Marc Andreessen is the co-founder of Andreessen Horowitz. We cover what DeepSeek's R1 model means for the tech arms race between the US and China, the trade-off between competition and cooperation in geopolitics, and the factors behind the sweeping backlash against traditional media and political institutions.
Principles & Lessons:
1) The inevitability and democratizing power of open-source AI. Marc highlights how DeepSeek’s open release “guarantees that none of the AI risk people or regulators... can lock down and control AI,” stressing that open source “makes both LLM reasoning... and also R1, which is actual reasoning, reasoning for things like math and code and science, universally available.” He underscores that while this move may threaten proprietary AI labs, it drives innovation and dramatically lowers costs: “If the cost of your key input drops by 30x, all of a sudden you can... offer services cheaper or free.”
2) The new Cold War lens and how China’s openness subverts expectations. Marc consistently describes a “Cold War 2.0” dynamic between the US and China, but notes that “the funny, ironic thing, is that that has flipped and sitting here today, China took the lead last week in having the best open system.” His view is that this move by DeepSeek ironically uses Western strengths against it. He argues that while China “runs a closed political and economic system,” it surprised the world by “delivering on the promise of actually OpenAI, actual open source.”
3) The tension between national security interests and technological globalization. Marc emphasizes that “there are rational arguments that need to be discussed” about export controls and bans on Chinese software or hardware, while also warning about unintended consequences: “If China doesn’t care about US copyright... and the American companies are prohibited ultimately by law from being able to train in the books, then the US may just lose on AI.” He also raises the risk that restricting China’s access to chips “accelerates the timeline of China doing a military invasion of Taiwan,” because “they already want it, but in part because it has these magic fabs.”
4) The concept of “counter-elites” reshaping power structures. Citing his reading of thinkers like Burnham, Marc frames US politics and global power shifts as struggles between incumbents and emerging counter-elites: “A new counter elite... will compete to take over from the current elite.” He applies this to media and politics, describing the mainstream press and political establishment as threatened by “the next rabble-rouser... that just starts hurling stones at whoever’s in power.” He notes how “there’s a revolt within the elites” in both major parties.
5) Venture capital as a personal, long-term partnership rather than mere analytics. While acknowledging AI’s potential for “analytical horsepower” in investing, Marc stresses that, especially at the early stage, “so much of what we do... is really very deep evaluation of individual people.” He highlights how their firm always returns to the idea that “most of the great companies get built over 10 years, 20 years, 30 years,” and the key is “face to face with the founders,” adding that “if I meet the algorithm that can do that better than I can, I will instantly retire.”
6) The complicated role of global supply chains and the limits of easy reshoring. Marc repeatedly explains that a “modern self-driving electric car is much more like a rolling phone on wheels,” with parts from “40 different countries.” He shows that unwinding these deeply interwoven supply chains is not straightforward: “We’re talking about unwinding a supply chain that has 40 countries involved... that’s one reason an iPhone costs $1,000 and not a billion dollars.” He sees this tension of interdependence and national interest creating “a constant thing, with no easy answers.”
7) Automation and warfare as a rapidly shifting landscape. Marc points to the Ukraine conflict as a “living laboratory” for drone warfare, stating that “a squad of well-trained human soldiers with drones... should be able to hold off thousands or tens of thousands of regular troops,” and calling this drone technology “the biggest innovation in defense since the stirrup.” He anticipates future wars “won by the people with the most money and the most technology,” where machines, not soldiers, “are fighting with each other.”
8) Biotech as the next dramatic frontier with profound societal implications. Marc highlights breakthroughs such as gene editing, embryo selection, fertility technologies, and life extension as areas where “there’s just a tremendous amount that’s becoming possible.” He references controversial experiments like the “Chinese guy who did a homebrew CRISPR thing” to produce HIV-immune babies, which generated a global furor, yet also points to George Church’s counterargument “that people need to do this.” For Marc, biotech’s “cluster of life extension, reproductive technologies, and genetic optimization” may redefine what is socially and ethically acceptable in the coming decades.
Transcript