The AI Podcast – Episode Summary

Episode Title: AI App Crisis, OpenAI Does Math, Big Nvidia Deal
Date: March 11, 2026
Host: Jaden Schaefer

Episode Overview

In this lively birthday episode, host Jaden Schaefer explores three pressing themes in the world of artificial intelligence:

• The troubling retention crisis facing AI-powered apps
• OpenAI’s new features for interactive math and science visualizations in ChatGPT
• Thinking Machine Labs’ billion-dollar compute deal with Nvidia

Jaden weaves together industry data, personal anecdotes, and future trends to deliver a value-packed analysis for both AI newcomers and professionals.

Key Discussion Points & Insights

1. AI Apps and the Retention Crisis

[02:10 – 12:00]

• AI-powered apps are experiencing high churn and trouble with long-term user retention, even though they monetize quickly.
• Data from Revenue Cat’s 2026 State of Subscription Apps (analyzing ~75k developers, $1B+ in subscriptions):
  • While 27% of subscription apps are now AI-powered, a whopping 73% remain non-AI.
  • AI apps see much lower 12-month retention:
    • AI: 21%
    • Non-AI: 30.7%
  • For monthly plans:
    • AI: 6.1%
    • Non-AI: 9.5%
  • AI apps struggle to build lasting value and user trust, in part due to industry-wide overpromising and underdelivering.
• Notable quote:
  • “As someone that has built AI-powered apps in the past…I can understand where a lot of this challenge is. I think with AI coming out and the power…being so incredible, we had a big wave…where a lot of people overhyped or oversold their apps and I think that’s going to be the primary driver of low retention.” — Jaden [03:10]
• Even old, non-AI app subscriptions are held onto more than newer AI ones—users dip into AI tools for novelty, but few become indispensable.
• Sectors with the highest AI inclusion:
  • Photo & video apps: 61% AI adoption
  • Gaming apps: only 6.2% AI adoption
  • Travel apps: 12.3%
    • “It is so wild to see that only 12% of the travel apps are actually using this. While every single AI company is basically using this category as the main demo.” — Jaden [07:50]
  • Business: 19.1% AI adoption

2. Value, Refunds & Monetization of AI Apps

[12:10 – 14:00]

• High rates of experimentation and shifting user expectations drive up refund and churn rates for AI apps.
  • AI app refund rates: as high as 15.6% (compared to 12.5% for non-AI)
  • Median refund for AI: 4.2%
• Despite retention struggles, AI apps extract higher lifetime value per user:
  • Monthly LTV: AI – ~$18, non-AI – ~$13
  • Annual LTV: AI – $30, non-AI – $20
• Notable quote:
  • “AI features are going to help apps monetize really quickly. But sustaining that long-term is going to be the challenge.” — Jaden [13:45]

3. OpenAI’s Dynamic Visual Explanations in ChatGPT

[14:15 – 15:30]

• OpenAI rolled out a new feature: “Dynamic Visual Explanations” in ChatGPT—users can now manipulate interactive diagrams for 70+ math and science concepts (e.g., move triangle sides to see the hypotenuse change real-time).
• Aimed at deeper learning: enables not just answers but comprehension.
• “I actually think it’s an incredible tool for education that’s going to make us smarter…OpenAI says more than 140 million people use ChatGPT weekly for math and science help.” — Jaden [15:00]
• Other AI giants (e.g., Gemini) are following suit, making educational AI a new competitive frontier.
• Tied to broader industry trends: “the race to build the next generation of AI infrastructure is going to be interesting.” — Jaden [15:40]

4. Thinking Machine Labs’ Billion-Dollar Nvidia Compute Deal

[15:45 – 18:05]

• Thinking Machine Labs (founded by former OpenAI exec Miriam Maratti) inked a multi-year compute partnership with Nvidia, including a commitment to at least 1 gigawatt of Nvidia’s latest AI systems starting in 2027.
  • Nvidia is also a strategic investor; TML’s valuation is over $12B.
  • Focus: building more reliable, replicable AI models (with first product: Tinker API)
• Competitive pressure for compute is “heating up”:
  • Industry-wide trend: OpenAI’s $300B Oracle compute partnership cited as precedent
  • Nvidia CEO Jensen Huang predicts AI infra spend will top $3–4 trillion by decade’s end.
• Notable quote:
  • “Honestly, even just going to sign in a gigawatt deal—it’s a lot of confidence that their product is going to be incredibly useful.” — Jaden [17:00]
  • “AI companies are really having to be very aggressive in how they compete for access to this computing power.” — Jaden [17:35]

Memorable Quotes & Moments

• “Right now, you have one shot really for someone to try your tool…and for it to wow them. If they try it and it flops…I haven’t gone back.” — Jaden (relating personal app experience) [04:00]
• “We should probably normalize…being able to under-hype your app, but it being super useful.” [14:00]
• “When you build a product…make sure it works really good on launch and then you’re going to be able to keep your churn up.” [18:00]

Major Segment Timestamps

• [03:00] – The AI app retention crisis: Overview & revenue report highlights
• [06:15] – AI adoption rates by app sector
• [09:45] – User retention and refund behavior
• [13:45] – Monetization insights & challenges for AI apps
• [14:15] – OpenAI’s new dynamic interactive explanations for math/science
• [15:45] – Thinking Machine Labs–Nvidia mega deal and AI infrastructure arms race
• [17:40] – Strategic context: compute competition and lasting app value

Tone & Style

Jaden’s tone is enthusiastic, direct, and relatable, blending industry data with real-world experience (“…I try probably 10 times as much software as I have over the last five, ten years…” [03:50]) and the occasional streak of humor and friendly candor—especially around app hype, subscription annoyances, and his own 30th birthday.

Takeaways for AI Enthusiasts & Builders

• Retention, not just revenue, is the new challenge for AI apps.
• Overhyping features leads to short-term novelty but long-term churn.
• Premium AI app users pay more, but expect more—and churn faster if disappointed.
• Education is a key battleground, with GPT and Gemini innovating interactive learning features.
• Race for compute power is fierce—big money, long-term deals, and high stakes.
• Ultimately, successful AI products will quietly deliver durable, indispensable value—not just launch with flashy new tricks.