We Study Billionaires – Infinite Tech

Episode: TECH009 — Data Centers in Space, AI Education, Haptic Touch Robotics, and More with Seb Bunney
Host: Preston Pysh
Guest: Seb Bunney
Date: December 17, 2025

Episode Overview

In this episode of Infinite Tech, host Preston Pysh and recurring guest Seb Bunney dive into the frontiers of exponential technology. They discuss:

• How AI is driving personalized healthcare and the crucial issue of genetic privacy
• Google and SpaceX’s moonshot projects: data centers in space and the cost challenges
• The revolution in education: AI-powered personalized learning and its impact on traditional schooling
• Breakthroughs in haptic technology, especially for robotics and remote touch
• The broader societal implications: regulation, AI bias, and figuring out what is ‘signal’ vs. ‘noise’ in a data-saturated world

Throughout, the conversation is lively and speculative, with both hosts bringing personal anecdotes, thoughtful skepticism, and excitement for the future.

Key Topics & Insights

1. AI-Driven Personalized Healthcare & the Privacy Dilemma

[02:40–13:01]

• Personalized Health Breakthroughs:
  Seb introduces the work of Gary Brecka, who pioneered holistic, genetics-based health evaluation (“ultimate human”). Rather than a generic approach, Brecka leverages methylation pathways and key genes (MTHFR, MTR, COMT, CBS) to create deeply customized supplement protocols.

  “What’s so cool about what is happening in the world today is we’re starting to get… personalized healthcare. We’re starting to have a personalized supplement protocol as opposed to… just blindly throwing darts at a board.” —Seb [06:32]

• AI + Genomics:
  Preston and Seb discuss using AI analysis to process and interpret vast personal genetic data—down to running personal genetic info through ChatGPT for tailored health advice.

  “I basically made a Gary Brecka bot. And I said, I want you to be able to look through my genetics… and try and find the MTHFR gene, the MTR gene, …go and see if there's a mutation… Just by taking B12, I noticed a huge difference in my health.” —Seb [10:25]

• Privacy & Risk:
  With advances come serious privacy concerns. Uploading DNA to third parties or LLMs (like ChatGPT) could expose sensitive data, an issue notably raised by the bitcoin community and in light of breaches like 23andMe.

  “How do you possibly go about it in a way that protects the privacy of the people? …It gets to be somewhat concerning and a lot of people don’t want to talk about that side of it.” —Preston [09:32]

• Doctors and AI:
  Increasingly, doctors themselves are using AI tools like ChatGPT to double-check symptoms, as the information AI can glean is incredibly broad compared to any human’s specialized knowledge.

• ‘AI Slop’ and Validity:
  A potential downside: mass production of health and self-help literature using AI without true expertise—raising issues around information reliability and “AI slop.”

  “We’re putting increasing amounts of trust into this thing and we’re not actually looking at the validity of the information being produced.” —Seb [16:01]

2. Space-Based Data Centers: Moonshot or Future Reality?

[20:12–28:54]

• Google’s Project Suncatcher:
  Google aims to launch prototype orbital data centers by 2027, hoping to harness solar energy at 8x Earth efficiency. Elon Musk has signaled interest.

  “There’s no doubt to me that a decade or so away we’ll be viewing it as a more normal way to build data centers.” — Sundar Pichai (Google CEO, clip) [20:35]

• Cost Barrier:
  The critical factor: Launch costs must fall by 10x for viability, echoing difficulties faced by space bitcoin mining advocates.

  “The cost for a space transport to get the hardware just into space had to drop 10x. … Before this would even be viable to even begin doing this for real.” —Preston [22:01]

• Technical Hurdles:
  Reliability risks (hardware degrading under rapid temperature swings) and communication latency (especially for tasks needing real-time speed) are significant obstacles.

  “If you mine a block in space, by the time you actually propagate that block to the blockchain, someone on Earth may have already found a block… So you've got a disadvantage already just by being in space.” —Seb [24:38]

• Use Cases:
  While unsuitable for real-time or latency-sensitive tasks, orbital data centers might excel at large-scale AI training and batch processing.

• Environmental & Regulatory Concerns:
  Discussion of “Kessler Syndrome,” the risk of runaway space debris as more hardware is launched.

  “Is there a point where… sending all this stuff up into space, we’re impeding our ability in the future to be able to go further afield?” —Seb [28:08]

3. AI & Personalized Education: Upending the Classroom

[29:22–42:29]

• AI-Powered Tutoring:
  Tools like Google’s “Learn Your Way” offer dynamic, individualized tutoring—creating audio lessons, quizzes, mind maps, and tailoring delivery to a student’s interests and pace.

  “Imagine being in a classroom, but essentially having a one on one teacher at all times being able to support you.” —Seb [29:42]

• Impact on Learning:
  Early studies show recall improvement, though the real promise is how AI can adapt content to each learner’s style and passions—moving away from the “Victorian” one-size-fits-all model.

  “The AI is going to learn these different techniques that people have for what's most optimal for them… the best way to frame it.” —Preston [35:27]

• In-Person vs. Digital:
  Personal anecdotes highlight the limitations of screen-based learning (as seen during COVID-19) and the value of peer/social interaction and nervous system co-regulation that human teachers provide.

  “There is also the human connection… we’re also learning how to regulate our nervous system… Are they [AI teachers] going to have that empathy and that capacity?” —Seb [38:16]

• The Future: VR and Robots?
  Will an “AI humanoid robot” as teacher be enriching—or dystopian?

  “My wife hearing me say… ‘yeah, you're going to have an AI humanoid robot like teaching’… she would just be disgusted.” —Preston [39:19]

• Unlocking Human Potential:
  Both hosts observe how, when allowed to pursue personal interests post-schooling, their real intellectual growth began—underscoring the need for education to pivot toward individual curiosity.

4. AI Bias, Regulations, and LLM Interpretation

[42:39–56:57]

• Simulators, Not “Thinkers”:
  AI expert Andrej Karpathy’s advice: don’t ask LLMs what “they” think; have them role-play experts or summarize viewpoints for diverse perspectives, to minimize inherent bias.

  “Next time try: what would be a good group of people to explore X, Y and Z? What would they say?... If you force it via the use of ‘you,’ it will give you something by adopting a personality embedding vector.” —Karpathy, paraphrased by Preston [43:34]

• Geopolitics & Regulation:
  The Biden administration is pushing for federal-level regulation on AI, preempting state-specific rules to create consistency (often at the behest of big tech).

  “From a states rights standpoint, one of the advantages is it creates competition between the states… but with AI… the product that’s being built here, I don’t know that it has a benefit or a negative for that state over another…” —Preston [49:41] “At what point do we need certain regulation? Because people are making decisions that are far more detrimental on a bigger scale.” —Seb [51:45]

• Regulation Dilemma:
  Both hosts note the tension between free-market growth and the need for oversight, especially when risks are not local but global. Issues of regulatory capture and conflicts of interest are highlighted.

5. LLMs, Memory, and Information Theory

[52:55–61:45]

• Google Titans: Memory Breakthroughs
  New “Titans” architecture allows LLMs to maintain high accuracy over tens of millions of input tokens by learning to focus on “surprise”—i.e., what’s unexpected compared to their entire training dataset.

  “In Titans, the AI uses a similar surprise signal… Deciding to store that useful bit in its long term memory.” — Preston [54:17] “I distinctly remember him saying… information theory… is just surprise. Like, his algorithm is just looking for surprise.” —Preston [56:57]

• But What Matters?:
  Both question how an AI “knows” what is valuable info to remember and whether, like humans, it ends up with rules of thumb (heuristics) that are optimal for something—but what? Survival, utility, or arbitrary gradients?

  “What is surprise to the AI? … Is that actually relevant?” —Seb [59:35]

• Limits of Automation:
  Even with such advances, domain experts are still needed to distinguish “signal from noise.”

  “You still need to be a specialist and deeply understand the topic to understand the validity of the output from a lot of these models.” —Seb [60:48]

6. Long-Distance Haptic Touch & the Next Human-Machine Interface

[62:03–69:51]

• Breakthroughs in Remote Touch:
  New wearable haptic patches (~1.1 mm) can transmit tactile sensations—pressure, vibration—long-distance, enabling next-level physical (and emotional) presence in AR/VR or remote robotics.

  “…If someone else was, say, wearing the same glove and they went and touched textures… you could feel what they are feeling through these haptic touch sensors.” —Seb [63:29]

• Applications:
  Training humanoid robots (so they can grip or handle fragile objects) and enabling more immersive virtual or remote human interaction.

  “The amount of pressure that it’s applying to the objects that it’s interacting with become really important… haptics are a huge part of humanoid robots and where a lot of that’s going to be going.” —Preston [65:05]

• Philosophical Musings — Are We in a Simulation?
  Haptic technology and hyperrealistic AI-generated environments prompt speculation: what if our reality is itself a simulation, refined through ever-advancing tech?

  “Is AI and robotics… being created or is it being rediscovered, as in, it already exists?” —Seb [67:02] “It’s one of Elon Musk’s biggest talking points… when this topic of simulation theory comes up.” —Preston [69:29]

Notable Quotes & Moments

• On Custom Health AI:

  “I actually haven’t had a cold in four to five years. I got one this weekend, and it’s the first one. Because you were going to be talking about it.” —Seb & Preston [06:32]

• On Education:

  “School, to me, I never felt like I fit in. But once I left school and I was able to find audiobooks, I was able to find my ability to find how I learn. Oh my God. It was… profound.” —Seb [40:42]

• On Physical vs. Digital Instruction:

  “There is also the human connection… And I think part of during our developmental years while we’re in school, we’re also learning how to regulate our nervous system.” —Seb [38:16]

• On Reality vs. Simulation:

  “How do you know what is real and what you’re experiencing is reality? … What you think, you know, is not that at all. It’s very, very, very different from a quantum mechanics standpoint.” —Preston [71:06]

Timestamps for Key Segments

• 02:40–13:01 — Personalized AI-Driven Healthcare & Privacy
• 20:12–28:54 — Space-Based Data Centers: Barriers & Opportunity
• 29:22–42:29 — Personalized Education: AI Tutors & The Future of Learning
• 42:39–56:57 — AI Bias, Regulation, and Finding the Signal in the Noise
• 56:57–61:45 — Google Titans: LLM Memory, Surprise, and Information Theory
• 62:03–69:51 — Haptic Touch Robotics & Simulation Theory

Tone, Language & Style

• Engaged, speculative, and enthusiastic
• Frequent self-deprecation and personal anecdotes
• Technical concepts broken down for non-experts, but with respect for detail
• Open musings about regulation and philosophy, with a skeptical but optimistic slant

Takeaways for Investors & Tech Enthusiasts

• Customization, powered by AI (in health and education), is set to disrupt legacy models—but privacy and data security will be crucial.
• Even moonshot projects (like orbital data centers) hinge on seemingly “boring” factors like launch costs.
• The boundary between the digital and physical is blurring fast—watch the interface layer, from VR to haptics to humanoid robots.
• As AI helps us process more information, how we ask questions and who creates regulations may matter as much as the technology itself.
• Information theory’s core (surprise, signal, noise) is shaping both how humans and AIs “learn”—and will dictate whose intelligence reigns.

**For further reading and deeper dives, listeners are encouraged to check out Seb’s work (@sebunney), his book The Hidden Cost of Money, and follow Infinite Tech for subsequent discussions on exponential tech trends.