Intelligent Machines (Audio) — IM 860: "You Gotta Get Computer - Claude Surges to No. 1"

Host: Jason Heiner (filling in for Leo Laporte)
Co-hosts: Paris Martineau, Jeff Jarvis
Special Guest: Dan Patterson (Blackbird AI)
Date: March 5, 2026
Main Theme:
This consequential episode of Intelligent Machines dissects a week jam-packed with transformative news in the AI world: the Pentagon's dramatic standoff with Anthropic, Sam Altman and OpenAI's opportunistic maneuvering, the dramatic surge of the Claude chatbot, Amazon data centers as wartime targets, and the debut of Perplexity's powerful new AI agent. The episode explores both the ethical dilemmas and technical innovations surrounding advanced AI as it permeates society and geopolitics.

Episode Highlights

1. The Evolving Landscape of Information Warfare and Narrative Risk

[02:16 – 22:32]

Guest Segment: Dan Patterson, Blackbird AI

• Dan shares Blackbird AI’s mission: protecting governments, organizations, and executives from narrative-based disinformation attacks, now turbocharged by generative AI tools.
  • “We protect organizations, executives and governments from narrative-based disinformation attacks that can cause operational, financial, of course, physical, sometimes harm.” — Dan Patterson [05:00]
  • Blackbird AI goes beyond 'social listening' to track how narratives travel and are amplified across social media, chat apps, and even the dark web.
• Perception as Attack Surface:
  • “Perception itself has become an attack surface.” — Jason Heiner [11:49]
  • Blackbird can forecast risk by clustering and visualizing conversational data, pinpointing when narratives escalate to threats (e.g., doxxing, doxx-to-harm).
• Not Just Doomsaying: Blackbird also advises organizations when a negative narrative is likely to burn out, and that a response would only amplify it [14:52].
  • “Sometimes it's telling people this is not worth rolling in the mud on...Let it play out.” — Jason [15:06]
  • Dan cautions that while they educate teams for fast decision making in the face of attacks, they avoid prescribing one-size-fits-all comms advice.
• Narrative Intelligence Evolution: Blackbird’s clientele has evolved from crisis comms to governments and NATO, necessitating more sophisticated tools for detecting deepfakes, context-driven claim checking (Compass), and assessing threat levels to both individuals and institutions.
• Robustness Amidst Noise:
  • Paris asks: How does Blackbird avoid being corrupted by the very disinformation it analyzes?
  • Dan explains that their detection is dynamic and multi-source, relying not on a whitelist but on understanding attacker behaviors, dark web activity, and tactics provided by partners like NATO. [27:06]
• AI Agents and Modeling:
  • Blackbird builds and trains its own models; strong internal literacy about the capabilities and risks of AI agents is mandatory for leadership.
  • “If you’re a leader and not thinking about AI agents, you’re already behind.” — Quoting Wasim Khaled (Blackbird CEO), relayed by Jason [29:06]

2. The Anthropic vs. Pentagon Showdown & Ethical Tech Boundaries

[38:45 – 95:49]

Key Event Recap:

• Friday Bombshell: Trump orders all federal agencies to cease use of Anthropic after tension reaches a boiling point — stemming from two "red lines" Anthropic set: no AI for autonomous weapons, and no mass domestic surveillance.
  • “On Friday, Trump directed every federal agency to cease use of all Anthropic technology...the culmination of a simmering brouhaha between Anthropic and the DoD.” — Paris [39:46]
  • Both moral (don't want to be complicit) and practical (the tech isn’t ready for such high stakes) arguments discussed.
• Wild Twist:
  • OpenAI and Sam Altman step in to say they’ll fill the Pentagon’s requirements, despite previously echoing Anthropic’s rules [43:51].
  • “Sam Altman proves himself to be a two-faced traitor of ass kisser to the government.” — Jeff Jarvis [45:25];
  • Paris: “Is this not the sort of behavior that led to Sam Altman's original ouster?” [45:36]
• Public & Employee Response:
  • Grassroots #DeleteChatGPT movement, Claude downloads surge to #1 worldwide [46:04], employee discontent at OpenAI and Google, reminiscent of the Google Project Maven revolt.
  • “Hundreds...of adamant ChatGPT fans are now switching en masse to Claude.” — Paris [46:43]
  • Significant not just for users, but as a potential turning point in the talent wars between AI leaders [54:00].
• Ethics Debate:
  • Who should draw moral lines on new technology — government, companies, or individuals?
  • “If Dario and Anthropic say, ‘we can’t enable autonomous killing or mass surveillance’, that’s not necessarily moralizing, it’s based in practical knowledge of AI’s risks.” — Jason [55:25]
  • “You are responsible, you are accountable, you have to be moral… Now Anthropic comes along and says, ‘yeah, we have a moral line’, and they’re accused of being dictators. No, they’re being responsible in exigent times.” — Jeff [60:35]
  • Lt. Gen. Shanahan: “No LLM, anywhere in its current form, should be considered for use in a fully lethal autonomous weapon system. It’s ludicrous to even suggest it.” [62:05]
• International and Economic Fallout:
  • Is Anthropic “dead” in the US government market? Or is it winning the long game by building public trust?
  • Claude’s popularity explodes; the company’s principled contracts (which warn employee equity can go to zero for moral reasons) are lauded as rare corporate backbone [87:43].

3. AI Consumer Shifts & Product Innovation: The Rise of Claude, AI Agents, and Perplexity Computer

[66:17 – 113:11]

• Claude Surges to #1:
  • Unprecedented: Claude, once niche, becomes the most downloaded app in the world overnight, toppling ChatGPT amidst the Pentagon drama (anecdotally, many new users prefer its capabilities and tone).
  • “All of a sudden it went from like, it was like 40th to 100th...to exactly the top number one within hours.” — Jason [66:09]
  • Even AI skeptics and non-techies are now jumping in: “Claude helped me reevaluate my life...I’ve got to ask my new best friend, Claudio.” — recounted by Paris [68:58]
  • But is this a fleeting boycott or a durable shift? Debate ensues.
• Poll Result:
  • Deep View poll: “Should Anthropic have acquiesced to the Pentagon’s request to remove safety restrictions?” 79% say no.
  • “I was surprised — I expected the no’s, but not by this margin. Perhaps a sign of a larger cultural moment.” — Jason [75:47]
• Amazon Data Centers as Wartime Targets
  • Drone strikes hit Amazon facilities in the UAE/Bahrain, underlining how US tech giants are icons and potential targets in global conflicts.
  • Gulf states have promoted themselves as a “third hub” for AI, but are not prepared to physically defend data assets in conflicts — complicating the global AI power balance [82:56].
  • “Tech companies now are as much a symbol of America as Coca-Cola or Nike once were.” — Jason [81:07]
• Perplexity Computer Launch: The Agentic Shift
  • Perplexity launches “Computer,” a powerful AI agent platform enabling users to build, deploy, and share AI apps quickly, without coding or API hassles.
    • “Perplexity Computer...lets the average person describe what they want, and it spits out a thing you can send to anyone. That’s really powerful.” — Jason [110:13]
    • First to allow cross-model capability, best-of-breed routing (Claude, Gemini, Grok, etc.), and auto-deployment/web-sharing for custom agents [107:00].
    • “My issue is that it’s a terrible name, because Jeff just said: did you get to play with Computer?” — Paris [104:42]
    • “If people have to use Terminal to run things, you won’t scale. You want to just say, ‘look what I made,’ with a link.” — Jeff [109:57]
  • Open-source agent platforms like OpenClaw have primed the market, but Perplexity makes it mainstream-friendly and is riding significant viral buzz.

4. Quick Hits & Additional Insights

• EU and Regulatory Pressure:
  • European regulators may codify prohibitions on AI for autonomous weapons and surveillance, applying pressure on US tech companies globally [86:03].
• Anthropic’s Unique Employee Agreements:
  • Unprecedented moral clause: “We may make choices that reduce your equity to zero, based on our values and principles.” [87:43]
• Guardrails Reality Check:
  • Are guardrails on general AI possible? Jeff: “Guardrails are a lie...it’s a general machine...so all you can do is warn users, not prevent all misuses.” [93:39]

5. Picks of the Week [113:35 – End]

• Paris Martineau: Superior Scrabble Transcription and Strategy – Breakingthegame.net
• Jeff Jarvis: “Walkman Land” digital collection; Bell Labs retrospective; podcast overtaking AM/FM in spoken word listening (per Edison Research)
• Jason Heiner: Whisper Flow — an AI-powered dictation tool that outperforms legacy voice assistants, making transcription and ideation via voice radically more useful (macOS/iOS)
  • “If Apple would just buy Whisper Flow, Siri could actually work.”

Notable Quotes & Moments

• “Perception is the attack surface.” — Wasim Khaled, via Jason [11:49]
• “AI is tremendous...but it also brings in challenges…We're confident that zscale[r] is going to help us ensure that we’re not slowed down by security challenges.” — Siva, Zwora (ad break) [96:54]
• “Hundreds of adamant ChatGPT fans are now switching en masse to Claude.” — Paris [46:43]
• “No LLM anywhere in its current form should be considered for use in a fully lethal autonomous weapon system. It's ludicrous to even suggest it.” — Lt. Gen. Jack Shanahan [62:05]
• “Tech companies now are as much a symbol of America as Coca-Cola or Nike once were.” — Jason [81:07]
• “If people have to use Terminal to run things, you won’t scale.” — Jeff [109:57]
• “The iPod thing is weird because for the longest time, we were like ‘just put this in my phone please’...and now we’re just separating it again.” — Jeff [121:36]

Useful Timestamps

• Dan Patterson on Blackbird AI: [02:16–22:32]
• Anthropic/Pentagon saga begins: [38:45]
• Public/employee revolt and #DeleteChatGPT: [46:35]
• OpenAI and Altman’s maneuver: [43:51, 53:18]
• Poll results on Anthropic vs Pentagon: [75:00]
• Amazon/war news: [80:17]
• Perplexity Computer product discussion: [98:01–113:11]
• Picks of the Week: [113:35–End]

Conclusion

IM 860 is a landmark episode, capturing how AI innovation is increasingly synonymous with political power, business ethics, and citizen activism. The team’s layered discussion, from Blackbird’s defense against weaponized narratives to the public surge behind Claude, exposes how quickly public sentiment and technology can pivot when ethical lines are drawn. Add in Perplexity’s clear-eyed focus on making agentic AI mainstream and you have an episode both timely and prescient.

Key Takeaways:

• AI is now a frontline issue in national and international conflict, both technically and culturally.
• Corporate moral choices, rare in tech history, can catalyze massive user (and employee) shifts.
• The agentic shift — from chatbots to deployable AI agents — is redefining what “using” AI even means.
• The industry and the world will be parsing the fallout from this week for months if not years to come.

Listen if:
You want to understand why AI is no longer just about tools but about power, ethics, and the shape of tomorrow’s society.