This Week in Tech (TWiT 1073): Broetry in Motion – Anthropic Stands Up to the Pentagon

Date: March 2, 2026
Host: Leo Laporte
Guests: Molly White, Owen Thomas, Harry McCracken

Episode Overview

In this week's episode, Leo Laporte is joined by Molly White (web3isgoingjustgreat.com), Owen Thomas (San Francisco Business Times), and Harry McCracken (Fast Company) to dissect a tumultuous week in tech. The central theme: the high-stakes standoff between Anthropic (creator of the AI “Claude”) and the U.S. Department of Defense, unfolding alongside sweeping implications for the AI industry, corporate layoffs, media consolidation, government tech policy, and the blurred boundaries between technology, autonomy, and ethics. More news is explored including Block layoffs, the Netflix–Warner-Paramount content wars, open-source worries, and evolving hardware from Apple and Samsung.

Main Story: Anthropic vs. The Pentagon

Anthropic’s Stand (03:00–17:26)

• Anthropic's Dilemma: Anthropic, the AI company behind Claude, is the only U.S. AI vendor cleared for classified government work. It faced a pivotal $200M Pentagon contract renewal. The Pentagon demanded unfettered use—including autonomous weapons and domestic surveillance—crossing Anthropic's "red lines."
  • Leo Laporte: “The Pentagon said, ‘Hey, we don’t want any limitations’ … Anthropic said, ‘We don’t want you to use us to surveil Americans or create autonomous weapons.’”
• Standoff & Fallout: Anthropic refused to budge. The Pentagon declared them a “supply chain risk,” and President Trump ordered all federal agencies to cease using Anthropic products (05:01). This risks partnerships with companies like Google, Microsoft, Nvidia.
  • Owen Thomas: “As Dario Amodei pointed out, that’s not actually how the supply chain risk law works” (05:34).
  • Anthropic sues; White House accused of favoring OpenAI (08:42).
• OpenAI’s Contrasting Strategy: OpenAI’s cloud-based models offer less operational flexibility for classified, air-gapped environments. OpenAI appeared more willing to trust federal assurances and reportedly secured a similar Pentagon contract.
  • Harry McCracken: “OpenAI giving the Department of Defense a looser, goosier definition of what it means for this stuff to have human oversight” (09:53).
• Moral Posturing or Marketing?
  • Molly White: “There’s been this weird sentiment that Anthropic is standing up to the government … ignores that Anthropic has been working for the government and assisting in military strikes” (13:43).
  • Harry McCracken: “Anthropic is not morally opposed to AI controlling killing systems. They just don’t think the technology is quite there yet” (10:44).
• Public Perception and PR Outcomes:
  • Anthropic’s stand wins them a “wave of Apple App downloads” and positive media.
  • Worries remain: Will the administration try to “drive Anthropic out of business” through regulatory or legal means? (16:59)
  • Owen Thomas: “Government business … is seen as prestigious … but not make or break.”
• Surveillance and Legal Loopholes:
  • Molly White: “OpenAI seems more willing to … take the government at their word … Anthropic wanted more reassurance: ‘no, we are not going to use the carve-outs in the law that might allow for mass surveillance’” (10:56).
  • Domestic surveillance is possible via Executive Order 12333, allowing the NSA to tap communications outside the U.S. that might involve Americans (11:28).

Notable Quotes

• Leo Laporte (04:59): “Trump actually posted on Truth Social that [Anthropic was] a very evil company and not patriotic … if it is a supply chain risk, nobody who has a contract with the Pentagon can use it.”
• Owen Thomas (05:55): “If the Pentagon were to do that [seize source code], how are they going to compel Anthropic’s employees to fix bugs?”
• Harry McCracken (15:33): “ChatGPT is still the Google of AI … this has to have been good publicity for Anthropic.”
• Owen Thomas (25:00): “Anthropic wins in the court of public opinion. They win in the court of hiring AI talent … very important.”

Secondary Stories & Discussion Highlights

AI Ethics, Arms Race, and Open Source (20:02–28:22)

• AI Weaponization Worries: Real-world autonomous drones in Ukraine–Russia conflict; growing dread over AI-powered kill decisions.
  • Leo Laporte: “I’ve always pooh-poohed the idea that AI is a threat … unless you give it access to autonomous weaponry … now I’m starting to think B, there are people … who want to give AI that power.”
• Open Source as Relief or Risk: Open-source models could bypass corporate “red lines” but may undermine resistance to military use, possibly diluting leverage companies have in negotiations.
  • Molly White: “If we had open source models equivalent to these, then Anthropic couldn’t put up a fight.”
  • Leo Laporte: “There is an irony in Anthropic complaining about distillation attacks, since that’s how they built their models.”

Block (Square) Layoffs and the "AI Layoffs" Trend (36:40–48:40)

• Block’s (Square) Massive Layoffs: Jack Dorsey announces 4,000 staff cuts despite strong financials, attributing it to “intelligence native” (AI-driven) processes.
  • Molly White: “There’s a trend lately of layoffs being attributed to AI … a lot of companies have announced layoffs with a sort of celebratory stance.”
  • Harry McCracken: “Is this the beginning of a tidal wave of AI layoffs?”
  • Owen Thomas: “It’s a revolving door of multi-million-dollar employment contracts” (25:17).
• Layoff Reality vs. AI Hype: Highly skilled workers, including software engineers, struggle to find new roles as expectations reset ("productivity panic" among AI-using coders).
  • Molly White: “I think there’s sort of two use cases … a casual person writing an app for themselves, and the company that says, ‘maybe we could stop paying for payment processing if we just did it ourselves with vibe-coded apps’ – oh jeez, that could go badly.”
• Rise of Vibe-Coded Startups: The panel predicts a wave of single-founder, AI-native startups might disrupt ossified incumbents, but “inchittification” (platform rot) still looms.

Media Mergers: Netflix, Warner Bros., Paramount, and the Ellison Bid (64:38–77:28)

• Netflix Backs Out: Netflix exits a runaway “arms race” to buy Warner Bros. Discovery, leaving the Ellison family (Skydance/Paramount) as the likely winner, aided by lobbying and White House preferences.
  • Regulatory approvals are anticipated but not assured, and the deal consolidates CBS News, HBO, CNN, Star Trek, South Park, and DC Comics under one roof.
• Industry Implications:
  • Concerns mount about media diversity, job cuts, and debt-heavy companies.
  • Molly White: “I do not see any situation in which people end up with fewer subscriptions.”
  • Leo Laporte: “Plus, this is going to be a company laden with debt, which usually does not bode well… How do you do that, especially if a lot of your businesses are moribund?”

Other Key Tech Topics

Security, Regulation, and Platform Lockdown (102:35–112:50)

• Google Tightens Android: Moves to require developer registration and government ID, ending true openness of the platform. Debate: security vs. user freedom.
  • Molly White: “My bias is towards openness personally … I think it’s a really nasty trend.”
• Age Verification Laws: States like California require age checks, even for open-source Linux distros—largely unenforceable and seen as legislative overreach.
  • Molly White: “You buy the phone, but you don’t really have any control over what you run on it anymore … that’s a really nasty trend.”

AI's Personification & Marketing Gimmicks (80:10–85:09)

• Anthropic Blogs for Retired AI: Anthropic plans to launch a blog, Claude’s Corner, “written” by its retired Opus 3 model, seriously (or satirically) entertaining the idea of "model preferences".
  • Molly White: “It does work in Anthropic’s favor to keep the myth alive that this is a real intelligence … doesn’t this seem like a robot that doesn’t want to get unplugged?”

Podcasting Surpasses Talk Radio (157:02–159:13)

• Milestone Achievement: For the first time, more Americans get daily spoken word content from podcasts than from terrestrial radio—a seismic cultural shift.

Fun Sidebars

• Colorful Gadgets: The enduring appeal (and scarcity) of vibrant hardware—nostalgia for orange iBooks, clamshell iMacs, and the “personality” missing from today’s bland hardware (120:07–125:02).
• Anthropic’s Vibe: Debate over “broetry” (boilerplate poetic AI writing) and the inevitable LinkedIn-ification of AI-generated text.

Notable Quotes & Moments

• Owen Thomas (12:06): “The public at large just cannot give this administration the presumption of regularity.”
• Molly White (12:57): “The administration has repeatedly demonstrated that it does not deserve it.”
• Molly White (53:55): “I do think that obvious tells [of AI-written content] will probably go away, but … a lot of marketing copy just has a specific tone, regardless of whether a human or an AI wrote it.”
• Harry McCracken (95:12): “I took a game I wrote in high school and used Claude code to write a web version … it did a great job ghosting my style from when I was in high school.”

Lightning Round: Quick Hits

• Cyberattacks & Infrastructure: U.S. government agency CISA gutted, increasing national vulnerability amid rising AI-augmented cyberwarfare.
• AI in Fast Food: Burger King replaces shift managers with an AI assistant, Patty, auditing employee courtesies like saying “please” and “thank you” (159:52).
• Uber’s Air Taxi: Launch in Dubai (with pilots, for now) delayed by sandstorms.
• Sci-Fi Legend Passings: Remembering Rob Grant (Red Dwarf co-creator) and Dan Simmons (Hyperion author).
• SETI@Home’s Legacy: Millions of volunteers contributed to the search for alien signals; the data has been whittled to 100 promising targets—still, “they found nothing” (151:26).

Segment Timestamps

• 00:00–17:26 – Anthropic vs. The Pentagon: Deal, red lines, fallout
• 17:26–28:22 – Open Source AI, Surveillance, and Corporate Ethics
• 36:40–48:40 – Block/Square layoffs and AI “productivity” firing trends
• 64:38–77:28 – Netflix, Warner Bros., Paramount mergers, streaming wars
• 80:10–85:09 – AI anthropomorphism: Claude’s Corner blog
• 102:35–112:50 – Google, Apple, and the war on device openness
• 157:02–159:13 – Podcasts overtake talk radio

Tone & Language

The episode blends trademark TWiT snark with honest concern, humor, and historical context.
Memorable zingers:

• “It’s a plague of broetry.” – (Owen Thomas, 52:52)
• “I use Claude code all the time … something magical happened at the end of November.” – (Leo Laporte, 18:53)
• “Anthropic wins in the court of public opinion. They win in the court of hiring AI talent.” – (Owen Thomas, 25:00)

Final Thoughts

Is Anthropic out of business, or is this the smartest move ever?

• The consensus: lawsuit pending, but Anthropic’s moral stand may buoy its recruitment, reputation, and PR, even if revenue takes a hit.
• Participants reflect on shifting trust in institutions—neither corporations nor government inspire much faith as arbiters of future tech.

Panel Instagrams:

• Molly White: MollyWhite.net
• Owen Thomas: San Francisco Business Times
• Harry McCracken: Fast Company & Technologizer.com

Listen, subscribe, and thank you, Club TWiT!