Building easy AI apps is a bad idea. Here's 8 better bets
Audio Brief
Show transcript
Using artificial intelligence to make digital content and subscription software is now a race to the bottom, according to Shaan Puri. In this episode released on November 9, 2026, hosts Shaan Puri and Sam Parr discuss where the biggest business and technological opportunities will emerge over the next 5 to 10 years. By the end, you will know which physical industries are ready for an AI takeover and how to reposition your career into them.
The biggest returns are moving from the world of digital bits to the physical world of things. Anthropic created two trillion dollars of value in about five years, yet Puri argues that applying AI to digital bits, like content creation and software as a service tools, now yields very low returns.
This physical shift is already helping patients bypass the traditional pharmaceutical pipeline. Citing a tweet by Jeff Tang, Puri notes that the founder of GitLab sequenced his own tumor, tested and synthesized drugs, and shrank his tumors, while about seven others are doing the same do it yourself process.
Biology is also turning food production into programmable technology. Puri recalls a Y Combinator startup called Nitroduck that bypasses this by modifying fruit genetics through pollen grains rather than stem cells.
The same push toward cheap, automated hardware is reshaping modern warfare. Cautioning that he is speaking out of turn without direct war front knowledge, Puri observes that conflict in places like Ukraine is shifting from expensive tanks and jets to cheap, autonomous drones.
Back on the home front, toy manufacturing precision is coming to robotic home building. Puri points to his friend Nick Mowbray and his brother at Zuru Tech, who spent 20 years running automated Chinese toy factories before investing one billion dollars of their own money into a factory making one modular luxury home a day.
Beyond housing, AI and hardware breakthroughs are spreading across heavy industries. That same self-driving tech is now automating warehouse forklifts, skyscraper window cleaners, ship barnacle cleaners, and bridge maintenance, while Puri's friend Tyler at Atom Limbs is engineering down the cost of mind-controlled robotic arms to replace stiff plastic prosthetics.
Powering all of these machines requires a new American industrial revolution built on data centers. Citing history books, Parr explains that America's massive growth between 1880 and 1970 relied on capitalism, natural resources, and shipping enough iron ore on the Great Lakes by the 1950s to build 50 million cars a year. Parr argues that today's massive wave of US data center construction outpaces almost every other country and lays the groundwork for a comparable revolution.
To catch these waves, stop choosing your next job through the narrow peephole of your past resume. Instead, write down the five sectors where you would invest 10 million dollars today if you kept all the profits, and reposition yourself there. I think AI in the world of things is where the puck is going over the next 10 years, that's where a lot of the opportunity is going to go.
Episode Overview
- Shaan Puri and Sam Parr explore where the biggest business and technological opportunities will emerge over the next 5 to 10 years as AI makes traditional digital software and content creation hyper-competitive.
- While foundation models like Anthropic have generated immense digital wealth, the hosts argue that building AI for software, SaaS, and digital content is now a low-return race to the bottom.
- The true next frontier is applying AI to the physical world across eight specific sectors: pharmaceuticals, agriculture, warfare, robotic home building, material science, autonomous machinery, ocean technology, and human prosthetics.
- The conversation connects these technological shifts to the historical foundations of American industrial dominance, comparing the current boom in US data center construction to the natural resource explosions of the late 19th and mid-20th centuries.
- Note that Puri acknowledges his presentation of these physical AI sectors is a high-level overview rather than deeply researched analysis.
Key Concepts
- Shifting from the World of Bits to the World of Things: Anthropic has created $2 trillion of value in about 5 years, with its 7 co-founders owning roughly 14% collectively (or about 2% to 2.5% each). However, Puri argues that applying AI to digital bits, such as content creation and SaaS tools, now yields very low returns. Just as the "Internet of Things" was pushed by O'Reilly Media and others about 25 years too early, "AI in the world of things" is finally arriving and represents the real opportunity for the next decade.
- Bypassing the Pharmaceutical Pipeline with AI: Historically, developing a drug required 10 years and $2 billion, with an 80% or 90% failure rate. Today, companies like Eli Lilly are buying compute like tech firms or finance companies such as Citadel and Jane Street, and DeepMind founder Demis Hassabis is focusing on Google's Isomorphic Labs to solve major health issues. Individuals are also bypassing the traditional system entirely. Citing a tweet by Jeff Tang, Puri notes that the founder of GitLab sequenced his own tumor, ran through possible drugs, synthesized and compounded them, and shrank his tumors, while about 7 other people are doing the same DIY process (including individuals designing stage 4 lung cancer vaccines for $100k out of pocket, targeting rare liver and breast cancers, and shrinking dog mast cell tumors). Parr notes that a friend signed up for the GitLab founder's service and is doing great, and both hosts share that their friend and former producer Ben Wilson saw dramatic cancer improvements after AI enthusiasts connected him with experimental drugs.
- Accelerating Agricultural Genetics: Food production is being treated as programmable technology. While Parr notes it took 20 years to invent the Honeycrisp apple so it could be available 365 days a year, new biotech is compressing crop timelines that normally take 8 to 16 years. Puri highlights a Y Combinator company he recalls as Nitroduck, which modifies fruit genetics via pollen grains rather than stem cells (overcoming a limitation where traditional CRISPR only works on about 15% of crops), alongside entrepreneur David Friedberg's work on engineering higher-yield, more nutritious potatoes and corn.
- Automated Warfare and Directed Energy: Although Puri cautions that he is speaking out of turn without direct knowledge of war fronts, he observes that modern warfare in places like Ukraine is shifting from expensive tanks, boats, and fighter jets to high-volume, cheap, autonomous drones. This shift includes land defense companies like Anduril, a Joe Lonsdale company building an underwater "Navy version of Anduril," and a hacker Puri met at 11:00 PM in a friend's San Francisco hardware space building lasers to shoot down high-altitude drones. Furthermore, Parr cites news and eyewitness reports from the January 13, 2016 raid that captured Venezuela's leader, where US forces reportedly used a mysterious sonic or directed energy weapon—which Donald Trump called a "discombobulator"—that caused guards to feel like they were exploding, bleed from the nose, vomit, and collapse, resulting in what Parr believes was zero or close to zero deaths.
- Applying Toy-Manufacturing Precision to Robotic Home Building: With mortgage rates at 7.25% and local construction costs estimated by Puri at roughly $600 to $800 per square foot, automated factories are stepping in to build high-end prefab homes at 20 times cheaper the cost. Puri points to his friend Nick Mowbray and his brother, founders of Zuru Tech, who spent 20 years running automated toy factories in China before investing $1 billion of their own capital (and $0 in external funding) into a Chinese factory that manufactures 1 modular luxury home a day, complete with 100 prototype homes outside for workers and an indoor earthquake simulator that tests 5,000-square-foot structures against the 5 biggest earthquakes of the last 50 years. Following the LA wildfires, the Mowbrays bought roughly $60 million to $80 million worth of burned lots (with on-screen sources noting $65 million across 9 to 11 lots) to assemble $20 million, $30 million, and $40 million modular homes built in China. This mirrors Elon Musk's manufacturing philosophy: Musk admires Lego most for its precision at scale and once used a 2-inch Hot Wheels car in a meeting to inspire replacing a 200-piece Tesla underbody with 1 single injected mold using custom giant molding machines, a feat no other automaker matched for about 7 years.
- Physical Autonomy, Material Science, and Human Hardware: AI and hardware breakthroughs are spreading across multiple heavy industries. Puri admits material science is way above his pay grade, but notes San Francisco researchers are inventing new materials and claiming to make gold from other elements. In autonomy, Puri says his Tesla already drives 90% to 95% of his miles, and that same self-driving tech is now automating warehouse forklifts, skyscraper window cleaners, ship barnacle cleaners, and bridge maintenance. Meanwhile, ocean technology (discussed with past podcast guest Will O'Brien) is poised to follow the SpaceX boom, and Puri's friend Tyler at Atom Limbs is working on engineering down the cost of mind-controlled robotic arms to replace the stiff plastic prosthetics currently given to amputees and war veterans.
- Data Centers as the New American Industrial Revolution: Citing history books, Parr explains that across America's 250-year experiment, a massive growth explosion occurred between 1880 and 1970 due to individual liberty, private property, capitalism, 50 tariff-free states, and immense natural resources. In the 1880s and 1890s, Rockefeller's oil, Carnegie's steel, and Duluth, Minnesota's iron ore and copper mines allowed so much material to ship through the Great Lakes to New York that by the 1950s, the US transported enough iron ore on the Great Lakes to build 50 million cars a year. Combined with remaining intact after WWII destroyed Europe, North Africa, and Asia in the 1940s, these factors secured US dominance. Parr argues that today's massive wave of US data center construction outpaces almost every other country and is laying the groundwork for a comparable industrial revolution.
Quotes
- At 1:07 - "Do you realize that amongst the seven co-founders, I think they've been around for 2500 hour— 2500 days, they have basically created, collectively, the seven founders, $96 million of net worth every single day since founding." - Sam Parr illustrating the unprecedented speed of wealth creation at AI foundation model company Anthropic.
- At 3:01 - "I think that using AI in the world of, um, you know, bits, which is basically like digital stuff, I think has very low returns. I think using AI, anybody who's like, 'I'm using AI to make content,' it's a race to the bottom there. I think using AI to make, make software, SaaS, right, make tools, I think is, you know, kind of a race to the bottom." - Shaan Puri explaining why traditional digital software and content applications of AI are rapidly losing their competitive edge.
- At 15:01 - "If you've ever seen how like the iPhone factory works, where it's this incredibly automated factory in China where they're just spitting out these incredible devices that a billion people will use and it's like the highest quality manufacturing you can do, these guys are doing that, but they're making homes instead of iPhones." - Shaan Puri describing how Zuru Tech is applying high-precision Chinese factory automation to mass-produce luxury modular houses.
- At 25:43 - "I think that what we're missing now is good common sense leaders who can say things like, 'America's amazing, and we have problems that we should fix. Capitalism is amazing, and I understand why you guys are upset that housing is so co— uh, expensive, let's try to fix that.'" - Sam Parr discussing the need for pragmatic leadership that embraces technological and capitalist growth while solving real-world affordability bottlenecks.
- At 34:05 - "Why don't I put kind of my investor hat on and think like an investor of like, where is the best body of water that I should go be in? As Mark Pincus said on our podcast, like go find the right body of water. That's far more important than trying to be the best sailor in a shitty body of water." - Shaan Puri explaining his mental framework for choosing a career or business sector based on macro tailwinds rather than past experience.
- At 36:50 - "I think AI in the world of things is where the puck is going over the next 10 years, that's where a lot of the opportunity is going to go." - Shaan Puri summarizing the core thesis of where entrepreneurs and builders should focus their attention for the next decade.
Takeaways
- Run the "Investor Hat" career exercise: Stop choosing your next project or job based solely on the narrow peephole of your past resume, current industry, or local network. Instead, write down the 5 companies or sectors where you would invest $10 million today if you got to keep all the profits, and use that list to reposition yourself into an industry with massive macro tailwinds, even if you have no prior experience in it.
- Pivot away from digital-only AI wrappers toward physical industries: Avoid the hyper-competitive race to the bottom in AI content generation and basic SaaS tools. Look for ways to apply automation, robotics, and compute to heavy, high-cost physical bottlenecks like construction, agriculture, specialized maintenance, and hardware manufacturing.
- Apply consumer-grade manufacturing precision to legacy industries: Study how high-volume consumer goods like Lego, Hot Wheels, and iPhones are manufactured, and look for opportunities to bring single-piece injection molding and automated assembly lines to fragmented, slow-moving sectors like custom home building or medical prosthetics.