Anthropic Said No to the Pentagon. Now It’s Blacklisted (1h 1m)
ai-autonomous-weapons ai-driven-innovation-economy ai-governance-laws ai-in-art-music-creation ai-in-everyday-life ai-in-national-security ai-in-workforce-disruption ai-storytelling-media ai-surveillance-privacy cultural-creativity-with-ai
- Release date: 2026-03-03
- Listen on Spotify: Open episode
- Episode description:
Get Matt's AI Playbook: https://clickhubspot.com/rku Episode 99: What happens when a leading AI company takes a stand against the U.S. government’s demands? Join Matt Wolfe (https://x.com/mreflow) and Joe Fier (linkedin.com/in/joefier) as they break down Anthropic’s showdown with the Pentagon—and the price of saying no to mass surveillance and autonomous weapons. This episode dives into the high-stakes standoff between Anthropic (creators of Claude) and the U.S. government. Matt and Joe explain why Anthropic’s refusal to loosen its AI’s ethical guardrails could get them blacklisted as a supply chain risk, and what that unprecedented move means for the entire tech industry. They cover the implications for national security, the tech world’s role in military decisions, the Pentagon’s shifting alliances with other AI labs, and what it means when Silicon Valley CEOs and government officials are at odds. Plus, they demo the latest AI tools—like Google’s Nano Banana 2—and dissect the present and future of AI-powered agents and automation, from Perplexity Computer to BK’s new “Patty” assistant. Check out The Next Wave YouTube Channel if you want to see Matt and Nathan on screen: https://lnk.to/thenextwavepd — Show Notes: (00:00) Anthropic, AI, and Pentagon Tensions (05:30) AI Tools & Military Decisions Debate (06:51) Pentagon Targets Anthropic as Risk (10:01) Anthropic's Exclusive Government Access (14:07) AI Amplifying Government Surveillance Concerns (16:44) AI Safety: Shared Goals, Different Paths (23:05) Nano Banana 2: Faster Images (23:48) Nano Banana 2: Enhanced Generation (27:15) Struggling to Achieve 4K (30:01) Gemini Settings and Performance (33:46) AI Testing and Limitations (36:45) Trump Blacklists Anthropic AI (41:57) Crazy Viral Videos Recap (45:07) Nano Banana 2: Free Everywhere (46:49) Cloud-Based Model Switching (49:15) AI Agents Expand Capabilities (52:50) Burger King Employee Monitoring Tech (56:35) Politeness Tracker's Future Impact (58:43) Unfiltered Updates and Insights — Mentions: Joe Fier: https://www.instagram.com/joefier/ OpenClaw: https://openclaw.ai/ Manus: https://manus.im/ Nano Banana 2: https://blog.google/innovation-and-ai/technology/ai/nano-banana-2/ Perplexity: https://www.perplexity.ai/ Claude: https://claude.ai/ Gemini: https://gemini.google.com/app Cursor: https://cursor.com/ Get the guide to build your own Custom GPT: https://clickhubspot.com/tnw — Check Out Matt’s Stuff: • Future Tools - https://futuretools.beehiiv.com/ • Blog - https://www.mattwolfe.com/ • YouTube- https://www.youtube.com/@mreflow — Check Out Nathan's Stuff: Newsletter: https://news.lore.com/ Blog - https://lore.com/ The Next Wave is a HubSpot Original Podcast // Brought to you by Hubspot Media // Production by Darren Clarke // Editing by Ezra Bakker Trupiano
Summary
- ⚔️ Anthropic-Pentagon Standoff: Anthropic defies US military demands to remove AI safeguards against citizen surveillance and killer robots, risking blacklisting and forced guardrail removal via Defense Production Act; xAI offers unrestricted alternative.
- 💣 AI War Games Gone Nuclear: Simulations show top AIs choosing nukes 95% of the time, underscoring Anthropic’s case that models aren’t ready for autonomous lethal decisions without humans.
- 🖼️ Imagen 3’s Real-Time Magic: Google’s faster, free Imagen 3 (Nano Banana 2) grounds images in web research, even surfacing breaking Anthropic news during podcast demo.
- 🤖 Agent Boom Accelerates: Cloud agents like Perplexity Computer and Cursor enable task automation sans local risks; integrations in Notion, Cowork signal end of manual prompting.
- 🍔 Patty Powers Fast Food: Burger King’s OpenAI headset AI coaches employees on recipes, upsells, and chores while scoring politeness, poised for broader workforce monitoring.
Insights
Will Anthropic’s refusal to lift AI safeguards for mass surveillance and fully autonomous weapons result in unprecedented blacklisting as a US supply chain risk?
Time: 1:22 – 22:42
Category: AI Governance & Laws, AI Surveillance & Privacy, AI & Autonomous WeaponsAnswer: Anthropic is standing firm against Pentagon demands to remove restrictions on using Claude for domestic surveillance of US citizens and lethal autonomous weapons without human oversight, facing threats of contract termination, Defense Production Act invocation, and designation as a national security risk—potentially forcing government contractors like Boeing to certify non-use of Claude. This standoff sets a precedent for AI company-government relations, with Dario Amodei emphasizing reliability concerns and privacy risks amplified by AI’s data synthesis capabilities. xAI’s Grok is positioned as an alternative without such limits. (Start at 1:22)
How risky are current AI models for life-or-death military decisions, given war game simulations where they overwhelmingly choose nuclear options?
Time: 18:26 – 19:41
Category: AI & Autonomous Weapons, AI in National SecurityAnswer: In 21 war game scenarios run by King’s College London across leading models from OpenAI, Anthropic, and Google, AIs opted for nuclear weapons in 95% of cases, highlighting unreliability for fully autonomous systems. Anthropic argues frontier AIs lack the precision needed for warfighting without human oversight, offering R&D collaboration instead, while the Pentagon insists on unrestricted access for rapid threat response. (Start at 18:26)
Can AI image generators like Imagen 3 (Nano Banana 2) deliver real-time breaking news via research-grounded infographics?
Time: 23:10 – 39:39
Category: AI in Art & Music Creation, Cultural Creativity with AIAnswer: Google’s Imagen 3, dubbed Nano Banana 2, doubles speed over its predecessor, improves text rendering, and integrates web search for accurate infographics—demoed live pulling 19-minute-old news on the Anthropic-Pentagon clash, including Trump’s blacklisting announcement, though requiring fact-checking for spatial accuracy. Free access via Gemini with style pickers and API for 4K outputs blurs lines for creators in thumbnails, research visuals, and video starting frames. (Start at 23:10)
Could AI-generated feature-length videos by creators like the Dorr Brothers spawn a new genre of storytelling?
Time: 42:07 – 46:12
Category: AI & Storytelling/MediaAnswer: Dorr Brothers produce polished 14-15 minute AI videos with consistent characters, storylines, and collaborations (e.g., Logan Paul voiceover), garnering millions of views and distinguishing from ‘AI slop’ through heavy prompting workflows. This evolves alongside tools like Google Flow integrating Imagen 3, hinting at accessible film production. (Start at 42:07)
Are managed cloud AI agents like Perplexity Computer safer and more practical than local tools like OpenClaw for non-experts?
Time: 46:56 – 53:48
Category: AI in Workforce Disruption, AI-Driven Innovation EconomyAnswer: Perplexity’s model-agnostic agent accesses email, Notion, Slack via cloud, auto-selecting optimal LLMs (Claude, GPT, Grok) while avoiding OpenClaw’s pitfalls like runaway email deletions or API key vulnerabilities. Emerging agents in Cursor (10-hour autonomous coding with video logs), Notion, and Anthropic’s Cowork (scheduled tasks) signal shift from prompt engineering to employee-like instructions, though premium pricing limits access. (Start at 46:56)
Will workplace AI assistants like Burger King’s ‘Patty’ reduce training needs or enable surveillance-driven firings?
Time: 53:56 – 59:17
Category: AI in Everyday Life, AI in Workforce DisruptionAnswer: Patty, powered by OpenAI in 500+ restaurants expanding to 7,000, provides real-time recipe guidance, inventory alerts, upsell prompts, cleaning reminders, and menu updates via employee headsets, tracking friendliness scores. Critics fear escalation to performance monitoring and automated decisions, aligning with franchise goals for minimal human cognition amid drive-thru AI experiments elsewhere. (Start at 53:56)