Moonshots with Peter Diamandis | EP #199

"OpenAI vs. Grok: The Race to Build the Everything App"

Date: October 8, 2025
Guests: Emad Mostaque, Dave Blundin, AWG (Alex)

Episode Overview

This episode dives into the rapid escalation of AI platform capabilities—spotlighting monumental reveals from OpenAI's recent Dev Day, the comparative advances of Grok (Elon Musk’s platform), agentic workflows, robotics, data centers, and the future “everything app.” The panel debates how these exponential technology shifts will rewire the global economy, labor, creativity, infrastructure, and human attention. The tone is energetic, bullish, often humorous, and deeply analytical—reflecting giddiness and awe at accelerating change.

1. The OpenAI Dev Day Aftershock (00:00–08:39)

• Unprecedented Growth:

  • OpenAI’s user base and developer ecosystem are exploding.
  • “We had 2 million weekly developers and 100 million weekly ChatGPT users [in 2023]... Today, 4 million developers... over 800 million use ChatGPT every week... over 6 billion tokens per minute on the API.” — Dave (03:03)

• Land Grab & Attention Wars:

  • OpenAI is in a global “land grab” (India, UK, Greece), vying for the dominant interface to human attention.
  • Competing with Meta (WhatsApp, Instagram), Google, and X (Elon Musk) for “Everything app” status.
  • “[The] battle here is human attention is finite... Everyone’s trying to be the Everything app.” — Emad (09:22)

2. Accelerating Toward Superintelligence & Economic Transformation (05:27–08:39)

• Token Volume as AI Milestone:

  • “6 billion [tokens/min] is 3 quadrillion a year. All humans together speak 50 quadrillion tokens a year. I expect [OpenAI] to hit 30 [quadrillion] next year, overtaking all human words spoken.” — Emad (05:27)
  • The imminent crossing point where AI out-communicates all humanity.
  • “We’re just at this tipping point and the tipping point is in the next six months.” — Emad (00:50)

• Planetary Economic Changes:

  • As AI agents proliferate, expect “transformative economic changes at a planetary scale.” — Alex (06:43)
  • Tension between compute limits and ballooning demand.

3. Apps, Agents, and the New Operating Systems (08:39–15:06)

• Agentic Platform Shakeup:

  • OpenAI launching an Apps SDK; ChatGPT talking directly to external apps (“Book me this trip,” “Figma, diagram this”)
  • “What is the Tencent WeChat super app that's coming together?... Everyone’s trying to be the everything app.” — Emad (09:22)

• Entrepreneurial Implications:

  • Debate: Does OpenAI’s expansion “kill off a million startups”?
  • Dave: “That’s absolutely not true... If you have a great team and you’re in AI, you will succeed every single time...you pivot so quickly.” (15:11)

• Transitory Tech Phases:

  • This “app store atop a new OS” is transitional. “At some point, every pixel is going to be generated... this is a passing phase... we’re living the Knowledge Navigator future [Apple’s 1987 concept].” — Alex (10:49, 11:53)

4. From Mind to Materialization — Autonomous Creation via AI (13:35–15:06)

• Startups at Warp Speed:

  • Now possible: ideating, branding, building, and pitching a business in minutes with ChatGPT.
  • “We’re not too many steps removed from ‘ChatGPT, start this business for me and wire the revenues.’” — Peter (13:35)

• Creation, Curation, & Attention:

  • “Consumption was expensive, it became cheap; creation was expensive, becoming cheap; now the valuable thing is curation and attention.” — Emad (14:37)

5. OpenAI Demo Highlights: Agent Builder, Voice Mode, & Sora 2 (16:08–32:53)

Build Agents Visually (16:08–19:52)

• Agent Builder: No-code Multi-step Workflows
  • Visual, node-based construction of AI-powered workflows.
  • Critique: UI will move beyond "spaghetti nodes" to voice/image-driven interfaces (Ironman's Jarvis analogy).
  • “Designing programs by drawing lines and boxes... has been around since 1980. In the age of AI, where you can talk to the AI—no, no, no.” — Dave (17:46)

Voice Mode & CodeX: Conversational Computing (20:15–27:39)

• Voice-Activated Code Execution
  • Demo: Control AV systems and physical devices by voice via Codex.
  • “AI is easy, AV is hard. Well, hopefully AI can solve that... show that video on that screen, connect me over Zoom.” — Peter (21:22)
  • The brand and product naming explosion as AI systems blend and multiply capabilities.
  • “Now it's going to be much more like these thematic branding like Codex is a grab bag. GPT is a grab bag...” — Dave (25:33)

Sora 2: Sketch-to-Video API (28:02–33:31)

• Mind to Materialization:

  • Sketch a product; Sora 2 turns it into a photorealistic video and 3D model.
    • “The holodeck is getting closer.” — Emad (30:00)
    • “This is mechanical design getting solved—an entire discipline by generative AI.” — Alex (31:57)
    • “Star Trek replicators aren’t 24th century, they're now just 2025.” — Alex (33:31)

• API Costs, Deflation, and Access:

  • Sora 2 API priced at $0.10/sec; rapid approaches to price deflation and democratization.
  • “Suddenly it’s far cheaper to outsource mechanical product design to an API call to Sora 2 than to a human. That’s an entire hyper-deflationary field being solved overnight.” — Alex (31:57)

6. Monetization, Competition & the AI Land Grab (34:11–36:39)

• OpenAI’s Evolving Business Models

  • Old ChatGPT subscriptions under pressure—token prices are plummeting due to new rivals (Grok4, DeepSea).
  • “The basic chat experience is a couple bucks a year now...you have to think about agentic workflows...because competitors are going to release your $20/mo product for free.” — Emad (34:11)
  • Global expansion and competition with open-source Chinese models.

• Sam Altman as Chief Strategist

  • “Paul Graham said, if you dropped [Sam Altman] on an island full of cannibals... he’d be running the island. He’s going after installed base and data centers. If you control the endpoints, everything else fills in.” — Dave (35:54)

7. Anthropic, Grok & The Next AI Wars (37:38–45:56)

Anthropic: Superhuman Computer Use (38:04–41:29)

• OS World benchmark: measuring full-control AI agents on computer tasks.
• Anthropic, OpenAI pushing toward “Jarvis for your computer.”
• “Generalist models are getting good enough to do most human standard tasks...this is the takeoff point.” — Emad (40:30)

Grok Updates & Video Reasoning (41:30–45:13)

• Grok’s video generation (15s clips), speed-focused, targeting gaming/entertainment.
• “The battle for human attention is the next battle for revenue... video reasoning will end up being a killer app.” — Alex (44:00)
• “OpenAI did $4.3B in revenue in the first half [of the year]; video games did $200B last year. Gaming is a massive market to go after.” — Emad (44:14)

8. Chips, Data Centers, and the AI Industrial Boom (47:30–58:31)

• OpenAI & AMD Partnership

  • OpenAI to deploy 6 gigawatts of AMD GPUs; stock surges, unheard-of corporate dealmaking.
  • “OpenAI will get 10% ownership [of AMD spin-up] for basically no price... This is a serious win-win.” — Dave (47:30)

• Supply Chain: Beyond the Chips

  • Chip production (TSMC, Intel, Samsung) sets the upper bound.
  • “Underlying materials...silicon boules, glass...all just going to get sucked into the vortex.” — Dave (49:03)

• Data Center Land Rush

  • BlackRock buying up to 78 data centers (5GW, $40B); Corning set to benefit from fiber optics.
  • “Half of US GDP growth this year is AI, which is insane.” — Peter (52:57)
  • “The data center build out can continue to the point of trillions of dollars of capex—as long as transformation and revenue generation keep up.” — Alex (55:55)

• Energy Bottlenecks:

  • Surging demand for energy, regulatory and community resistance, need for off-grid and innovative solutions (small modular reactors, space-based data centers).
  • “We’ll have gigawatt data centers in space... actually makes sense in a few years.” — Emad (58:07)

9. Robotics: End-to-End Automation, FSD, and Beyond (58:31–68:44)

• Tesla FSD 14.1 and Embodied AI

  • 10x parameters, neural net navigation—Tesla’s “feels alive” and advances toward full autonomy.
  • Long-term convergence of vehicle, robot, and digital agent stacks: “vision language action” (VLA) models as general brains.
  • “We’re seeing the emergence of VLA models... works in cars, in Optimus robots—end to end embodiment.” — Alex (59:57)

• Google Gemini’s Robotics & Asimov Safety Benchmarks

  • Vision-language models enabling robots to reason and act efficiently.
  • “Mechanical design and reasoning getting solved on edge compute... all specialist chips and models with tremendous capabilities.” — Emad (61:33)
  • “DeepMind is benchmarking the safety of Asimov’s three laws against better constitutions for robotic AI. The fact we’re now at this point is amazing.” — Alex (62:33)

• The Coming Automation Flywheel

  • Recycling, sorting, labor, and construction become autonomous.
  • “So many tasks no human would want can just be automated.” — Alex (65:51)
  • “The fact it's all neural network–based... the pace of development and dexterity is going to skyrocket.” — Dave (66:18)

• Singularity Riffs:

  • Recursive self-improving loops—robots building data centers and more capable robots.
  • “It’s a hyper-exponential. I can feel the singularity coming.” — Peter (68:09)

10. Notable Quotes & Memorable Moments

• “This is the vaudeville on Hollywood screens stage of AI for software development... on the one hand it’s great... On the other, it's a passing phase.” — Alex (18:58)
• “Everything you see and hear—pixels, noises—is up for grabs; curation and attention become the final scarce resources.” — Emad (14:37)
• “If you have a great team and you’re in AI, you’ll succeed every single time. Maybe something gets crushed by OpenAI, but you pivot quickly.” — Dave (15:43)
• “[Sam Altman] will be running the island full of cannibals… he’s going after user base and data centers. If you control the endpoints, everything in the middle will fill in.” — Dave (35:54)
• “Star Trek replicators aren’t 24th century, they're now just 2025.” — Alex (33:31)
• “The music [of investment] can continue as long as the transformative applications continue ... and the revenue that results from that.” — Alex (55:55)

11. The Race to Everything and What’s Next

• AI “everything apps” (OpenAI, Grok, Meta, X) are converging toward seamless, multimodal agentic experiences.
• Breakneck deflation in AI costs is unlocking new classes of application and market entry.
• Robotics is fast approaching generalist labor at scale.
• The hardware and industrial base is rushing to keep up with AI’s insatiable demand across the entire supply chain, from chips to fiber, to energy to gigawatt-scale data centers.

12. Timestamps for Key Segments

| Segment | Start | |-------------------------------------------------------------------|-----------| | OpenAI Dev Day Reactions / Impact | 00:00 | | Token Volumes & Tipping Point | 05:27 | | Apps SDK, Agentic Future, Entrepreneurial Shifts | 08:39 | | Mind to Materialization: AI Startups in Minutes | 13:35 | | Agent Builder Demos / Critique of Visual Programming | 16:08 | | Voice Mode: From AV Control to API Naming Chaos | 20:15 | | Sora 2 API: Sketch-to-Video and Industry Deflation | 28:02 | | OpenAI’s Business Model & Global Expansion | 34:11 | | Anthropic OSWorld Benchmark: Superhuman Computer Use | 38:04 | | Grok Video Generation & Gaming Industry Focus | 41:30 | | Chips, Data Centers, Optics, and Energy Constraints | 47:30 | | Robotics: Tesla FSD, Gemini Robotics, Optimus & Labor | 58:31 | | Singularity, Recursive Loops, and Closing Thoughts | 68:09 |

13. Final Thoughts & Outlook

Peter:
“I don’t know how you asymptotically approach infinity, but we’re going to watch it happen.” (69:04)

Emad:
“We’re just at this tipping point, and the tipping point is in the next six months across just about all of these.” (66:51)

Alex:
“Increasingly, the innermost loop is going to look like recursive self improvement of robots building data centers, training better robots.” (67:19)

Dave:
“If you have a great team and you’re in AI, you’ll succeed every single time... You pivot so quickly...” (15:43)

14. Tone & Takeaways

The episode radiates high optimism, urgency, and marvel: humanity’s on the verge of “mind to materialization” via AI agents, reconstructed labor, and a new phase of economic and technological acceleration. The challenge is no longer “can it be built,” but where to allocate attention, compute, and capital fast enough as the world remakes itself under exponential change.

For more: Follow Peter on X, subscribe to his newsletter, and check out the next episode of Moonshots for ongoing coverage at the edge of the future.