Frontiers Podcast

John and Steve ask: What happens when a computer can read your brain, and eventually write to it? If you could download French like an app, would you still need to learn it? And if two brains could talk directly to each other, what’s left of language, privacy and the private self?

Topics:

• Visions of the future: Elon Musk’s AI-generated video of a spacefaring civilization, space elevators, permanent bases on Mars and the Moon, and why hard tech outside of IT has barely moved (curvier cars aside).
• Reading the brain: Aleph Neuro’s non-invasive neurovascular ultrasound, the most detailed vascular image of a living human brain ever captured, and rebuilding what someone sees (a red London double-decker, a seagull) from brain activity alone.
• Medicine’s final frontier: Why dementia diagnosis is still so subjective, why “do more Sudoku” headlines are false hope, and the case for a yearly brain scan as routine as a blood pressure check.
• The BCI landscape: Neuralink, Synchron, Sam Altman’s Merge Labs and the invasive vs. non-invasive divide, plus why the field is waiting for its Raspberry Pi moment and still wide open for young engineers.
• The next computing interface: From the command line to the GUI to voice tools like Wispr Flow, then to thought, and whether ambient voice computing turns us back into an oral society.
• Brain to brain: Sapir-Whorf, Russian shades of blue, Chinese counting, and what happens to the everyday performance of relationships when your thoughts are no longer private.
• Would you get one?: Downloading languages and whole books, a local open-source model wired to your mind, and why nobody should drill a hole in their skull to turn on the dishwasher.
• Augmentation and human potential: The many agents in your head, freestyle chess and Tyler Cowen’s Average Is Over, J.C.R. Licklider’s “Man-Computer Symbiosis,” and seeing beyond the visible spectrum, just as Tang dynasty painters and the Impressionists each saw the world anew.

Subscribe for more deep dives into technology and human progress.

(0:00): Opening(0:30): Visions of the future: Elon’s AI video, Mars and space elevators(1:45): Brain-computer interfaces and restoring lost function(4:00): The BCI landscape: Neuralink, Synchron, Merge Labs and Aleph Neuro(8:50): Dementia and the CNS as medicine’s final frontier(12:15): Early innings: trials, hardware and the Raspberry Pi moment(15:30): The next computing interface: buttons, voice and bi-directional BCI(19:45): Oral vs. written societies(21:50): Language, Sapir-Whorf and brain-to-brain communication(26:00): Would you get one?(29:30): What does learning mean when you can write to your brain?(34:00): The many selves in your head(37:30): Augmentation, freestyle chess and man-computer symbiosis(41:00): Rewiring old wounds, removing the governor and the da Vinci question(43:00): Expanding the senses: animal communication, Tang painters and the Impressionists(45:30): Outro

John and Steve ask: Has the cost of making a film effectively dropped to zero, and what does that do to storytelling? Is taste the one creative muscle AI still can't flex? And why might a robot folding your laundry matter more to the economy than the next interest-rate decision?

Topics:

• The bedroom auteur: A product designer working under the alias "Gossip Goblin" releases The Patchwrite, a 20-minute AI-generated cyberpunk short. What it actually took to make, why it's the opposite of "AI slop," and how the barrier to filmmaking just collapsed.
• Taste as the final frontier: Why AI remixes but doesn't create, why editorial judgment and curation still resist automation, and what the Coen brothers vs. improvisational directors tell us about making films with a machine.
• GPT Image 2 vs. Nano Banana: A real leap on the LM Arena leaderboard, Contra's attempt to benchmark actual creativity, and why "is this AI-generated?" is becoming a pointless question.
• Mythos & the cybersecurity arms race: What happens when AI can find zero-day exploits in the software that runs everything, why defense and offense stay in balance, and why we still don't know who Satoshi is.
• Robot roommates & the economy: Figure's humanoid demos, the "in-law or the robot?" test, what you'd actually pay for one, and why automation might be the real antidote to inflation and an aging population.
• Optimism vs. doom: Why the most useful take on AI sits between Jensen Huang's hope and Dario Amodei's warnings, and why the optimistic story is the one that rarely gets told.

Subscribe for more deep dives into technology and human progress.

(0:00): Opening
(0:50): Welcome & the Frontiers mission
(4:10): The AI narrative: optimism, doom & where the truth sits
(11:00): AI-generated film & the future of storytelling
(42:45): GPT image generation & the AI content debate
(49:45): Mythos, cybersecurity & online privacy
(1:01:30): Humanoid robots, automation & the economy

John and Steve ask: If a data engineer with no biology background can design a cancer vaccine for his dog using ChatGPT and AlphaFold, what does that say about the future of medicine, and who gets to do science? What are world models, and why do some think they're the missing piece for robotics? And why are our visions of the future always so dystopian?

Topics:

• World models explained: How they differ from large language models, why companies like World Labs and AMI Labs are betting billions on spatial intelligence, and what this means for the future of robotics
• The Rosie story: A data engineer in Sydney uses AlphaFold and ChatGPT to design a custom mRNA cancer vaccine for his dog, and it works. What he actually did, and what it means for personalized medicine
• The n-of-1 problem: If individualized cancer vaccines are trivially easy to make, why can't terminally ill humans try them? The case for rethinking regulatory frameworks
• Stated vs. revealed preferences: Why LLMs are trained on what we say, not what we do, and why your relationship advice from Claude might be Reddit's worst takes at scale
• Dystopian defaults: From Black Mirror to Brave New World, why our stories about the future are always grim, and Peter Diamandis's XPRIZE to fund optimistic storytelling
• The Amex lounge aesthetic: How we went from 1950s color maxing to fifty shades of gray, and a call for more optimism in art, design, and culture

Subscribe for more deep dives into technology and human progress.

(0:00): Opening

(4:26): World models, spatial intelligence, and the future of robotics

(24:00): AI and the job market, the lump of labor fallacy

(28:00): Life vlogging, data collection, and the GoPro farms

(33:00): Revealed vs. stated preferences and LLM training data

(40:00): Data curation, filter bubbles, and Bernie Sanders vs. Claude

(46:00): One man and his dog: the Rosie mRNA cancer vaccine story

(55:00): AI for science and man-computer symbiosis

(59:20): The case for utopian sci-fi and Peter Diamandis's XPRIZE

(1:07:00): Storytelling, aesthetics, and the call for color maxing

(1:10:50): Outro

John and Steve ask: Are we in a paradigm shift where no one writes code anymore — or is that just hype? What happens when you can reprogram aging at the cellular level? And if AI agents can run your life in the background, what's left for humans to actually do?

Topics:

• The coding agent revolution: Claude Code, Opus 4.5, and whether engineers are becoming system architects or just prompt engineers
• Vibe coding reality check: Why non-technical people are shipping apps, and the era of personalized computing
• Hyper-personalization: From algorithmically fractured media diets to genetics-informed nutrition via your phone
• GLP-1s and the longevity frontier: Why these molecules do more than suppress appetite, and the case for leapfrogging the Anglosphere back to baseline health
• David Sinclair's ER-100: The information theory of aging, Yamanaka factors, and why the eye is the least reckless place to start reprogramming cells
• The clinical trials bottleneck: We don't need more molecules — we need faster trials. Why AI in drug discovery is mostly hype
• AI personal assistants: ClawBot, ambient agents, and the future where your phone does everything while you focus on what matters

Subscribe for more deep dives into technology and human progress.

(0:00): Opening

(0:50): The coding agent shift — interviews, Claude Code & Opus 4.5

(9:00): Vibe coding and personalized computing

(12:00): Hyper-personalization and the fragmentation of culture

(15:00): Health data, genetics, and the future of nutrition

(21:00): Getting back to baseline — longevity and GLP-1s

(25:00): The information theory of aging and GLP-1 deep dive

(39:00): David Sinclair's ER-100 gene therapy

(50:00): Clinical trials, the bottleneck problem, and AI in drug discovery

(1:13:00): AI personal assistants and ambient agents

John and Steve ask: Why do pessimists always sound smart but optimists make money? How did one chemical process (Haber-Bosch) enable half of humanity to exist? Why is playing whack-a-mole with progress actually a good sign?

Topics:

• Solutionism: William Crooks, Paul Ehrlich & the false dichotomy of optimism vs. pessimism
• The Haber-Bosch process: Why a third of humanity owes their existence to ammonia synthesis
• France's nuclear triumph: When electricity prices dropped to zero and civilizational capacity pays off
• AI coding reality check: Hundreds of hours of vibe coding, the bug whack-a-mole problem, and what actually works in finance
• Safety as progress: From Underwriters Laboratories to the FDA, why solutions create new problems
• Why doesn't this exist yet: Colic research, liquid biopsies, and the false positive cascade in cancer screening

Subscribe for more deep dives into technology and human progress.

(0:00): Opening

(2:30): What happened this week

(2:30): GPT 5.2

(7:20): Specialisation of AI Models

(12:30): AI Enterprise Use Cases

(20:50): Maximising LLM value

(24:00): Free Energy in France

(31:00): Crawford's Techno-Humanist Manifesto - Solutionism, Part 1

(55:00): Solutionism, Part 2

(1:12:00): Why doesn't this exist yet?

John and Steve ask: If modernity gives us more choice than any generation in history, why do so many people feel worse off? What does a “good life” actually look like in 2024?

Topics:

• AI scientists & the future of discovery: Can tools like Cosmo (Edison/Future House) remake how science is done?
• Wealth taxes, billionaire flight & network effects: Will New York and London actually depopulate, or are they too “sticky” to leave?
• Jason Crawford’s “good life”: Beyond GDP and happiness surveys — what values should progress really be judged on?
• Work, love, knowledge & beauty: From bullshit jobs and trad-farm aesthetics to dating apps, Duolingo and AI art.
• Modernity as a bundle: Supermarket abundance, Ozempic, distraction-free tech and “protect me from myself” products.
• Millennials, housing & fertility: Are we genuinely worse off than our grandparents despite a century of progress?
• Choice, agency & templates: Why some people thrive on infinite options — and why others need structured paths.
• Why doesn’t this exist: A cognitive fitness tracker and a minimalist Spotify “iPod” for deep work.

Subscribe for weekly deep dives into technology, philosophy, and human progress.

(0:00): Intro

(2:01): NYC elections & exodus of wealth

(7:21): AI, Edison scientific, & equalising the scientific playing field

(18:00): Jason Crawford and defining the ‘good life’

(24:11): Freedom and autonomy as the core of the good life

(26:51): Core values of the good life

(35:21): Progress has afforded us a bundle of choice

(39:30): Progress is relative to the periods of comparison

(52:52): Why doesn’t this exist?

(57:20): Wrap-up

John and Steve explore: Are humans a cancer on Earth, or the most remarkable species? What future are we building with AI?

Topics:

• Sora vs Thinking: OpenAI's slop engine vs Anthropic's manifesto — who has the mandate of heaven?
• The death of humanism: When did "human extinction might be good actually" become respectable?
• Technology as human nature: Why the iPhone is as natural as the spear
• Malthusians were wrong (again): From famine predictions to obesity epidemics
• Agency vs inevitability: Are we passive observers or active creators of the future?
• The responsibility of builders: Every choice about what to build is a moral choice

Subscribe for weekly deep dives into technology, philosophy, and human progress.

(00:00) Introduction

(00:31) Welcome to Episode 2

(01:11) OpenAI integrations & the last-mile problem

(03:00) Commoditization hell & frontier model economics

(04:52) Concrete examples: Spotify, Zillow, Expedia

(06:40) LLM personalities & brand loyalty

(09:36) The prompt window & memory

(10:40) Sora 2 launch & AI-generated slop

(11:49) Sam's avatar & CCTV memes

(13:40) Meta's Vibes vs Sora's social features

(15:49) Entertainment as slop

(18:31) Anthropic's "Keep Thinking" campaign

(19:54) Say no to slop — the mandate of heaven

(21:00) Machines of Loving Grace manifesto

(22:20) Think Different redux

(22:42) AI encroaching on human territory

(26:00) Honda's Power of Dreams

(27:02) Jason Crawford's "Ode to Man"

(28:00) The death of humanism

(29:20) When did anti-natalism become mainstream?

(30:28) Climate change & anti-human sentiment

(31:20) Malthusians & the fertility trap

(32:20) Haber-Bosch & human ingenuity

(33:00) Progress as stress vs progress as solution

(34:20) Obesity, not famine: The Malthusian surprise

(36:20) Technology as expression of human nature

(37:40) The cumulative human project

(38:20) Brutality without technology

(39:00) Cloud seeding & playing God

(40:04) Backbreaking work & technological relief

(41:16) Lapvona & life before modernity

(42:00) Egyptian agriculture & cosmic theology

(43:40) Controlling our environment

(44:00) Agency vs fatalism

(45:20) Doom vs E/acc vs Crawford's third way

(45:45) How much agency do we really have?

(46:20) Capitalism demands progress

(47:20) Agency can be taught

(48:20) Free will & self-fulfilling prophecy

(48:40) Progress isn't linear: The medieval dark ages

(50:00) Agency is accessible to everyone

(50:38) The Sovereign Individual

(51:32) Who decides what's responsible?

(52:20) Decision funnels & the Network State

(53:40) The Museum of Passion Projects

(55:00) Moral obligation of builders

(56:00) Why doesn't this exist: Neurodegenerative diseases

(57:29) Alzheimer's & personal genetics

(58:20) The 2006 amyloid hypothesis fraud

(59:40) Academic incentives gone wrong

(01:00:27) The man who can't get Alzheimer's

(01:01:20) Gene editing vs mRNA vaccines

(01:02:20) Food allergies & pollen misery

(01:03:40) He Jiankui & the CRISPR babies

(01:05:00) mRNA as the better path

(01:06:00) Pandora's Box & AGI parallels

(01:06:58) Why doesn't this exist yet: AI to-do list app

(01:08:00) Minimalist phone apps

(01:09:00) ChatGPT forgets my weekend tasks

(01:11:00) Proactive AI & location awareness

(01:11:32) Accountability & push notifications

(01:12:40) CBT: Everything is your fault

(01:13:20) Memes vs productivity nudges

(01:13:40) AI doesn't know what time it is

(01:14:20) Wrap-up & Crawford's thesis

John and Steve kick off by asking: What are the real frontiers of progress, and what’s holding us back?

Topics:

• Apple vs Google

• E/ACC vs Philosophy of Progress: Is acceleration always good, or should tech serve human flourishing?

• Why innovation thrives in extremes (war, prosperity) and stalls in the middle.

Subscribe for deep dives into technology, society, and the future.

(00:00) Introduction
(01:00) Life-fit vs self-fit and Charlie Munger’s “inner scorecard”
(03:00) the “Great Lock-In,” and escaping the permanent underclass
(05:00) Apple vs Google — stagnation, innovation, and Banana by Google
(07:00) Product plateau and the equilibrium state of hardware
(10:00) Apple as luxury brand; design vs technology
(12:00) Enduring design
(14:00) Flip phones vs. brain interfaces
(16:00) Dangers of AI autonomy
(17:00) “Techno-Humanist Manifesto”
(18:00) 1960s optimism
(20:00) Techno-humanism vs effective accelerationism (E/acc)
(23:00) Doom vs E/acc vs Crawford
(26:00) Human well-being as the metric for progress
(29:00) Luddites vs. automation
(31:00) Taking modern infrastructure for granted
(33:00) Automation & translation

(36:00) Why physical tech lags behind software
(40:00) Defense tech as the engine of innovation
(44:00) Conflict as a driver of technological progress
(47:00) Forgetting pre-modern misery
(49:00) Bridging the global development gap
(50:00) Modernity as a bundle
(52:00) Slopware vs. real innovation
(54:00) Hard science vs SaaS
(57:00) Funding moonshots
(59:00) Oddball geniuses
(01:01:00) Anglo-Celtic inventiveness and prosperity cycles
(01:05:00) Ingenuity in prosperity