Overview of The Jaeden Schafer Podcast
Host Jaeden (Jaden) Schafer covers three AI industry stories on his 30th birthday episode: (1) new data showing AI-powered apps struggle with long-term retention, (2) ChatGPT launching interactive, dynamic visual explanations for math and science, and (3) Thinking Machine Labs announcing a large multi-year compute partnership with NVIDIA. The episode mixes data-backed analysis, product context, and practical takeaways for builders and users.
Key takeaways
- RevenueCat’s 2026 subscription analysis finds AI apps monetize faster but retain users worse than non-AI apps.
- ChatGPT’s new dynamic visual explanations let users manipulate variables and see equations/diagrams update in real time across many math/science concepts — a major step for AI-powered learning.
- Thinking Machine Labs struck a multi-year strategic partnership with NVIDIA to deploy large-scale compute (multi-gigawatt-class deployments), signaling continued massive investment in AI infrastructure.
- Practical lesson for builders: novelty and hype drive quick monetization, but durable retention requires product utility, reliable UX, and realistic promises.
1) AI apps: fast monetization, poor long-term retention
Data highlights (RevenueCat — 2026 State of Subscription Apps)
- Dataset: subscription infrastructure across
75,000 developers; >1 billion in-app subscription transactions ($11B annual developer revenue). - Penetration: ~27% of apps were categorized as AI-powered; AI positioning is accelerating (≈1 in 4 apps now market themselves as AI-driven).
- Retention:
- 12-month retention: AI apps ≈ 21% vs non-AI apps ≈ 30.7%
- Monthly retention: AI apps ≈ 6.1% vs non-AI apps ≈ 9.5%
- Weekly retention: AI apps ≈ 2.5% vs non-AI apps ≈ 1.7% (AI slightly better for short-term/weekly models)
- Monetization & LTV:
- Median download monetization: AI 2.4% vs non-AI 2.0%
- Median monthly realized LTV: AI ≈ $18–$19 vs non-AI ≈ $13–$13.50
- Median annual realized LTV: AI ≈ $30 vs non-AI ≈ $20
- Refunds: AI apps show materially higher refund rates (median and high-end figures higher than non-AI apps; high-end examples cited ~15.6% vs ~12.5%).
Causes Jaeden highlights
- Overhype / overpromising AI capabilities leads to disappointment and churn.
- High experimentation by users — many apps are tried and dropped.
- Early product quality issues (bugs, missing features) increase initial churn.
- Strong competition and fast iteration mean “first mover” novelty fades quickly.
Implications
- AI features can boost initial revenue but sustaining subscriptions requires delivering consistent, durable value.
- App categories vary in AI adoption (photo/video high; gaming and travel low).
- Pricing models matter: weekly subscriptions often annoy users and may not fit true use cases.
2) ChatGPT: dynamic visual explanations for math & science
What’s new
- ChatGPT introduced interactive visual explanations that let users manipulate variables and see live updates to equations/diagrams.
- Supported across 70+ math and science concepts (examples cited: Pythagorean theorem, compound interest, exponential decay, linear equations, Coulomb’s law, Ohm’s law, kinetic energy, Hooke’s law).
Why it matters
- Moves beyond static textual answers to exploratory learning — useful for students, tutors, self-learners.
- Reinforces the role of AI as a teaching aid (not just an answer engine).
- Competes with similar interactive features from other assistants (e.g., Gemini).
Context & scale
- OpenAI reports very high weekly use of ChatGPT for math/science help (host cites 140M weekly for math and science).
- This feature aligns with broader trends of AI enhancing educational UX and deeper conceptual understanding.
3) Thinking Machine Labs — NVIDIA compute partnership
The announcement
- Thinking Machine Labs (Miriam Miratti / Miriam Maradi mentioned by host as a spinoff from top AI talent) announced a multi-year strategic partnership with NVIDIA to deploy large-scale compute systems starting in 2027.
- The deal includes deployment of at least gigawatt-scale NVIDIA AI systems and a strategic investment from NVIDIA.
- Thinking Machine Labs: raised >$2B since founding; cited valuation >$12B; launched first API product (Tinker).
Industry significance
- Reinforces the compute arms race: companies securing direct hardware commitments to scale model training and inference.
- Signals large capital commitments to infrastructure — Jensen Huang and others project trillions in AI infrastructure spending this decade.
- Aligns with other big compute arrangements in the market and underscores competition for GPU/accelerator capacity.
Notable insights & quotes (paraphrased)
- “AI features help apps monetize quickly — but sustaining that long-term is the real challenge.”
- “You have one shot to wow users in an era where people try many tools; if it flops, they move on.”
- Product advice: under-hype, over-deliver — focus on core utility, reliability, and user experience at launch.
Actionable recommendations (for founders & product teams)
- Prioritize core reliability and remove friction/bugs at launch to reduce early churn.
- Set user expectations conservatively: avoid marketing hyperbolic capabilities that can’t be reliably delivered.
- Focus on durable product value (regularly-used workflows) rather than novelty demos.
- Re-evaluate subscription cadence — prefer monthly/annual over frequent weekly billing unless it genuinely fits the product.
- Monitor refunds and early churn closely; treat them as signals for product-market fit and UX issues.
- Plan infrastructure commitments with realistic growth forecasts — secure compute partnerships only when product sustains demand.
Mentions & resources
- RevenueCat — 2026 State of Subscription Apps report (source for the retention/monetization data discussed).
- ChatGPT interactive visual explanations — new ChatGPT feature for math/science.
- Thinking Machine Labs — announced NVIDIA multi-year compute partnership; product: Tinker.
- Host’s product plug: AIbox.ai — access to 40+ models (mentioned as a place to try models discussed).
Final note from the episode: Jaeden celebrated his 30th birthday on this episode and asked listeners to leave a rating and review as a birthday gift.