TBPN — Anthropic vs DoW, Ben Thompson Joins, Ellison Says The Biggest Number

Podcast: TBPN
Hosts: John Coogan & Jordi Hays
Aired: March 2, 2026

Overview

This densely packed TBPN episode tackled one of the most dramatic weekends in recent tech and global news—from the U.S. government banning Anthropic’s AI models amid a rapidly escalating conflict with Iran, to landmark media acquisitions, legal shakeups on tariffs, and new funding across frontier tech. Special guests Ben Thompson (Stratechery), John B. Quinn (Quinn Emanuel), James Beshara (Magic Mind), and more each brought expertise on technology, law, media, and business innovation.

Main Segments and Key Insights

1. U.S. Government Bans Anthropic: AI, War, and Power

[00:56–26:00]

Topic Introduction & Context

• Over the weekend, President Trump ordered all federal agencies to immediately phase out use of Anthropic’s Claude models, citing national security as tensions with Iran flared into open conflict.
• Civil and criminal penalties are threatened if Anthropic doesn’t cooperate with the transition; agencies have six months to move off Claude.

Key Questions Raised

• How should private tech companies interface with government—especially when products like AI toe the line between ‘general-purpose tech’ and ‘national-security-grade’?
• Does the government’s assertive take reflect genuine supply chain risk, or a broader demand for control?
• How unique or dangerous is this AI moment, and is it comparable to the early nuclear era?

Discussion and Analysis

• Jordy: "If I'm the CEO of Ford…should I really refuse to sell cars to the military if I disagree with Washington’s policies? But if they ask for armored vehicles, that crosses into a new contract, not just a regular purchase." [07:12]
• John: "The negative externality should be internalized by the government who's asking for that particular contract." [07:13]
• Debate centers on whether AI companies should retain veto power over use cases when providing to government—drawing analogies to defense contractors, carmakers, and the autonomy they do (or don't) have.

Notable Quotes

• Jordi: "It gets to this question: who gets to decide how the technology’s used?" [08:43]
• John: "For Dario to say anthropic models aren’t good enough for certain DoD contexts is reasonable... but at the same time, the government must assess efficacy, which evolves quickly." [09:30]
• Adam Simon: "Palantir senior executive notified the Pentagon after an Anthropic employee inquired about Claude's role in the Venezuela raid… It's probably unfair to say Anthropic as a whole is against Claude being used there." [16:10]

Supply Chain Risk Designation

• Ben Thompson (summary): The "supply chain risk" talk is currently just a tweet—there's no formal ruling yet, but Kalshee prediction market gives ~42% odds it becomes official by April 1st. Previous such designations (Huawei, Kaspersky) were for foreign adversaries.
• Jordi: "There’d be outrage if Anthropic got supply chain risk before DJI, which reportedly has backdoors on everything from drones to robot vacuums." [19:36]

2. Ben Thompson (Stratechery): Nationalization, Nuclear Parallels, and AI Power

[57:45–91:41]

Framing the Debate

• Thompson’s article "Anthropic and Alignment" argues we’re now living through the theoreticals imagined by AI risk and effective altruism communities: advanced AI has forced hard questions about sovereignty, national security, and corporate power.
• The reality: if AI is as powerful as nukes, government will not relinquish control to private actors—mirroring the history of nuclear weapons nationalization.

Notable Insights

• "If you analogize AI to nukes…imagine a world where a private company invents nuclear weapons. What’s the government’s logical response?" [Ben, 58:10]
• "Alignment, in the short term, is about aligning with the nation-state within which you exist. All rights are subject to the agreement of those governed and—ultimately—enforced by those who wield real power." [Ben, 67:28]
• "If AI is power, and you want private executives to decide how it’s used, that’s an intolerable position for those in power and for democracy. Congress needs to pass law about digital surveillance, not have CEOs dictate national policy." [Ben, 75:59]

On Hypocrisy and Process

• Discussion over Dario Amodei (Anthropic CEO) pushing for regulation and then resisting government dictates as hypocritical.
• Ben argues the best path is for companies to sell to government and lobby politically for changes they believe in: "If you don’t, ultimate power shifts to unelected tech executives, not the democratic process." [80:22]

Nuclear Parallel

• Nationalization didn’t kill innovation or safety: "Maybe it’s controversial, but we built the bomb and have avoided nuclear war for 70 years—maybe that was the best outcome." [Jordy, 71:27]

3. Tariff & Legal Shakeups with John B. Quinn (Quinn Emanuel)

[120:02–148:57]

Key Legal News

• U.S. Supreme Court struck down tariffs imposed under the Emergency Economic Powers Act; $670B in tariffs were collected.
• President Trump used a new legal authority to immediately reimpose up-to-15% tariffs for 150 days while Congress must act to extend.
• Companies are already filing lawsuits to recover improperly collected tariffs—an unusually straightforward process, per Quinn.

AI and Law

• Legal industry is transforming: junior associate tasks are rapidly automatable with LLMs; Quinn’s firm built AI tools to surface the best evidence from huge discovery troves.
• Potential “litigation explosion” as AI enables new class-action identification and reduces case time.

4. Funding Rounds & Lightning Announcements

[149:22–202:56]

Major Funding Announcements

• WorkOS (enterprise SaaS infra): $100M Series C on a $2B valuation. Powers enterprise identity for OpenAI, Anthropic, Perplexity, Cursor and many new AI lab clients. [150:05]
• Flux AI (hardware design): $30M Series B to accelerate AI tools for electronics, aiming for “prompt-to-PCB” flows. [174:50]
• Quiver AI: $8.3M led by a16z for prompt-to-SVG code generation.
• SMAC Technologies (frontier AI defense models): $32M to bring deep reinforcement learning to national security decision-making. [200:48]
• Cal.ai (food image-to-calorie tracking): Acquired by MyFitnessPal after $30M+ ARR, all built by a teen founder.

Product News

• OpenAI device spotted in the wild; speculation about hardware form factor and strategic leaks.
• Launch of TBPN Simulator (in-studio experience).

5. Conversation Highlights, Lightning Round, and Memorable Moments

Debate: Who Decides AI Use—Democracy vs. Corporate CEO?

• Palmer Luckey (reading): "Do you believe in democracy? Should our military be regulated by our elected leaders or corporate executives? A missile company enforcing their own policies on war is a minefield… At the end of the day, the American experiment is ongoing." [30:52–33:51]

Netflix–Paramount–Warner Media Mega-Deal

• David Ellison’s Skydance takes over Paramount, then launches a successful $81B bid for Warner Bros Discovery, out-bidding Netflix.
• Netflix played “3D chess”—forcing a rival to overpay, then earned a $2.8B breakup fee and possibly gets more licensed content regardless. [41:57–43:26]
• John: "Somehow Netflix was able to force one of its rivals to overpay...and got paid $2.8B for it." [43:46]

Entertainment/Venue Trends

• Adam Simon (ex-IPG Media) predicts "Immersion Economy": entertainment, events, and sports will be deeply enhanced by tech—personalized audio feeds, AR/VR experiences, responsive venues (Sphere, Cosm), not just more screentime. [161:37–174:13]
• Demand for IRL experiences remains robust; post-pandemic “revenge spending” is sticking.

Career Reflections & Missed Opportunities

• James Beshara (Magic Mind founder) shares his heartbreak at passing on OpenAI’s seed round, then discusses iterative product building and hiring only senior staff.
• "I did the math—the miss is like I missed out on 40 Googles..." [115:01]

Notable Quotes & Memorable Moments

• Ben Thompson [59:03]: "A failure to grapple with a world of guns… If AI is nuke-level tech, the people with guns will demand a say. It's not fun, but it's reality."
• Jordy [80:58]: "That's a huge change—I just spawn in and believe in democracy and now it's people discussing an entirely different world of governance."
• John Quinn [139:31]: "Big law firms have a pyramid structure; much of the junior work is now being done by large language models. That's going to disrupt the profession."
• JJ Beshara [118:53]: "The humility that comes with building—you only get that by doing. If you're just number two at a big company, you're living in your head. It's by putting yourself out there that you really grow."

Key Timestamps

• Anthropic ban/government conflict: [00:56–16:34]
• Private company vs. democracy debate: [27:09–35:10]
• Ben Thompson deep-dive: [57:45–91:41]
• Tariff legal segment with John Quinn: [120:02–148:57]
• Startups, funding, and product news: [149:22–174:36]
• Entertainment/Immersion Economy: [161:24–174:13]
• SMAC Technologies (AI for defense): [193:01–202:56]

Tone & Takeaways

The episode maintained TBPN’s energetic, live “temple of technology” spirit and featured genuinely candid, occasionally irreverent debate. Technical nuance, hard policy questions, and inside-industry knowledge flowed freely, making it must-listen content for policy wonks, VCs, founders, and anyone tracking America’s rapidly shifting relationship with Silicon Valley.

Major takeaways:

• The U.S. government’s relationship with top AI labs is now existential—contracts, ideology, and supply chain control are being openly contested.
• The parallel to nuclear tech isn’t just metaphorical; nationalization and control are playbook moves.
• Legal and commercial frameworks are being rewritten in real time—across tariffs, IP, and even the very structures of big law.
• Venture funding and product velocity remain at fever pitch, even as macro and regulatory risks mount.
• The next frontiers of tech are immersive, physical, and—despite all odds—still very human.

For the full drama—ranging from PetSmart conspiracy debates to tales of missed billion-dollar investments—listen to the archived episode.