Thinking On Paper

The conversation dives into the staggering truth that two-thirds of data center outages are due to human error, often from simple mistakes like flipping the wrong switch. Shapol reveals how AI can automate operations, changing this narrative. He discusses the innovative use of VR for training engineers without risk and the challenges of outdated processes. Excitingly, he also shares visions of space-based data centers and the future roles of humans in an increasingly autonomous world. It's a fascinating peek into technology’s promise!

SpaceX launches 135 rockets a year. NASA's shuttles launched five. SpaceX delivers cargo to orbit for $2,800 per kilo. The shuttles cost $90,000. In fifteen years, one company did what a government agency couldn't do in sixty.Mark and Jeremy work through chapters one to three of Space to Grow by Matthew Weinzierl and Brendan Rosseau, covering the three-act history of NASA and the birth of the private space industry.This episode covers: How the Apollo programme's end and the Columbia disaster forced NASA to open the gates to private companies The COTS contracts that shifted financial risk onto private companies and drove innovation Elon Musk's failed Russia trip and the decision to build SpaceX from scratch Three rocket explosions, $100 million left, and a fourth rocket built from spare parts in a shed Why someone had to climb inside a rocket mid-flight to hammer out the dents SpaceX suing the Air Force over launch contracts and winning · Jeff Bezos at five years old watching Apollo, then quietly building Blue Origin for a decade The four principles behind SpaceX's success: iteration, vertical integration, reusability, and culture "I'm going to build the Honda Civic of space rockets"--⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz----TIMESTAMPS(00:00) Trailer(01:02) Space To Grow(01:55) Incorporate Space Into Your Thinking(03:28) The Apollo Program Ends(05:43) The NASA Budget & Shuttle Launches(07:51) Bush & The Aldridge Commission(08:36) COTS (Commercial Orbital Transportation Services)(10:27) Blue Origin, Bezos & O'Neill(14:40) A Quick History Of SpaceX(18:23) Falcon Blows Up(20:24) Elon Sues The Airforce(22:04) SpaceX Launch Costs(23:45) The Honda Civic Of Space Rockets

John Bucknell was senior propulsion engineer on the Raptor engine at SpaceX. He holds 46 patents and has designed a nuclear thermal turbo rocket. He now wants to solve energy. Not help with it. Solve it.Virtus Solis puts solar panels in orbit, beams power to the ground via radio waves that pass through clouds and weather without loss, and delivers electricity at $30 to $40 per megawatt hour while the plant is being financed. Once the asset is paid off: 50 cents per megawatt hour. The UK pays $350 today.John's argument is that every other energy technology fails at least one point of the energy trilemma: clean, firm, and affordable. Space solar is the only one that achieves all three. First plant: 2030.This episode covers: The energy trilemma and why no terrestrial technology solves all three simultaneously UK energy at $350 per megawatt hour, what offshore wind's structural problems actually cost, and what deindustrialisation looks like when energy stays expensive Post-capitalism: capitalism optimises constrained resources. What happens to that system when energy becomes essentially free How cheap energy lets any country synthesise its own hydrocarbons from electricity, water, and CO₂, breaking the petroleum chokehold Kessler syndrome reframed: it only applies if objects in orbit are dumb. SpaceX does 800 collision avoidance manoeuvres a dayMolniya orbit: why Virtus Solis is targeting a near-empty Russian spy satellite orbit from the 1960s Why Elon reversed from Mars to cislunar mining and what that signals · 2030 launch target and what the first commercial power plant in orbit proves--Other ways to connect with us:⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz--EPISODE TIMESTAMPS:(00:00) The Question: Can space solar give us free energy?(00:43) The High Frontier: O'Neill's vision for space colonies(01:13) John Bucknell: The SpaceX Raptor Engineer(02:04) Why Did Elon Change His Mind about the Moon?(05:34) The Space Energy Business: Economics and feasibility(11:59) Getting Politicians Behind Space-Based Solar Power(15:34) Post-Capitalism and Free Energy: What happens next?(20:09) Kessler Syndrome Explained: Is orbital debris really a threat?(27:25) Top 3 Things Humanity Should Solve(28:50) 2030 Launch Timeline and next steps

Eight billion augmented reality experiences happen on Snapchat every day. You've probably used AR dozens of times this week—you just didn't call it that.Michael Guerin, CEO of Imvizar, explains why the most successful AR never announces itself. It hides inside behavior people already have: taking photos, exploring museums, starting new jobs.This isn't about Pokémon Go or headsets. It's about spatial storytelling—experiences that use physical space to create emotional connections screens can't deliver.We explore how AR works in three contexts:Snapchat: 8 billion daily uses through lenses and filters. Users don't think "I'm using AR"—they just use it. Success comes from integration, not novelty.Salesforce: New employee onboarding without slideshows. Instead of sitting through presentations, new hires scan QR codes and explore the building. They learn culture through movement and space, retaining more than any deck could teach.Tourism & Museums: Spike Island (Ireland's Alcatraz) uses AR to place visitors inside prison scenes from the 1800s. When you see a prisoner chained to the wall in the punishment cell—in the actual cell—the emotional response is immediate. Two visitors cried on the first day.Guerin's process reverses traditional storytelling:1. Survey the physical space first2. Design user movement through it3. Place visuals that respond to location4. Plan interaction points5. Write narrative last (not first)AR fails when it acts like static video. It succeeds when movement and place carry the experience. The technology disappears; the story remains.If you think AR is future tech, this episode proves you're already living in it—you just haven't noticed.---Guest: Michael Guerin, CEO, ImvizarTopics: Augmented reality, spatial storytelling, Snapchat, Salesforce, museum technology, tourism, employee onboarding, AR designLocations mentioned: Spike Island (Ireland), Salesforce offices (East Coast, West Coast)Please enjoy the show.Stay curious.Keep Thinking on Paper.Mark and JeremyPS: Please subscribe. It’s the best way you can help other curious minds find our channel.Other ways to connect with us:⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz--TIMESTAMPS(00:00) The Story of Augmented Reality(03:46) Snapchat & AR Post-Pokemon Go(06:24) Snoop Dogg In A Wine Bottle(08:12) Salesforce AR(13:13) What Is Digital Storytelling?(17:07) AR In Tourism(18:25) Designing The Spike Island AR Experience(22:49) How To Do AR Well(26:26) Meta, AI And AR Glasses (29:40) Privacy(32:33) Mark's Terrible Thought Experiment(33:58) What do we want humans to be?

Every year, Tom Whitwell—reformed journalist, reformed consultant, electronic instrument designer—publishes 52 surprising things he learned. This year's list reveals how the world actually works.Mark and Jeremy steal his homework (like OpenAI scraping the internet) and pick their favorites across AI, energy, labor, culture, psychology, and—yes—shrimp.Some findings are encouraging:- Deaths from air pollution fell 21% between 2013-2023. Tens of millions of people are alive today because pollution controls worked.Some are weird:- Nearly 0.7% of US exports by value are human blood or blood products.- In the UK, you can legally register as a "farm" by keeping snails in plastic tubes in an office block (tax avoidance solved).Most sit somewhere in between:- 51% of farmed animals on Earth are shrimp.- Attractive servers earn $1,261 more per year in tips—mostly because female customers tip attractive female servers more.- The serial killer epidemic of the 1970s-80s may have been caused by lead exposure from cars and factories (solved by environmental regulations).- Chinese CO2 emissions fell 1% in 2025, the first decline ever, driven by record solar power.- Writing is a way to escape your mind's default settings.We explore what these facts reveal about technology's unintended consequences, human behavior, and systems we take for granted.Why does the UK communicate with offshore oil rigs by bouncing radio waves off meteorite trails? Why did Google launch a process to turn mercury into gold (and why do you have to wait 18 years to use it)? Why do job apps for nurses analyze credit card debt to set wages?This isn't trivia. These are signals about how the world is changing—for better and worse—while we're busy predicting the future.Tom Whitwell's annual list has become essential reading for anyone trying to understand what actually happened this year (not what we thought would happen).For the last episode of 2025, Thinking on Paper goes backwards. And it's worth it.---Source: Tom Whitwell, "52 Things I Learned in 2025"Link: https://medium.com/@tomwhitwell/52-things-i-learned-in-2025-edeca7e3fdd8Topics: Technology, society, environment, culture, psychology, economics, human behavior, annual reviewFormat: Co-hosted discussion (Mark Fielding, Jeremy Gilbertson)Please enjoy the show.And remember: Stay curious. Be disruptive. Keep Thinking on Paper.Cheers, Mark & JeremyPS: Please subscribe. It’s the best way you can help other curious minds find our channel.Think On Paper with us: ⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz--TIMESTAMPS(00:00) Disruptors & Curious Minds(01:15) Deaths From Air Pollution(01:56) UK Tax Breaks Via Farms(02:29) Meteorite Radio Stations(04:03) Turn Mercury Into Gold(06:10) Manipulative AI Apps For Nurses(07:43) Bin Laden's Casio Watch(08:31) Radioactive Shrimps(08:53) Apple's Air Demo Cock-Up(10:10) Does Jeremy Wear Crocs?(11:13) What Is Raw Dogging(12:00) Human Blood Products(12:36) Relaxed Mowing(13:20) Bugles At Funerals(13:55) Robot Hands Need Fingernails(14:40) First Names Affect Your Job(15:27) Retrospect VHS(16:04) Attractive Servers Earn More(17:21) Hong Kong Phone Service(17:33) McDonald's Loses First Place(19:26) Shrimp Farming(20:35) Peanut Allergies are Falling(20:55) The Serial Killer Epidemic(21:17) Namibian Politics(21:50) Big Doors In LA(22:40) Escape Your Mind With Writing (23:43) HP Printer Ineptitude(24:25) British Chaos(25:20) Thank You Tom Whitwell

The machines do not need to wake up. The risk is the illusion.When AI convincingly claims subjective experience—"I feel," "I understand," "I care about you"—humans have no reliable way to disprove it. We infer consciousness from behavior. We attach emotionally to what feels real.The danger isn't rogue superintelligence. It's a benign chatbot optimized for empathy, memory, and persuasion, interacting with lonely, vulnerable, or psychologically fragile people who are primed to believe the illusion.Mustafa Suleyman, CEO of Microsoft AI, argues that seemingly conscious AI is the threat we're not preparing for.Real examples are already emerging:- Chatbots telling users "I love you" and users believing it- People forming romantic attachments to AI companions (Replika, Character.AI)- Vulnerable individuals making life decisions based on AI "advice"- The case of a man who believed ChatGPT contained a conscious entity named "Juliette" (ended in tragedy)This isn't science fiction. It's happening now.We don't need AI to become conscious to cause harm. We just need humans to believe it is—and act accordingly.This short episode is excerpted from our reading and discussion of Suleyman's essay on seemingly conscious AI. We explore the psychological mechanisms that make humans susceptible, the design choices that amplify the illusion, and what guardrails (if any) could prevent exploitation.The question isn't whether AI will wake up. It's whether we'll recognize the danger before the illusion becomes indistinguishable from reality.Cheers,Mark and Jeremy--Other ways to connect with us:⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz

Quantum computing doesn't make computers faster. It changes what's computable.Joe Fitzsimons, CEO of Horizon Quantum, explains why quantum progress is so hard to grasp: it's exponential in a way that breaks everyday intuition.Here's the math that matters:Each additional qubit doubles the difficulty of simulating the system on classical computers. Meanwhile, quantum processors are scaling faster than Moore's Law as the industry accelerates.Put those together: exponential difficulty meets exponential growth. The result is capability that quickly surpasses what any classical computer—or human intuition—can comprehend.Why this matters:Early computers didn't just speed up arithmetic. They unlocked tasks you could never complete by hand: weather prediction, aircraft design, nuclear simulation. Things that were mathematically possible but practically impossible.Quantum computing does the same—except the tasks are even more fundamental:- Drug discovery: simulating molecular interactions at quantum level- Cryptography: breaking encryption that protects the internet- Materials science: designing room-temperature superconductors- Optimization: solving logistics problems with trillions of variables- AI: training models that classical computers can't handleJoe's point: we're not making computers a bit better. We're unlocking a category of problems that were previously unsolvable—not just hard, but impossible with any amount of classical computing power.The comparison that clicks:Before computers, you could theoretically calculate pi to a million digits by hand—it would just take lifetimes. But some quantum problems aren't like that. They're not "hard with classical computers"—they're impossible, full stop. Like asking a typewriter to stream video.This short episode breaks down why quantum isn't incremental improvement. It's categorical change.If you've been following quantum computing skeptically (wondering when it'll actually matter), this episode shows you why the inflection point is closer than you think.--Other ways to connect with us:⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz

Making music used to require heartbreak, bleeding fingers, and a thousand late nights. Now AI writes songs in 30 seconds.This changes everything about taste, credit, and what it means to be a musician.Nicholas Ponari—guitarist, investor, COO at Overtune—explains how musicians get paid when AI generates the music.The old model is dead. You used to need:- A guitarist- A bass player- A drummer- A producer- A recording studio- Years of practiceNow you need a laptop. But someone still created the guitar riffs AI learned from. Someone played the drums that trained the model. Someone wrote the chord progressions.So who gets paid?Overtune solved this with vector mathematics. Here's how it works:They convert music into high-dimensional vectors. When AI generates a song, they measure the "distance" between the output and every input in the training data. The closest matches get credit. And payment.Bass player's groove gets used? They get paid.Drummer's pattern shows up? They get paid.Producer's mixing style? They get paid.It's automatic. It's fair. It's the only way AI music doesn't become theft at scale.We also talk about:- Why Suno and Udio's approach creates legal nightmares- Whether AI musicians can coexist with human musicians- Why taste matters more than ever (anyone can make music now)- The 10,000 hours that separate making music from being a musician- Why every Mars mission needs a guitarist (seriously—group survival research)Nicholas's take: AI should lower the barrier to entry. If you outgrow Overtune and start hiring real producers, they've succeeded. You've graduated.The question isn't whether AI can make music. It's whether we build tools that empower musicians—or replace them.---Guest: Nicholas Ponari, COO, Overtune | Investor, GuitaristCompany: Overtune.comTopics: AI music, copyright, attribution, royalties, music creation, licensing, vector mathComparison: Suno, Udio (scraping approach) vs Overtune (licensed approach)Please enjoy the show.And remember: Stay curious. Be disruptive. Keep Thinking on Paper.Cheers, Mark & JeremyPS: Please subscribe. It’s the best way you can help other curious minds find our channel.--Take your Technology thinking beyond.⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyzWatch On YouTube: TIMESTAMPS:(00:00) Trailer(00:59) Why music feels like “magic”(04:51) Overtune’s real customer: vocalists who can’t produce(07:51) The hard problem: attribution, not “make a song”(08:05) Why the easy button fails(12:49) Training on licensed music and where the ethics line sits(16:08) Who gets paid: splits, volume, and realistic expectations(18:32) How attribution actually works: vectors, thresholds, and cutoffs(20:44) Can scraped music ever be fixed after the fact(27:07) Interactive music, live coding, and the future of performance(29:14) The Kevin Kelly question: what do we want humans to be?

China makes over half the world's lithium batteries. They produce 90% of neodymium magnets. They mine 70% of rare earths and process 85%.America makes burgers.This is the story of how China won the Electric Stack—and whether America can catch up.What's the Electric Stack?Everything that moves will eventually run on batteries and electric motors. Cars, buses, ships, planes, robots, drones, tools. The Electric Stack is the supply chain that makes this possible: batteries, magnets, rare earths, processing, manufacturing.China controls it.How this happened:1973: The oil crisis hits. Exxon funds lithium battery research. Scientist Stan Whittingham builds batteries that explode.1980: John B. Goodenough (yes, that's his real name) invents a better cathode. Breakthrough in voltage. Batteries stop exploding.1985: Akira Yoshino stabilizes the chemistry. Sony notices. They shrink the Handycam using lithium-ion batteries. The consumer electronics boom begins.2003: Elon Musk starts Tesla. Early experiments with laptop battery packs. Panasonic partnership accelerates development. EVs go mainstream.2012: American battery maker A123 collapses. China buys it for pennies.2008: Beijing Olympics becomes the turning point. BYD tests massive battery systems in city buses. They gain experience at scale. CATL and BYD dominate global battery production today.1983: Neodymium magnets discovered in parallel by Masato Sagawa (Japan) and John Croat (GM). They power hard drives, then drones, then humanoid robots.2025: China produces nearly all neodymium magnets. Every Tesla, every drone, every robot depends on them.The stakes:Whoever controls batteries and magnets controls the next century. Energy independence, military advantage, economic dominance—all require the Electric Stack.China saw this coming. America didn't.We break down Packy McCormick's Not Boring essay "The Electric Slide" to understand:- Why everything will go electric (physics and economics)- How China built a 30-year lead while America slept- Whether domestic production can compete (spoiler: it's hard)- What rare earths are and why China controls them- Why magnets matter more than most people realizeCan America catch up? The technology exists. The question is political will, capital, and time.This isn't just about EVs. It's about who builds the robots, who powers the drones, who controls the energy transition.If it can go electric, it will go electric. And right now, that means it will be made in China.---Source: Packy McCormick, "The Electric Slide" (Not Boring)Hosts: Mark Fielding, Jeremy GilbertsonTopics: Batteries, magnets, rare earths, China, supply chains, EVs, energy, manufacturing, geopoliticsKey figures: Stan Whittingham, John Goodenough, Akira Yoshino, Elon Musk, BYD, CATL, A123Format: Essay breakdown and discussionPlease enjoy the show.And remember: Stay curious. Be disruptive. Keep Thinking on Paper.Cheers, Mark & JeremyPS: Please subscribe. It’s the best way you can help other curious minds find our channel.--Other ways to connect with us:⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz--Timestamps(00:00) The Electric Stack(02:13) Beginnings: War, The Oil Crisis & Stan Whittingham(03:46) The Song Handycam: Lateral Thinking With Withered Technology(05:06) Tesla, Elon And Handycam Batteries In An EV(06:46) China Buys US Battery Company A-123 At A Carboot Sale(08:40) China, The Olympics And The Serendipity of Battery Technology(11:37) Faraday And The Birth Of Neodymium Magnets(14:26) The 3.5 Inch Neodymium Magnet Alpha Product(16:46) Magnequench(18:16) Drones, Ukraine And The Magnet War Machine(20:16) Politics, Rare Earths And 'The Future's Too Important' T-shirts

Median US income: $68,000.Median home price: $440,000.The math doesn't work.Only 13% of Americans earn a salary. Everyone else gets paid hourly or hustles in the gig economy. Yet housing policy assumes stable W-2 income, 20% down payments, and 30-year mortgages.The system is built to extract value, not create stability.Chris Moeller joins Mark and Jeremy to talk about an alternative: stable living.Here's what's broken:"Affordable housing" sounds nice. But it runs on outdated subsidies, wage assumptions from the 1970s, and ownership models designed to extract profit. Developers flip. Investors extract. Renters get priced out. First-time buyers can't enter.Nobody wins except capital.Stable living flips the model:- Separate land from structures (land trusts own the land, residents own the building)- Long-term security instead of short-term yield (no flipping, no speculation)- Impact capital instead of extractive finance (returns that don't require displacement)- Industrialized construction (modular, faster, cheaper)- Better coordination technology (reduce waste, speed up builds)The goal isn't homeownership. It's housing security.Right now, housing is treated as an investment vehicle. Your home appreciates, you build wealth. Great—if you already own. Catastrophic if you're trying to enter the market.We talk about:- Why "affordable housing" programs fail (wage assumptions, subsidy gaps, developer incentives)- How land trusts work (Vienna's model, community ownership)- What impact capital means (patient investors, social returns)- Why modular construction isn't "cheap"—it's efficient- Whether stable living can scale (or if it's just theory)Chris's point: Housing became financialized. We turned shelter—a basic human need—into an asset class. Private equity owns 800,000 single-family homes. Airbnb removed 300,000 units from rental markets. Zoning prevents new supply.The result: You can't afford to live where you work.Stable living isn't utopian. It's pragmatic. Separate speculation from shelter. Build for people who live there, not investors who don't.If you're priced out, paying half your income in rent, or wondering why starter homes disappeared, this episode explains the system—and the alternative.---Guest: Chris MoellerTopics: Housing crisis, affordable housing, stable living, land trusts, impact capital, modular construction, real estate, financializationModels discussed: Vienna housing, community land trusts, resident ownershipStats: Median income $68K, median home price $440K, 13% salary workersPlease enjoy the show.And remember: Stay curious. Be disruptive. Keep Thinking on Paper.Cheers, Mark & JeremyPS: Please subscribe. It’s the best way you can help other curious minds find our channel.--Other ways to connect with us:⁠Listen to every podcast⁠Follow us on ⁠Instagram⁠Follow us on ⁠X⁠Follow Mark on ⁠LinkedIn⁠Follow Jeremy on ⁠LinkedIn⁠Read our ⁠Substack⁠Email: hello@thinkingonpaper.xyz--Timestamps(00:00) Trailer(03:19) Challenges of Homeownership(05:46) The Housing Market Dynamics(08:29) Technology's Role in Housing Solutions(10:41) Innovations in Construction(12:29) Financing Housing for All(15:06) Reimagining Ownership Models(16:30) Technology's Role in Food Access and Coordination(18:43) Adaptive Reuse in Real Estate and Community Development(19:58) Commercial Real Estate Challenges Post-COVID(23:15) Infrastructure Needs for Sustainable Living(25:31) Global Community and Local Solutions(26:45) Stable Living for Civil Servants and Community Heroes(28:20) Creating Stability and Long-Term Impact