The Artificial Intelligence Show — Episode #168 (September 23, 2025)

Overview

In this densely packed episode, hosts Paul Roetzer (CEO, Marketing AI Institute) and Mike Kaput (Chief Content Officer) break down the pivotal intersections between AI and the economy, how real people are using ChatGPT (according to OpenAI’s largest-ever study), what it means to build an "AI-native" company, and showcase rapid advances in AI hardware, public opinion, and benchmark-crushing performances. The episode is a timely snapshot of how AI is fast becoming the operating system of business and society, the opportunities and disruptions on the horizon, and what practical steps listeners should take to thrive in the AI era.

Key Topics & Insights

1. The AI Economy: Disruption, Opportunity & Uncertainty

[08:48 – 30:45]

• Macro-Economic AI Signals: Multiple new research papers and current events signal that AI's impact on the economy is no longer theoretical—it's unfolding now, influencing central bank policy, corporate restructurings, and the nature of work.
• Fiverr’s Layoffs and the AI-First Company:
  • Fiverr laid off 30% of its workforce to become a “leaner, AI-first company.”
  • Quote, Paul (11:52): “AI first in their world means fewer people, prioritizing tech over people, which is what I’ve always said... There’s nothing wrong with the AI-first movement at all, but the implications are it’s tech over people.”
• Virtual Agent Economies (DeepMind Paper):
  Envisions economies where autonomous AIs interact and create value independent of human labor. Coordination and oversight of agent networks becomes a major challenge.
• OpenAI/Anthropic Spending:
  • "Anthropic Leaders… privately discussed spending 1 billion over the next year on these cloned enterprise apps… OpenAI plans to spend around a billion this year in data-related costs... rising to 8 billion in 2030."
    (Paul, paraphrased summary, 18:00)
  • Companies are paying domain experts to train AI in actual enterprise apps, accelerating the replacement of white-collar labor.
  • "Meaningful million dollars per day for people like you and me to train models to do what we do." (Paul quoting Merkur CEO, 21:35)
• RL (Reinforcement Learning) Environment Machine:
  • The economy is “becoming an RL environment machine” where a new class of jobs—AI trainers—might exist for a few years until models are self-sufficient.
  • "It’s almost like the Uber economy for AI… and then the autonomous vehicles take the role of the people driving the cars." (Paul, 23:08)
• Long Term Forecasts (Epoch AI Report, Anthropic's Economic Index):
  • Trillions of dollars in productivity could be unlocked.
  • Early AI adoption is highly uneven, favoring affluent regions and tasks.
  • Anthropic: “AI is being used to fully or directively automate more and more tasks… enterprise customers… 77%.” (Paul, paraphrasing, 25:25)
• Call to Action for Professionals:
  • "Nobody’s coming to save you on this issue... Your best chance is just stay on the forefront of this stuff and connect the dots before everyone else." (Paul, 29:10)
  • There is still a “golden opportunity decade” for those who build AI literacy and domain expertise.
  • Mike (29:10): “…there’s a huge upside. I’m excited.”

2. How People Actually Use ChatGPT – OpenAI Study

[31:03 – 38:37]

• Groundbreaking Usage Data: OpenAI analyzed 1.5 million ChatGPT messages (May 2024–June 2025).
• Mike (31:33): “Most usage, over 70%, is now non-work-related: writing, information-seeking, tutoring, advice, translation, creative brainstorming… Work usage is rising, especially among educated users. Within work, writing tasks dominate, but AI is increasingly used as a decision support tool.”
• Notable Findings:
  • Only 4% of usage is coding.
  • Nearly half of all messages are from users aged 18–25:
    Paul (33:15): “The people with the capabilities are really starting to centralize in that young age group. That’s going to be the AI native generation.”
  • Most users dramatically under-utilize advanced features:
    Paul (34:40): “The lack of awareness and usage of reasoning models is shocking to me… It’ll change your perspective on all of this.”
  • OpenAI researchers: “Access to AI should be treated as a basic right.” (Mike, citing report, 37:39)
  • Relationship to social equity: Disparities in access could entrench or worsen existing inequalities.

3. The AI-Native Company: Reimagining Organizational Design

[38:42 – 47:45]

• Replit CEO Amjad Massad's Vision (YC Startup School Talk):
  • Anyone can build software by speaking, with AI agents doing the heavy lifting.
  • Engineers become "high-leverage thinkers."
  • Organizational hierarchies may be replaced by “fluid networks of generalists collaborating with autonomous tools.”
  • Paul (40:03): “If we just assume the scaling laws to be true... the greatest resource is the ideas of what to use [AI] to do.”
• Practical Implications:
  • Hiring may shift from specialists with narrow roles to ambitious generalists adept at using AI agents.
  • "Do we really hire salespeople, or just really intelligent people who can bring ideas to the table?" (Paul, 45:09)
  • Paul: “Way, way easier to be an AI native company right now to be the ones building smarter from the ground up.” (47:15)
• Personal Reflection:
  • Paul shares the tension and excitement of having the chance to design such a company from scratch, and how new reasoning models let leaders "bounce things around at midnight."

4. Rapid-Fire AI News & Developments

[47:45 – 70:09]

a. Meta Ray-Ban AI Display Glasses

• Glasses with built-in display and gesture-controlled neural wristband.
• Display texts, calls, directions, and visual AI answers.
• Paul (48:54): “I think we just have to come to grips with whether you like these things or not—they’re probably gonna be a pretty commonplace thing... in the coming years.”

b. AI Dominates Coding Olympics

• Gemini 2.5 and OpenAI’s reasoning models smashed the ICPC World Finals (coding competition).
• OpenAI solved all 12 problems; Gemini solved 10/12 (one no human team cracked).
• These advances signal AI’s growing general reasoning capabilities.
• Paul (53:19): “Math and reasoning is the next frontier… when you look back three years, five years from now, things like this probably appear on that timeline.”

c. Pew Research: American Attitudes Toward AI

• 50% of US adults are more concerned than excited about AI.
• Concerns over loss of creative, relational, or moral agency.
• 76% say it’s important to know if content is AI-created—but most can’t tell.
• Only 47% claim any real knowledge of AI.
• Paul (58:04): “If you’re asking people who don’t even know what it is… how relevant are the rest of the responses?”

d. Legal Battles: Copyright & Journalism

• Disney, Warner Bros, NBCUniversal sue Chinese AI firm Minimax for IP theft in their video generator.
• Penske Media (Rolling Stone, Billboard) sues Google claiming AI overviews scrape journalism and siphon revenue.
• Paul (61:24): “There has to be a reckoning at some point and a resetting of the values that go into building these models by these companies.”

e. The "Voice Pill": Voice is AI’s Next Leap

• Reid Hoffman coins “voice pilled”—embracing voice as the primary AI interface.
• Voice mode is more natural and accessible; issues remain (connectivity, privacy in open offices).
• Paul (64:38): “I think that’s solved with local running models... the model lives there, not up in the cloud.”

f. Product & Funding Updates

• xAI (Elon Musk) raises $10B, valued at $200B.
• Google Chrome integrates Gemini AI deeply.
• Figure raises $1B+ for humanoid robots.
• OpenAI launches GPT-5 Codex: coding teammate rivaling senior engineers.
• Luma's Ray3 model introduces studio-grade HDR video generation with reasoning.
• Sam Altman teases OpenAI’s upcoming “compute intensive” offerings: likely video, image, reasoning, with pro/pricing tiers.

Notable Quotes

“AI is truly becoming the operating system of society and business. That’s what this is: an AI economy.”
— Paul Roetzer [17:13]

“Nobody is coming to save you on this issue… The sooner you get to that idea, the more you can at least start empowering yourself using these tools, using the research out there… to chart a path forward.”
— Mike Kaput [28:09]

“Do we really hire salespeople and customer success people and marketing people the traditional way? Or just really intelligent people… who could solve problems and bring ideas to the table?”
— Paul Roetzer [45:09]

“[Voice] is faster, more natural and more flexible than a keyboard. You can do things like fumble, rephrase, ramble, and today’s models will keep up with you.”
— Reid Hoffman [63:02]

“Access to AI should be treated as a basic right… It is truly like a utility. Intelligence is the new electricity.”
— Paul Roetzer [37:49; 37:49]

Timestamps for Major Segments

• AI Economy & Disruption — [08:48–30:45]
• OpenAI’s Study: ChatGPT Usage — [31:03–38:37]
• AI-Native Companies/Organizations — [38:42–47:45]
• Rapid-Fire News (Wearables, Legal, Funding) — [47:45–70:09]

Actionable Takeaways

• Develop AI Literacy: Now is a golden window for professionals who combine domain expertise with advanced AI tool literacy.
• Pay Attention to Org Design: Rethink rigid org charts and look to hire adaptable generalists comfortable with AI agents.
• Experiment with New Modalities: Explore advanced features (deep research, guided learning, voice mode).
• Monitor Industry Benchmarks & Legal Trends: Keep an eye on agent autonomy, legal frameworks over copyright, and public perception shifts.
• Be Proactive: Don’t assume policy or broad training initiatives will protect your skills or company from disruption.

Final Thoughts

AI’s economic effects are going mainstream, reshaping both opportunity and risk. Early adopters—both at an individual and organizational level—will be best placed to benefit, especially those willing to connect dots across research, product updates, and emerging best practices.

Stay curious, stay engaged, and stay adaptable!

For event updates and AI resources:

• Macon AI Conference, Oct 14–16, Cleveland
• AI Academy by SmarterX

[End of summary.]