China AI cold war.
2 (10:33) **Kimi K3: Demand, Compute Constraints & Market Reaction** - Moonshot pauses new subscriptions because demand has pushed GPUs to capacity, highlighting the world's compute constraint.
3 (14:47) **Benchmarks & Cybersecurity Implications** - Kimi K3 achieves top-tier results in cybersecurity benchmarks, refuting the "pure distillation" thesis.
4 (19:49) **Dean Ball's Analysis & The Open Source Debate** - Dean Ball goes back and forth on the timeline, noting the model is good but token-hungry, and questions why China allows open-sourcing.
5 (24:36) **The "Dumping" Argument & Geopolitical Strategy** - Growing Daniel calls China's strategy "dumping" – subsidizing to kill local competition, then controlling the market.
6 (31:49) **U.S. Government's Soft Power Approach** - Jacon reports the administration is taking a "slower and more durable" approach: procurement rules, entity list threats, and public pressure campaigns instead of a hard ban.
7 (33:42) **The Odyssey Opens to $264M** - Christopher Nolan's film is a big financial win, though the $250M pre-tax lifetime profit is only 0.2% of Comcast's market value.

+ Full timestamped outline available in the app

Guests on this episode

Tony Zhao Tyler Cowen

Show Notes

(01:35) - The AI Cold War
(33:40) - Odyssey Rips
(44:19) - 𝕏 Timeline Reactions
(55:54) - Tyler Cowen is an American economist, author, and professor at George Mason University, known for his work in cultural economics and as co-author of the blog Marginal Revolution. He discusses the inevitability of open-source AI, emphasizing that attempts to ban it are futile due to its global proliferation, particularly from China. He also highlights the challenges Europe faces in competing with Chinese industries, suggesting that imposing tariffs may not be an effective solution.
(01:31:26) - Danny Yeung, CEO and co-founder of Prenetics, discusses the rapid growth of IM8, a health supplement brand co-founded with David Beckham, achieving $100 million in annual recurring revenue within 11 months and projecting over $200 million in its second year. He attributes this success to a global launch strategy, shipping to 31 countries from day one, and highlights a $1 billion growth commitment from General Catalyst, emphasizing the importance of strong retention cohort data for predictable future revenue.
(01:37:48) - Connor Love, a former U.S. Army Captain with a deployment to Northern Iraq in 2019, has transitioned into venture capital, focusing on frontier technologies such as defense, space, manufacturing, and autonomous systems. In the conversation, he discusses his new role at Andreessen Horowitz, emphasizing the firm's commitment to supporting entrepreneurs in the American dynamism sector, and highlights the significant growth potential in defense and space markets, underscoring the need for substantial manufacturing capabilities to meet future demands.
(01:51:57) - Kahlil Lalji, CEO and co-founder of Natural, discusses the company's $30 million Series A funding led by Forerunner and its mission to build payments infrastructure for agents. He explains that Natural enables agents to store balances, make payments, and perform transfers, with plans to expand capabilities like collecting card information over the phone. Lalji emphasizes the potential for agents to handle a majority of global payment volume in the next decade, positioning Natural to become a comprehensive financial platform encompassing banking, payment processing, and networking services.
(02:00:06) - Tarek Mansour, co-founder and CEO of Kalshi, discusses the platform's significant engagement during the World Cup, noting it generated more brand impressions than major companies like Coca-Cola and Adidas, and even surpassed ChatGPT and Instagram in search volume. He attributes this success to a dynamic, adaptable strategy executed by a lean team, emphasizing rapid response to unfolding events. Additionally, Mansour introduces Kalshi's launch of GPU futures, aiming to establish a forward curve for compute resources, enabling stakeholders to hedge risks and make informed investment decisions in the evolving compute m