Two hosts explore where the next decade’s biggest AI-driven opportunities lie, arguing that the highest-leverage applications will move AI out of pure software (“the world of bits”) into the physical world (“the world of things”), and they map eight concrete industries where that shift is already beginning.
The core thesis: AI in the world of things, not bits
The hosts contend that applying AI to digital products — content, SaaS, coding agents — is becoming a race to the bottom with diminishing returns.
The real opportunity is using AI to manipulate atoms: biology, manufacturing, construction, defense, energy, and the human body itself.
They frame this as “skating to where the puck is going” — five years ago the puck was foundation models; today it is physical-world deployment.
1. Drugs and miracles: AI-native biotech and personalized medicine
Traditional drug development takes ~10 years, ~$2B, and has an 80–90% failure rate, limiting participation to massive incumbents.
AI collapses that loop: sequencing tumors, simulating drug candidates, synthesizing compounds, and iterating in months instead of years.
Concrete example: a GitLab executive used AI to design and take a custom therapy that shrank his own tumors; he now offers the service to others.
Eli Lilly is buying compute like a tech company; Google’s Isomorphic Labs (Demis Hassabis) is going all-in on AI drug design.
The hosts cite their friend Ben, a producer who accessed cutting-edge AI-guided treatment and showed dramatic improvement, as living proof the window is open now.
2. Food is technology: CRISPR crops, automated ag, and novel foods
Food production is fundamentally genetic engineering — every crop variety is a “genetic being” that can be optimized.
YC company Nitro Duck (stealth) uses CRISPR to create higher-yield, better-tasting, more nutritious fruits; the Honeycrisp apple took 20 years to breed — AI can compress that cycle.
David Friedberg (The Production Board) is building better seeds and crops for yield, flavor, and nutrition.
Adjacent opportunity: computer-vision-guided field robots, autonomous tractors, and fully automated greenhouses.
3. Warfare: drones, autonomy, and directed-energy weapons
The mental model of tanks, ships, and fighter jets is outdated; modern conflict is drones fighting drones — cheap, mass-produced, autonomous.
Ukraine is the proving ground: inexpensive loitering munitions defeat multi-million-dollar platforms.
Anduril (Palmer Luckey) leads in autonomous systems; Joe Lonsale’s company (Saronic) builds the naval equivalent.
Directed-energy / sonic weapons already deployed: a 2016 raid in Venezuela reportedly used a “discombobulator” that incapacitated guards without killing them.
A San Francisco hardware hacker demonstrated a laser system that can shoot down drones at extreme range — built in a converted auto-body shop.
4. Building homes with AI and robots: radical cost reduction via factory automation
Housing is the largest household expense; mortgage rates ~7.25%, construction costs $600–800/sq ft in desirable areas.
Prefab, factory-built homes can deliver 20× cost reduction by applying iPhone-factory-level automation to construction.
Zuru Tech (Nick Bore and brother) prints/assembles a luxury home per day in a Chinese factory they capitalized with $1B of their own money — no outside investors.
Process: modular concrete components cast with extreme precision, seismic-tested on a custom shake table simulating the five largest earthquakes of the last 50 years, then shipped and assembled like Lego blocks.
Founders’ background: 20 years building highly automated toy factories (Lego/Hot Wheels precision) — Elon Musk has cited toy manufacturing as the best training for gigacasting car bodies.
5. New materials: computational material science and “alchemy”
Teams in SF are using AI to discover or synthesize entirely new materials — some claim to have produced gold from base elements, others target room-temperature superconductors, ultra-light alloys, etc.
The hosts acknowledge this is “above their pay grade” but flag it as a frontier where AI + physics simulation + robotic labs could unlock step-change material properties.
6. FSD for everything: autonomy beyond passenger cars
Tesla FSD already handles 90–95% of driving miles; the stack is now being ported to forklifts, window-cleaning robots, ship-hull scrapers, bridge-inspection drones, and general maintenance robots.
The thesis: the physical world is full of dull, dangerous, dirty maintenance tasks that go undone because labor is scarce — autonomy unlocks a “maintain the world” layer.
7. Weather and oceans: the next SpaceX-scale frontier
Will O’Brien (previous guest) highlighted ocean tech — sensors, autonomous vessels, seabed mining, climate intervention — as the analog to space tech 15 years ago.
Oceans cover 70% of the planet, are poorly monitored, and hold massive resource/energy/climate leverage; AI + cheap robotics makes them newly addressable.
8. Human body hardware: neural interfaces and bionic limbs
Neuralink (brain-computer interface) and companies like Atom Limbs (Tyler’s project) are building mind-controlled robotic prosthetics that match or exceed natural limb function.
Current standard of care for amputees is a “stiff piece of plastic”; the engineering exists to give veterans and accident victims dexterous, sensory-feedback limbs — the gap is cost reduction and distribution.
This is AI in the world of things where “the things are us.”
America’s structural tailwinds for a new industrial revolution
The hosts draw a parallel to 1880–1970: abundant natural resources (iron ore via Great Lakes, oil, farmland), tariff-free internal market, WWII intact infrastructure, and a culture of individual liberty + capitalism created a compounding flywheel.
Today’s tailwinds: dominant data-center buildout, deepest AI talent pool, entrepreneurial culture, and a continent-scale resource base.
Risk factors: political polarization (socialist drift in major metros vs. anti-institutional right), hollowed-out manufacturing base making reshoring hard, and AI skepticism (either “takes all jobs” or “it’s a bubble”).
They long for a “common sense” political brand that can claim the moral high ground (like “America First” or “Make America Great Again” did) while advocating pro-growth, pro-technology, pro-affordability policies.
Career strategy for high-agency people: think like an investor, not a worker
The hosts’ cousin (smart, recent exit) was defaulting to “what jobs exist near me?” — a narrow peephole view.
Better exercise: write down the five companies you would invest $10M in if you had to keep all profits; that forces an investor’s “where is the puck going?” lens.
Mark Pincus’ “right body of water” metaphor: picking the expanding market (social media in 2003, AI-in-the-physical-world now) matters more than being the best operator in a stagnant one.
The eight industries above are candidate bodies of water — each offers impact, novelty, and financial upside simultaneously.
