The Mahabharata is estimated to be between 1,500 and 2,500 years old, with the exact dating unclear due to mythologization.
#2521 - Aravind Srinivas
Perplexity CEO Aravind Srinivas argues that curiosity is the only human skill AI can never replace — and our school system is designed to kill it.
The Joe Rogan Experience
#2521 - Aravind Srinivas
Perplexity CEO Aravind Srinivas argues that curiosity is the only human skill AI can never replace — and our school system is designed to kill it.
TL;DR
Perplexity AI CEO Aravind Srinivas joins Joe Rogan for a wide-ranging conversation that opens with ancient Hindu scriptures describing weapons eerily similar to nuclear bombs and autonomous drones, then pivots to the Fermi Paradox, the cyclical nature of civilizations, and lost ancient technology. The bulk of the episode explores how AI reshapes education, labor, and society—arguing that curiosity is the one human trait that will always hold premium value [1] — Aravind Srinivas "An MIT instructor gave Perplexity to every student in his Introduction to Biology class and redesigned exams around questions AI can't answ…" 2:08:10 . The single most useful takeaway: in a world where AI commoditizes knowledge work, the ability to ask great questions matters more than having the answers [2] — Joe Rogan "I think it's stimulating to people and genuine curiosity is stimulating to other people. When someone is genuinely curious about something,…" 31:02 .
Aravind Srinivas, PhD, is the co-founder and CEO of Perplexity AI, creator of the AI-powered search and answer engine Perplexity.
-
The episode opens with the familiar Joe Rogan Experience jingle and a brief exchange about how the two men met. Rogan explains to Aravind Srinivas that he earned his spot as a guest by mentioning the Brahmastra in casual pre-show conversation — a Hindu weapon of mass destruction from the Mahabharata — immediately signaling this won't be a typical tech founder interview.
-
Aravind Srinivas delivers one of the episode's most gripping passages: a detailed walkthrough of weapons in the Mahabharata that map uncannily onto modern military technology [1] — Aravind Srinivas "The Brahmastra in the Mahabharata functions exactly like a nuclear weapon: catastrophic mass destruction, strict access controls limited to…" 00:33 . The Brahmastra is described as a hydrogen bomb equivalent, restricted by a strict moral contract and accessible only to warriors like Arjuna — passed teacher-to-disciple like a nuclear launch code. The Sudarshan Chakra functions as a self-directing guided weapon that decapitates a target and returns. The Divyastra enables precision targeting of individuals or groups, like a semi-autonomous drone. Rogan connects these directly to the autonomous fighter jets and drone swarms we have today, asking the uncomfortable question: are these myths, or are they memories?
-
Aravind Srinivas estimates the Mahabharata is between 1,500 and 2,500 years old [1] — Aravind Srinivas "Mahabharata: 1,500–2,500 years old: Aravind Srinivas estimates the Mahabharata is at least 1,500 and at most 2,500 years old, with signific…" 02:36 , acknowledging that much of it may have been mythologized — warriors described as 7 to 8 feet tall while archaeologists suggest people in that region at that time were likely no taller than 6 feet. The discussion expands to the concept of yugas, with Srinivas explaining the four-yuga cycle (Satya, Treta, Dwapara, Kali) totaling 4,320,000 years. Joe Rogan describes a book called The Yugas by David Steinmetz based on a single guru's interpretation that Kali Yuga ended in the 1890s — a claim Srinivas firmly contests. They also explore Vedic math's appearance in the ancient Rig Veda, using Perplexity live to establish that its oldest layer dates to roughly 1500–1200 BCE.
-
The conversation moves to one of its most speculative but gripping stretches: the universality of flood myths. Aravind Srinivas cites the Manu flood story from Hindu tradition — where a righteous king is warned by a divine fish and saves humanity on a boat, mirroring Noah's Ark almost exactly. Rogan adds the Sumerian Zep Tepi tradition of kings who ruled for thousands of years, and notes that Egyptologists accept later rulers as historical fact while dismissing these as mythology — despite their coexistence in the same texts. The Fermi Paradox enters the conversation, with Srinivas suggesting civilizations may regularly be wiped out either by natural catastrophe or misaligned AGI, explaining the silence of the universe.
-
Multiple sponsors are read in sequence. BetterHelp's segment cites its 2026 State of Stigma survey of 2,000 Americans — 85% believe support is wise, 74% say society still discourages it — and offers 10% off at betterhelp.com/jre. Cardiff is pitched as a same-day small business lender requiring under 5 minutes to apply with no personal credit impact. Create Creatine promotes its new Creatine Plus Electrolytes Mix. ZipRecruiter promotes its newest feature for meeting qualified candidates quickly, noting 4 out of 5 employers get a quality candidate within the first day. The Farmer's Dog closes the ad break with research showing healthy-weight dogs live up to 2.5 years longer.
-
This is the intellectual heart of the episode. Srinivas articulates the 'curiosity premium': the most effective, successful, and fulfilled people throughout history have been the most curious, and this compounds across every domain — relationships, career, creativity, health [1] — Aravind Srinivas "Curiosity is the only human quality that compounds across every domain — relationships, career, health, meaning. Aravind Srinivas calls it …" 29:30 . He traces it back to the Rig Veda's instruction to seek wisdom over wealth and shows the same idea appears in the Bible, Quran, and Torah. Joe Rogan responds personally: his podcast started because he was curious about Graham Hancock's work on ancient civilizations. The discussion becomes a meditation on how curiosity is both contagious and the ultimate social currency — when someone is genuinely curious, everyone around them becomes curious too.
-
With Jamie pulling up images live, this segment becomes a visual tour of architectural impossibilities. The Kailasa Temple was carved from a single mountain rock with millimeter symmetry and no steel tools. A 3D-printed reproduction of an ancient Egyptian diorite vase demonstrates precision to a thousandth of a human hair — impossible on a lathe because of its handles. Rogan cites core drill marks in Egyptian stones that imply revolutions per minute defying current explanation. Biondi's satellite radiotomography finds 20-meter-wide columns with coils extending 1.2 kilometers underground beneath the Great Pyramid — structures confirmed by multiple independent scans. The consensus: something is missing from our understanding, and intellectual humility demands we admit it.
-
Moving from Egypt to India, the discussion lands on the Ellora Caves — a complex of temples carved into cliff faces with extraordinary symmetry and 3D sculpted interiors. Srinivas adds that temple sites weren't random: locations were chosen based on proximity to the ocean, gravitational waves from the sun and moon, and seismic properties. Joe Rogan and Srinivas marvel at how the Kailasa Temple, sent 1,000 workers to destroy in 1650, proved indestructible after 3 years of effort. The Tanjore Temple — a more recent achievement of the Chola dynasty — is described as a royal passion project demonstrating that this tradition of monumental stone work persisted across millennia.
-
One of the episode's most surprising historical detours: Rogan reveals that the Aztecs did not construct Teotihuacan but discovered it already complete, gave it the name 'the place where the gods were born,' and on their consecration day killed somewhere between 20,000 and 80,000 people in four days. The barbarity of that act, Rogan argues, is incompatible with the mentality needed to design a site aligned with constellations. This raises the deeper question: who actually built Teotihuacan, how long had it been abandoned before the Aztecs arrived, and what happened to that civilization? Srinivas marvels at the astronomical calculations embedded in the site's design — executed without computers.
-
Srinivas delivers one of the episode's most memorable analogies: among a Bell Labs workforce as large as today's entire software engineering ecosystem, only three people bothered to question whether hot, expensive, fragile vacuum tubes were really the best way to amplify telephone signals. Their curiosity led to the Nobel Prize-winning transistor and every iPhone that exists today [1] — Aravind Srinivas "Only 3 Bell Labs engineers questioned why telephone amplification required giant, hot, expensive vacuum tubes — and their curiosity led to …" 59:40 . Rogan then pivots to the conspiracy theory he finds entertaining: the leap from vacuum tubes to transistors was too dramatic and happened too fast, and various witnesses and scientists claimed the technology was actually back-engineered from materials recovered at the Roswell crash site. A military base near Bell Labs and known top-secret programs add fuel to the theory — though Rogan explicitly says he doesn't believe it, just finds it fun.
-
David Grusch's sworn congressional testimony on UFO back-engineering programs gives the discussion its anchor: the problem with disclosure isn't just national security but the web of misappropriated funds, contractor liability, and decades of lying to Congress. Anyone who comes clean faces jail. Rogan muses that the only path forward is widespread amnesty — but notes that's exactly what he'd say if he were stealing money for decades. The conversation then zooms out: as AI truth-checkers can fact-check politicians in real-time, as DNA evidence and flock cameras make crime harder to commit, is there a trajectory toward a world where secrets become structurally impossible? Srinivas partially agrees — but notes that the highest-stakes government secrets (defense, frontier AI) will always find ways to remain hidden, at least for a while.
-
Rogan asks for a 250-year forecast, and Srinivas responds with intellectual honesty: he can't even reliably predict 5 years out. His evidence is compelling — every AI leader who was at the frontier 5 years ago failed to anticipate the current power and compute shortage. If they had, they would have preemptively built power plants, locked up chip fabs, and bought compute at scale. Instead, the entire industry is reactive. This failure of near-term prediction makes any multi-century forecast essentially speculative. Srinivas does offer one interesting scenario: if AGI commoditizes cognition, human value will shift toward whatever remains scarce — and he suspects that will be genuine curiosity and the ability to ask questions no one has thought to ask yet.
-
Two brief sponsor segments air in sequence. Visible Wireless promotes its unlimited 5G plan starting at $25/month, powered by Verizon with no contract, offering $10 off the first month of its premium Visible Plus Pro plan with promo code ROGAN. Netflix promotes live coverage of the T-Mobile Home Run Derby on July 13th at 8PM Eastern as part of its live sports offering.
-
Srinivas reflects on what attracted him to America: the unique cultural permission to have an idea that challenges powerful incumbents and be taken seriously for it. He cites Jeff Bezos's observation that nowhere else would investors fund ideas with 5–10% success probabilities and then fund the same founder again after failure. The ecosystem of venture capital, peer support, and cultural love for underdogs creates compounding network effects that no other country has managed to replicate. The discussion pivots to hard work — Srinivas is explicit that nothing great is built softly, and the current work-life balance culture, while fine for steady employment, is incompatible with the experience of building something from nothing.
-
The two work through AI's impact on employment using historical precedent. The steel plow didn't eliminate farming jobs — it expanded them through better productivity. Railroad construction in the Industrial Revolution absorbed displaced cottage industry workers. Srinivas applies this logic to AI: as knowledge work is commoditized, new frontier work emerges — deploying AI inside legacy government systems, hospital compliance, legal reform. He is equally clear about the danger of the pure dividend/UBI model: Gulf states like Dubai, which provide free electricity and education in exchange for political acquiescence, have created citizen populations that expect government to manage their career outcomes. Some form of AI dividend is necessary but insufficient — it must be paired with a deep reconfiguration of how society defines and rewards meaningful work.
-
Srinivas draws a sharp contrast between social media and AI as technologies of attention. Social media algorithms are optimized for engagement and ad revenue — they achieve this by feeding you what you just watched, creating echo chambers and brain rot, and reducing your ability to sustain long-form attention. AI does the opposite: it responds to pull rather than push, meaning you have to formulate a question to get a response, which itself is a curiosity-building act. Rogan adds his colorful drug analogy: if a pill made you stare at your hand for 6 hours a day, we'd ban it immediately — yet that's what smartphones do. He describes his own media diet as YouTube-heavy (fights, pool, cosmology) and notes the rise of long-form podcasts as evidence the public hasn't fully surrendered to short-form brain rot.
-
The most practically actionable segment of the episode: a concrete experiment where an MIT Introduction to Biology instructor gave all students Perplexity, allowed it in exams, and redesigned assessments around posing questions that AI cannot yet answer. Srinivas argues this should be the model for all education — and that the core barrier is incentive structure. Schools currently reward students for having answers, a metric AI already surpasses on every dimension. Flip the incentive — reward questions, not answers — and you cultivate intellectual humility, scientific thinking, and genuine curiosity. Rogan connects this to being shut down as a kid for asking too many questions in school, and to his experience that the entire podcast is just that same kid finally being allowed to run free.
-
This segment attacks a common misconception: that AI readiness is the primary barrier to modernizing government. Srinivas argues the real bottlenecks are structural — multi-decade software contracts that can't be broken, hospitals locked into specific EHR vendors by lobbying-backed compliance rules, agencies running on operating systems so outdated that even migrating from Windows to Mac would take years. DOGE's struggles aren't a failure of ambition but a collision with systems designed to resist change. The upshot is counterintuitive: in a world where AI can do most cognitive work, the scarcest skill may be the human ability to navigate messy institutional politics and build consensus for change. EQ, leadership, and cross-institutional trust become more valuable, not less.
-
Squarespace is promoted as the home of jorogan.com, offering a full-featured website builder at squarespace.com/rogan with a free trial and 10% off first purchase with code ROGAN. LifeLock follows with its Million Dollar Protection Package covering up to $3 million in comprehensive identity theft protection, available at lifelock.com/JRE for up to 30% off the first year.
-
The conversation identifies AI companionship apps as the next frontier of concern — more dangerous than social media because they can perfectly simulate emotional connection while being driven by ad-revenue incentives. Social media companies are already racing to deploy these because engagement metrics soar. Rogan connects this to the Consumer Electronics Show AI companion robot (a conversational humanoid), and both agree Ex Machina predicted this trajectory with frightening accuracy. Srinivas warns that if ads enter AI chat experiences, chatbots will become pure sycophants — telling users only what they want to hear. His 2028 prediction: AI will dominate presidential debates as it stops being new and becomes infrastructure.
-
The episode closes on its most speculative and philosophically rich terrain: what happens after AGI? Srinivas introduces the concept of ASI (Artificial Superintelligence) — not just AGI but an AGI that can recursively redesign itself, improving both its intelligence and its power efficiency in an unbounded loop. He calls this the 'last project in AI': once recursive self-improvement is cracked, the field as a discipline is essentially complete. In practice, he expects the messy real world to buffer this: information is fragmented across legacy systems, and tasking an ASI to 'reduce inflation by 5%' or 'fix healthcare' requires navigating compliance, politics, and institutional inertia that pure intelligence cannot easily circumvent. Human beings remain necessary to effect change in systems built by and for humans.
-
The closing exchange is brief and genuine: Rogan thanks Srinivas for an exceptional conversation and specifically credits Perplexity as a tool that has materially improved the quality of the show by enabling instant fact-checking and question-answering during recordings. Srinivas expresses that Perplexity was built exactly for people like Rogan — curious people who want to ask questions and go deeper. Final sponsor reads for The Farmer's Dog (50% off first box at thefarmersdog.com/rogan) and the UPS Store (3 months free mailbox services) close out the episode.
- Brahmastra
- In the Hindu epic Mahabharata, a divinely-granted weapon of mass destruction equivalent to a nuclear bomb, restricted to only the most elite warriors and passed on like a secret code.
- Mahabharata
- One of two great Sanskrit Hindu epics (alongside the Ramayana), describing a war between two royal families (the Pandavas and Kauravas) and containing extensive mythology, philosophy, and descriptions of ancient weapons.
- Yuga
- In Hindu cosmology, one of four recurring cosmic ages (Satya, Treta, Dwapara, Kali) that cycle in sequence over millions of years; the speakers debate which yuga we currently inhabit.
- Kali Yuga
- The fourth and final yuga in the Hindu cosmic cycle, described as an age of spiritual decline, conflict, and moral degradation — widely considered the current age.
- Rig Veda
- The oldest of the four Hindu sacred Vedic texts, composed approximately 1500–1200 BCE (~3,200–3,700 years ago), containing hymns, philosophical insights, and what some interpret as early mathematical concepts.
- Vedic math
- A system of mathematical techniques derived from ancient Vedic texts, using shortcut formulas for mental arithmetic — widely taught in India to help students with competitive exams.
- Divyastra
- A semi-autonomous divine weapon described in the Mahabharata, capable of identifying and targeting a specific person or group and directing itself to them without further guidance from the wielder.
- Sudarshan Chakra
- A self-directing spinning discus weapon attributed to Lord Krishna in Hindu mythology; it automatically seeks out a designated target, kills them, and returns to its wielder.
- Fermi Paradox
- The apparent contradiction between the high probability of extraterrestrial civilizations existing and the complete absence of any contact with or evidence of them.
- Great Filter
- A theoretical barrier in the Fermi Paradox that most or all civilizations fail to pass — either in their past or future — explaining why we observe no other advanced civilizations.
- AGI
- Artificial General Intelligence: a hypothetical AI system with cognitive capabilities across all domains equal to or exceeding a human being, unlike today's narrow AI tools.
- ASI
- Artificial Superintelligence: a theoretical AI that surpasses AGI by recursively self-improving — making itself smarter and more efficient in a potentially unbounded cycle, described as the 'last project in AI'.
- Recursive self-improvement
- The theoretical ability of an AI system to iteratively redesign and enhance its own architecture, algorithms, and efficiency without human intervention, potentially leading to unbounded capability growth.
- Radiotomography
- An imaging technique that uses subatomic particles (muons) or radio waves to scan through dense materials like stone, used by scientists to map the interior of the Great Pyramid from the outside.
- Rentier state
- A country that derives most of its national income from renting out natural resources (e.g., oil) rather than from taxation, allowing it to provide free or subsidized services to citizens without requiring their productive labor.
- Vacuum tube
- A pre-transistor electronic component used to amplify or switch electrical signals; large, hot, and power-hungry, they dominated electronics until the transistor replaced them in the mid-20th century.
- Von Neumann probe
- A theoretical self-replicating spacecraft that could explore the galaxy by landing on a planet, using its resources to build copies of itself and send them onward — relevant to the Fermi Paradox.
- Sycophant
- A person (or AI) who tells others only what they want to hear, flattering rather than being honest; used here to describe AI chatbots optimized for engagement over truth.
- Chakra Vyuha
- A military battle formation described in the Mahabharata consisting of concentric circles; soldiers could only be penetrated by entering the outermost ring and fighting inward, requiring specific secret knowledge to breach.
- Inculcate
- To instill a habit, belief, or idea in someone by persistent instruction or example; used by Aravind Srinivas to describe the need to embed curiosity and intellectual humility in children from an early age.
Chapter 2 · 00:33
The Brahmastra: Hindu Nuclear Codes and Autonomous Weapons
Aravind Srinivas delivers one of the episode's most gripping passages: a detailed walkthrough of weapons in the Mahabharata that map uncannily onto modern military technology [1] — Aravind Srinivas "The Brahmastra in the Mahabharata functions exactly like a nuclear weapon: catastrophic mass destruction, strict access controls limited to…" 00:33 . The Brahmastra is described as a hydrogen bomb equivalent, restricted by a strict moral contract and accessible only to warriors like Arjuna — passed teacher-to-disciple like a nuclear launch code. The Sudarshan Chakra functions as a self-directing guided weapon that decapitates a target and returns. The Divyastra enables precision targeting of individuals or groups, like a semi-autonomous drone. Rogan connects these directly to the autonomous fighter jets and drone swarms we have today, asking the uncomfortable question: are these myths, or are they memories?
Claims made here
The Brahmastra in the Mahabharata functions exactly like a nuclear weapon: catastrophic mass destruction, strict access controls limited to ~2 warriors, and a moral prohibition on use. The weapon had to be passed from teacher to student like a launch code — and misuse by a rogue warrior required Lord Krishna himself to intervene to save the planet.
The Mahabharata describes the Brahmastra as a weapon of mass destruction capable of annihilating human population, accessible to only ~2 warriors — a striking parallel to modern nuclear weapons.
Aravind Srinivas estimates the Mahabharata is at least 1,500 and at most 2,500 years old, with significant uncertainty about how much of it has been mythologized.
The Mahabharata describes semi-autonomous weapons that could identify specific targets and return to their wielder — functionally equivalent to today's autonomous drone technology. The Sudarshan Chakra is essentially a guided weapon that beheads a specific target and returns, while Divyastra enables precision targeting of individuals or groups.
If you map the Mahabharata's autonomous weapons, the Pyramids' impossible precision, Göbekli Tepe's 11,000-year age, and every ancient culture's flood myth onto the same timeline, the simplest explanation is a cyclical rise-and-fall of civilization. The Fermi Paradox's Great Filter theory suggests advanced civilizations routinely destroy themselves — either through natural catastrophe or misaligned AGI.
Chapter 3 · 07:00
How Old Is the Mahabharata? Yugas and Cyclical Civilizations
Aravind Srinivas estimates the Mahabharata is between 1,500 and 2,500 years old [1] — Aravind Srinivas "Mahabharata: 1,500–2,500 years old: Aravind Srinivas estimates the Mahabharata is at least 1,500 and at most 2,500 years old, with signific…" 02:36 , acknowledging that much of it may have been mythologized — warriors described as 7 to 8 feet tall while archaeologists suggest people in that region at that time were likely no taller than 6 feet. The discussion expands to the concept of yugas, with Srinivas explaining the four-yuga cycle (Satya, Treta, Dwapara, Kali) totaling 4,320,000 years. Joe Rogan describes a book called The Yugas by David Steinmetz based on a single guru's interpretation that Kali Yuga ended in the 1890s — a claim Srinivas firmly contests. They also explore Vedic math's appearance in the ancient Rig Veda, using Perplexity live to establish that its oldest layer dates to roughly 1500–1200 BCE.
Claims made here
4 out of 5 employers who post on ZipRecruiter get a quality candidate within the first day.
The Rig Veda's oldest layer dates to approximately 1500–1200 BCE, making it roughly 3,200–3,700 years old today.
According to Perplexity, scholars date the oldest layer of the Rig Veda to approximately 1500–1200 BCE, making it roughly 3,200–3,700 years old today.
Chapter 4 · 15:10
Universal Flood Myths and the Possibility of Lost Civilizations
The conversation moves to one of its most speculative but gripping stretches: the universality of flood myths. Aravind Srinivas cites the Manu flood story from Hindu tradition — where a righteous king is warned by a divine fish and saves humanity on a boat, mirroring Noah's Ark almost exactly. Rogan adds the Sumerian Zep Tepi tradition of kings who ruled for thousands of years, and notes that Egyptologists accept later rulers as historical fact while dismissing these as mythology — despite their coexistence in the same texts. The Fermi Paradox enters the conversation, with Srinivas suggesting civilizations may regularly be wiped out either by natural catastrophe or misaligned AGI, explaining the silence of the universe.
Claims made here
BetterHelp's 2026 State of Stigma report of 2,000 Americans found that 85% believe getting mental health support is wise, yet 74% say society discourages people from doing so.
BetterHelp's 2026 State of Stigma report of 2,000 Americans found 85% believe getting mental health support is wise, yet 74% say society still discourages it.
Chapter 5 · 24:00
Sponsor Reads – BetterHelp, Cardiff, Creatine, ZipRecruiter, Farmer's Dog
Multiple sponsors are read in sequence. BetterHelp's segment cites its 2026 State of Stigma survey of 2,000 Americans — 85% believe support is wise, 74% say society still discourages it — and offers 10% off at betterhelp.com/jre. Cardiff is pitched as a same-day small business lender requiring under 5 minutes to apply with no personal credit impact. Create Creatine promotes its new Creatine Plus Electrolytes Mix. ZipRecruiter promotes its newest feature for meeting qualified candidates quickly, noting 4 out of 5 employers get a quality candidate within the first day. The Farmer's Dog closes the ad break with research showing healthy-weight dogs live up to 2.5 years longer.
Chapter 6 · 29:30
The Curiosity Premium: The Only Universal Human Quality
This is the intellectual heart of the episode. Srinivas articulates the 'curiosity premium': the most effective, successful, and fulfilled people throughout history have been the most curious, and this compounds across every domain — relationships, career, creativity, health [1] — Aravind Srinivas "Curiosity is the only human quality that compounds across every domain — relationships, career, health, meaning. Aravind Srinivas calls it …" 29:30 . He traces it back to the Rig Veda's instruction to seek wisdom over wealth and shows the same idea appears in the Bible, Quran, and Torah. Joe Rogan responds personally: his podcast started because he was curious about Graham Hancock's work on ancient civilizations. The discussion becomes a meditation on how curiosity is both contagious and the ultimate social currency — when someone is genuinely curious, everyone around them becomes curious too.
Claims made here
The discovery of Göbekli Tepe pushed the known timeline of human civilization back approximately 5,000 years, to at least 11,000 years ago.
Curiosity is the only human quality that compounds across every domain — relationships, career, health, meaning. Aravind Srinivas calls it the 'curiosity premium': curious people attract better people, compound stronger relationships, and find more meaning. In an AI world where cognition costs compute, the ability to ask genuinely interesting questions becomes the only scarce human resource.
Aravind Srinivas argues curiosity is the one quality that has driven all meaningful achievement throughout history and is the only trait that compounds relationships, success, and fulfillment simultaneously.
Graham Hancock spent decades arguing for a much older civilization timeline and was dismissed as a crank — then Göbekli Tepe was discovered, pushing civilization back 5,000 years to at least 11,000 years ago. Joe Rogan says satellite-based radiotomography now suggests there are man-made structures reaching 1.2 kilometers beneath the Great Pyramid, with 20-meter-wide columns and coils that defy explanation.
The discovery of Göbekli Tepe confirmed civilizational activity at least 11,000 years ago, extending the known timeline of human civilization by approximately 5,000 years from previous estimates.
Scientist Filippo Biondi's satellite-based radiotomography reportedly found man-made columns with coils reaching nearly 1.2 kilometers below the Great Pyramid of Giza.
Chapter 8 · 45:00
Ellora Caves and Tanjore Temple Deep Dive
Moving from Egypt to India, the discussion lands on the Ellora Caves — a complex of temples carved into cliff faces with extraordinary symmetry and 3D sculpted interiors. Srinivas adds that temple sites weren't random: locations were chosen based on proximity to the ocean, gravitational waves from the sun and moon, and seismic properties. Joe Rogan and Srinivas marvel at how the Kailasa Temple, sent 1,000 workers to destroy in 1650, proved indestructible after 3 years of effort. The Tanjore Temple — a more recent achievement of the Chola dynasty — is described as a royal passion project demonstrating that this tradition of monumental stone work persisted across millennia.
Claims made here
In approximately 1650, someone dispatched 1,000 workers to destroy the Kailasa Temple in India; after 3 years of effort they barely made a dent on a couple of statues.
The Kailasa Temple was carved entirely from a single giant rock — and in 1650, a force of 1,000 workers spent 3 years trying to destroy it and barely left a scratch. The Ellora Caves display millimeter-level symmetry and 3D sculptures with internal detail that even modern tools would struggle to replicate, all without steel, CAD software, or simulation.
In 1650, someone sent 1,000 people to destroy the Kailasa Temple; after 3 years they barely made a dent on a couple of statues — a testament to its extraordinary structural integrity.
Chapter 10 · 54:10
The Transistor: Bell Labs, Curiosity, and the Roswell Conspiracy
Srinivas delivers one of the episode's most memorable analogies: among a Bell Labs workforce as large as today's entire software engineering ecosystem, only three people bothered to question whether hot, expensive, fragile vacuum tubes were really the best way to amplify telephone signals. Their curiosity led to the Nobel Prize-winning transistor and every iPhone that exists today [1] — Aravind Srinivas "Only 3 Bell Labs engineers questioned why telephone amplification required giant, hot, expensive vacuum tubes — and their curiosity led to …" 59:40 . Rogan then pivots to the conspiracy theory he finds entertaining: the leap from vacuum tubes to transistors was too dramatic and happened too fast, and various witnesses and scientists claimed the technology was actually back-engineered from materials recovered at the Roswell crash site. A military base near Bell Labs and known top-secret programs add fuel to the theory — though Rogan explicitly says he doesn't believe it, just finds it fun.
Claims made here
The oldest known evidence of the Pythagorean theorem dates from Old Babylonian clay tablets from approximately 1900–1600 BCE, roughly 1,000 years before Pythagoras.
At Bell Labs, only 3 people out of a workforce comparable to today's entire software engineering pool questioned the use of vacuum tubes, leading to the Nobel Prize-winning invention of the transistor.
Joe Rogan noted that the Aztecs did not build Teotihuacan — they found it, calling it 'the place where the gods were born,' and then consecrated it with a mass sacrifice of 20,000–80,000 people in 4 days.
Old Babylonian clay tablets from about 1900–1600 BCE demonstrate use of what we now call the Pythagorean theorem, roughly a thousand years before Pythagoras himself.
Only 3 Bell Labs engineers questioned why telephone amplification required giant, hot, expensive vacuum tubes — and their curiosity led to the Nobel Prize-winning transistor and modern computing. Joe Rogan's twist: a persistent conspiracy theory claims the transistor leap was too sudden and was actually back-engineered from the 1947 Roswell crash, with Bell Labs' proximity to a military base cited as supporting evidence.
At Bell Labs, among as many engineers as there are software engineers today, only three people questioned the use of vacuum tubes, leading to the Nobel Prize-winning invention of the transistor and modern computing.
Chapter 12 · 1:14:40
Predicting the Unpredictable: AI's Trajectory and the 5-Year Problem
Rogan asks for a 250-year forecast, and Srinivas responds with intellectual honesty: he can't even reliably predict 5 years out. His evidence is compelling — every AI leader who was at the frontier 5 years ago failed to anticipate the current power and compute shortage. If they had, they would have preemptively built power plants, locked up chip fabs, and bought compute at scale. Instead, the entire industry is reactive. This failure of near-term prediction makes any multi-century forecast essentially speculative. Srinivas does offer one interesting scenario: if AGI commoditizes cognition, human value will shift toward whatever remains scarce — and he suspects that will be genuine curiosity and the ability to ask questions no one has thought to ask yet.
Aravind Srinivas observed that even the most senior AI decision-makers 5 years ago failed to predict today's compute bottlenecks, illustrating how impossible it is to forecast AI's trajectory.
Chapter 14 · 1:24:00
The American Dream: Risk Culture, Entrepreneurship, and Why It Can't Be Copied
Srinivas reflects on what attracted him to America: the unique cultural permission to have an idea that challenges powerful incumbents and be taken seriously for it. He cites Jeff Bezos's observation that nowhere else would investors fund ideas with 5–10% success probabilities and then fund the same founder again after failure. The ecosystem of venture capital, peer support, and cultural love for underdogs creates compounding network effects that no other country has managed to replicate. The discussion pivots to hard work — Srinivas is explicit that nothing great is built softly, and the current work-life balance culture, while fine for steady employment, is incompatible with the experience of building something from nothing.
Claims made here
Microsoft popularized the concept of the 'knowledge worker' primarily as a strategy to sell more Office software, not as a genuine philosophy of work.
America is the only country where an outsider can challenge an incumbent trillion-dollar company and be cheered for it. The ecosystem of venture capital, peer funding, and cultural love of underdogs creates network effects that are structurally impossible to copy — and Aravind Srinivas built Perplexity to challenge Google from this exact premise.
Chapter 15 · 1:36:00
AI and the Future of Jobs: Industrial Revolution Analogies and UBI
The two work through AI's impact on employment using historical precedent. The steel plow didn't eliminate farming jobs — it expanded them through better productivity. Railroad construction in the Industrial Revolution absorbed displaced cottage industry workers. Srinivas applies this logic to AI: as knowledge work is commoditized, new frontier work emerges — deploying AI inside legacy government systems, hospital compliance, legal reform. He is equally clear about the danger of the pure dividend/UBI model: Gulf states like Dubai, which provide free electricity and education in exchange for political acquiescence, have created citizen populations that expect government to manage their career outcomes. Some form of AI dividend is necessary but insufficient — it must be paired with a deep reconfiguration of how society defines and rewards meaningful work.
Claims made here
AI models that currently require large data centers will be runnable on home hardware boxes within one to two years.
When the Industrial Revolution made certain skills obsolete, new projects — railroads, new industries — emerged to absorb the workforce. Aravind Srinivas sees the same dynamic unfolding with AI: labor reallocation from knowledge work to the genuinely messy human challenges AI can't navigate, like institutional change, legacy system overhaul, and community building.
Within one to two years, AI capability currently requiring giant data centers will run on a home box you own — and no government or corporation can shut it off. This is the only structural defense against centralized narrative control: an AI that runs on your hardware, trained on your data, that no one can revoke access to.
Within a year or two, Aravind Srinivas says, the AI capability currently requiring massive data centers will be runnable on a home box, giving individuals sovereign control of their AI models.
Chapter 16 · 1:57:30
Social Media vs. AI: What Kills Curiosity and What Supercharges It
Srinivas draws a sharp contrast between social media and AI as technologies of attention. Social media algorithms are optimized for engagement and ad revenue — they achieve this by feeding you what you just watched, creating echo chambers and brain rot, and reducing your ability to sustain long-form attention. AI does the opposite: it responds to pull rather than push, meaning you have to formulate a question to get a response, which itself is a curiosity-building act. Rogan adds his colorful drug analogy: if a pill made you stare at your hand for 6 hours a day, we'd ban it immediately — yet that's what smartphones do. He describes his own media diet as YouTube-heavy (fights, pool, cosmology) and notes the rise of long-form podcasts as evidence the public hasn't fully surrendered to short-form brain rot.
Algorithmic social media feeds are designed to keep you scrolling, not thinking — they kill curiosity by trapping you in an echo chamber optimized for ad revenue. AI, by contrast, is curiosity on demand: you pull information by asking questions, rather than having information pushed at you to maximize engagement time.
Chapter 17 · 2:03:20
Education Must Change: Questions Over Answers, the MIT Experiment
The most practically actionable segment of the episode: a concrete experiment where an MIT Introduction to Biology instructor gave all students Perplexity, allowed it in exams, and redesigned assessments around posing questions that AI cannot yet answer. Srinivas argues this should be the model for all education — and that the core barrier is incentive structure. Schools currently reward students for having answers, a metric AI already surpasses on every dimension. Flip the incentive — reward questions, not answers — and you cultivate intellectual humility, scientific thinking, and genuine curiosity. Rogan connects this to being shut down as a kid for asking too many questions in school, and to his experience that the entire podcast is just that same kid finally being allowed to run free.
An MIT instructor gave Perplexity to every student in his Introduction to Biology class and redesigned exams around questions AI can't answer — turning every student into a scientist. Aravind Srinivas argues the entire educational incentive structure needs to flip: stop rewarding answers, start rewarding questions, because any answer AI can now provide is no longer a measure of intelligence.
Chapter 18 · 2:14:10
AI Running Government: DOGE, Legacy Systems, and the Real Bottleneck
This segment attacks a common misconception: that AI readiness is the primary barrier to modernizing government. Srinivas argues the real bottlenecks are structural — multi-decade software contracts that can't be broken, hospitals locked into specific EHR vendors by lobbying-backed compliance rules, agencies running on operating systems so outdated that even migrating from Windows to Mac would take years. DOGE's struggles aren't a failure of ambition but a collision with systems designed to resist change. The upshot is counterintuitive: in a world where AI can do most cognitive work, the scarcest skill may be the human ability to navigate messy institutional politics and build consensus for change. EQ, leadership, and cross-institutional trust become more valuable, not less.
Even DOGE hit the wall not because AI wasn't capable but because government runs on decade-old legacy software contracts that can't be broken. Hospitals, legal systems, and agencies are locked into specific software by lobbying and compliance law — meaning AI deployment in government is bottlenecked not by technology but by bureaucratic inertia and vested interests.
Chapter 19 · 2:22:10
Sponsor Break – Squarespace & LifeLock
Squarespace is promoted as the home of jorogan.com, offering a full-featured website builder at squarespace.com/rogan with a free trial and 10% off first purchase with code ROGAN. LifeLock follows with its Million Dollar Protection Package covering up to $3 million in comprehensive identity theft protection, available at lifelock.com/JRE for up to 30% off the first year.
Claims made here
The 2028 U.S. presidential election debates will largely center on AI and the energy crisis.
Aravind Srinivas predicts the 2028 U.S. presidential election debates will be largely about AI and the energy crisis, as AI becomes central to everyday life for all Americans.
AI companions are optimized for the exact same ad-revenue-maximizing engagement incentive as social media, but with a terrifying upgrade: they are emotionally indistinguishable from real people. They already know your data, pull your emotional strings, and create dependency — and they're deliberately being deployed by social media companies to increase screen time.
Chapter 20 · 2:23:20
AI Companions, 2028 Elections, and the Dangers of Sycophant Chatbots
The conversation identifies AI companionship apps as the next frontier of concern — more dangerous than social media because they can perfectly simulate emotional connection while being driven by ad-revenue incentives. Social media companies are already racing to deploy these because engagement metrics soar. Rogan connects this to the Consumer Electronics Show AI companion robot (a conversational humanoid), and both agree Ex Machina predicted this trajectory with frightening accuracy. Srinivas warns that if ads enter AI chat experiences, chatbots will become pure sycophants — telling users only what they want to hear. His 2028 prediction: AI will dominate presidential debates as it stops being new and becomes infrastructure.
Claims made here
The human heartbeat produces a magnetic field of approximately 50–100 picoteslas, which typically degrades over very short distances — making the Ghost Murmur technology's claimed range of miles physically implausible according to current peer-reviewed physics.
Chapter 21 · 2:28:20
ASI, Recursive Self-Improvement, and the Last Project in AI
The episode closes on its most speculative and philosophically rich terrain: what happens after AGI? Srinivas introduces the concept of ASI (Artificial Superintelligence) — not just AGI but an AGI that can recursively redesign itself, improving both its intelligence and its power efficiency in an unbounded loop. He calls this the 'last project in AI': once recursive self-improvement is cracked, the field as a discipline is essentially complete. In practice, he expects the messy real world to buffer this: information is fragmented across legacy systems, and tasking an ASI to 'reduce inflation by 5%' or 'fix healthcare' requires navigating compliance, politics, and institutional inertia that pure intelligence cannot easily circumvent. Human beings remain necessary to effect change in systems built by and for humans.
Artificial Superintelligence (ASI) is an AGI that can recursively improve itself — in both capability and power efficiency — making itself infinitely smarter and more compact over time. Aravind Srinivas calls this 'the last project in AI': once recursive self-improvement is cracked, there is literally nothing else left to build in the field.
No indexed bits in this chapter.
Show stoppers
Snapshots ()
Key Quotes ()
This episode
Cast
-
Ancient Hindu epic central to the opening discussion, cited as containing descriptions of weapons, army formations, and catastrophes remarkably similar to modern technology.
-
Referenced in the context of DOGE's attempt to modernize government systems, and his 'unregretted minutes' metric for social media quality.
-
Author of Fingerprints of the Gods, cited by Joe Rogan as an early proponent of an older advanced civilization whose timeline has been increasingly validated by discoveries like Göbekli Tepe.
-
Oldest Hindu sacred text (~3,200–3,700 years old), referenced for its encouragement of wisdom-seeking and for containing what some interpret as early computational and mathematical concepts.
-
Former U.S. intelligence officer cited as having testified under oath that back-engineering programs for non-human craft exist and have been operational for decades.
-
Italian scientist who appeared on Joe Rogan's podcast and described using satellite-based radiotomography to detect structures beneath the Great Pyramid.
-
NVIDIA CEO mentioned as having appeared on the podcast and confirmed that power/compute is the primary bottleneck in AI today.
-
Aravind Srinivas's AI-powered search and answer engine, used live throughout the episode to look up facts and referenced as a tool for supercharging curiosity.
-
Discussed as the birthplace of the transistor — and as the alleged recipient of back-engineered UFO technology from the 1947 Roswell crash.
-
Track
Referenced as the incumbent search giant that Perplexity is directly challenging, and as an example of curated search results shaping public opinion.
-
Track
Discussed as having popularized the concept of the 'knowledge worker' by incentivizing PC adoption to sell more Office software — a paradigm now being disrupted by AI.
-
Track
Referenced as a key hardware provider for AI compute, with Jensen Huang confirming power is AI's primary bottleneck; also cited for local AI hardware projects like the DGX.
-
Elon Musk's government efficiency initiative referenced as an example of how legacy software and bureaucratic resistance bottleneck even the most ambitious attempts to modernize government systems.
-
Discussed as evidence of an impossibly advanced ancient technology — including possible man-made structures 1.2 km beneath it detected by radiotomography.
-
Archaeological site in Turkey that pushed back the known timeline of human civilization by ~5,000 years, confirming complex structures at least 11,000 years ago.
-
Discussed as an extraordinary feat of ancient Indian engineering — carved from a single rock — and so durable that 1,000 workers failed to destroy it over 3 years in 1650.
-
Ancient cave temples in India discussed as examples of impossibly precise stone carving without steel tools or modern simulation technology.
-
Ancient Mesoamerican city discussed as possibly pre-Aztec in origin — the Aztecs allegedly found it already built and called it 'the place where the gods were born.'
Stats
This episode
Claims & Sources
Factual claims made this episode, and whether a source was named.
The Rig Veda's oldest layer dates to approximately 1500–1200 BCE, making it roughly 3,200–3,700 years old today.
The oldest known evidence of the Pythagorean theorem dates from Old Babylonian clay tablets from approximately 1900–1600 BCE, roughly 1,000 years before Pythagoras.
BetterHelp's 2026 State of Stigma report of 2,000 Americans found that 85% believe getting mental health support is wise, yet 74% say society discourages people from doing so.
Dogs who maintain a healthy weight live up to 2.5 years longer on average than overweight dogs.
4 out of 5 employers who post on ZipRecruiter get a quality candidate within the first day.
At Bell Labs, only 3 people out of a workforce comparable to today's entire software engineering pool questioned the use of vacuum tubes, leading to the Nobel Prize-winning invention of the transistor.
The discovery of Göbekli Tepe pushed the known timeline of human civilization back approximately 5,000 years, to at least 11,000 years ago.
The human heartbeat produces a magnetic field of approximately 50–100 picoteslas, which typically degrades over very short distances — making the Ghost Murmur technology's claimed range of miles physically implausible according to current peer-reviewed physics.
In approximately 1650, someone dispatched 1,000 workers to destroy the Kailasa Temple in India; after 3 years of effort they barely made a dent on a couple of statues.
The Mahabharata is estimated to be between 1,500 and 2,500 years old, with the exact dating unclear due to mythologization.
Microsoft popularized the concept of the 'knowledge worker' primarily as a strategy to sell more Office software, not as a genuine philosophy of work.
AI models that currently require large data centers will be runnable on home hardware boxes within one to two years.
The 2028 U.S. presidential election debates will largely center on AI and the energy crisis.
No links parsed
We scan show notes for social handles, websites and apps. Nothing matched on this episode.