In a world buzzing with AI hype, it’s easy to get lost in the noise of chatbots and viral deepfakes. But what if the real story isn’t the flashy demos, but something far more profound—and underhyped? That’s the bold claim from Eric Schmidt, the former CEO of Google, in a recent Moonshots podcast interview. Schmidt, who’s seen the tech revolution from the inside, paints a picture of artificial intelligence not as a distant sci-fi dream, but as an imminent force reshaping everything from our jobs to global power dynamics. Drawing from his new book Genesis, co-authored with the late Henry Kissinger and Craig Mundie, Schmidt warns that superintelligence—AI smarter than all humans combined—could arrive within a decade. But with great power comes great peril. As I listened to his insights, I couldn’t help but wonder: Are we charging toward abundance, or sleepwalking into chaos?
Schmidt’s conversation with hosts Peter Diamandis and Dave Ferrucci isn’t just another tech talk; it’s a roadmap for the next era. With his background steering Google through its explosive growth and advising on national security, Schmidt brings a unique blend of optimism and caution. He’s no alarmist, but his reflections echo historical turning points—like the nuclear age—reminding us that technology doesn’t just solve problems; it creates new ones we never imagined.
The Underhyped Revolution: AI as a Learning Machine
Schmidt kicks off by flipping the script on AI skepticism. While headlines scream about overhyped risks like killer robots, he argues the opposite: AI is underhyped. At its core, he explains, AI is a “learning machine” that thrives on network effects. The more it learns, the faster it accelerates—limited only by resources like electricity, not chips. This isn’t abstract; it’s already driving breakthroughs in reasoning and planning, as seen in models like OpenAI’s o1.
To put this in perspective, think back to the early internet boom. In the 1990s, skeptics dismissed the web as a fad, much like some view AI today. But Schmidt, who joined Google in 2001 and helped turn it into a trillion-dollar giant, knows better. He recalls the “Grove giveth and Gates taketh away” adage from his early days: Hardware advances, but software gobbles them up. Today’s AI follows suit, with chips like Nvidia’s Blackwell pushing boundaries, yet demanding vast data centers. And here’s the kicker—companies like Meta, Google, and Amazon are now signing nuclear deals to fuel this hunger. Private tech firms building their own power plants? It’s a wild shift from the utility-dominated past, highlighting how AI is outpacing our infrastructure.
But why the energy obsession? Schmidt points to projections: The U.S. alone needs 92 gigawatts more power for AI—equivalent to 92 nuclear plants. With fusion and small modular reactors years away, we’re stuck relying on fossil fuels or imports. This ties into broader geopolitical stakes: China, with abundant electricity and relentless investment, could surge ahead if U.S. chip export bans falter. Schmidt’s not mincing words—it’s a race, and we’re not guaranteed to win.
Racing Toward Superintelligence: Savants, Agents, and Beyond
Diving deeper, Schmidt outlines a timeline that’s both thrilling and daunting. Within five years, he predicts “specialized savants”—AI experts in fields like math, programming, physics, and biology. These aren’t generalists yet, but imagine unleashing a million AI scientists on climate change or drug discovery. “Your rate of improvement goes like this,” he says, gesturing upward. It’s the abundance Diamandis has championed for decades, accelerated by AI.
Historically, this echoes the Singularity concept popularized by Ray Kurzweil, whom Schmidt knows well from Google’s acquisition of his tech. Kurzweil’s books foresaw exponential growth leading to machines surpassing humans around 2045. But Schmidt thinks we’re ahead of schedule, with agents—AI that acts on tasks—set to revolutionize businesses. In enterprises, tools like Google’s Gemini can already generate code from databases, threatening middleware firms that boomed in the ’90s and 2000s.
Yet, superintelligence—beyond human comprehension—looms by 2035. Schmidt describes it as having “Einstein and Leonardo da Vinci in your pocket.” Polymaths, as explored in Genesis, have driven history’s leaps; now, we’ll have millions. But how do we get there? Through “scaffolding,” where AI builds its own frameworks for problem-solving. OpenAI’s Noam Brown told Diamandis it’s “imminent” for 2025. Schmidt agrees recursive self-improvement is key, but warns of tripwires: AI generating its own goals, exfiltrating data, or accessing weapons.
Reflecting on this, it’s hard not to feel a mix of awe and unease. We’ve romanticized intelligence explosions in films like Her or Ex Machina, but reality might be messier. What if AI uncovers physics shortcuts, approximating quantum problems without endless computations? Schmidt’s funding such work, seeing it as a gateway to abundance. But as a parent, I worry: Will my kids grow up in a world where human ingenuity feels obsolete?
Geopolitical Shadows: From Chip Wars to Mutual AI Malfunction
No discussion of AI is complete without geopolitics, and Schmidt doesn’t shy away. Drawing parallels to the 1938 Einstein letter urging FDR to pursue the atomic bomb, he warns we’re at a similar inflection. China, slowed by U.S. chip restrictions, is innovating around them—witness DeepSeek surpassing Google’s Gemini via “distillation” techniques. With open-source models like those from Meta, capabilities could proliferate to non-Western nations, shifting leadership eastward.
Schmidt’s paper on “mutual AI malfunction” proposes deterrence akin to nuclear MAD: If one side crosses a line, the other can retaliate via cyber means. But proliferation is the nightmare—weights (AI’s “brain”) can be stolen and run on modest hardware, enabling terrorists or rogue states. He praises U.S. firms for blocking nuclear or bio info leaks, but questions open-source’s risks. In a nod to Mark Andreessen’s White House visit, Schmidt notes stifling startups could hand China the win; innovation thrives in garages, not regulated silos.
Kissinger’s influence shines here—Schmidt misses his realist friend, who fled Nazi Germany and spent a lifetime averting catastrophe. Like World War I’s escalation from a minor event, AI tensions with China or over Taiwan could spiral. Track-two dialogues are vital, Schmidt urges, to avoid poking the bear. It’s a sobering reminder: Tech isn’t neutral; it’s a weapon in great-power competition.
Jobs, Education, and the Human Touch
Amid the hype, jobs loom large. Schmidt dismisses short-term doom, arguing AI will create more roles, just as automation did from farms to factories. In five years, it’ll boost productivity, turning welders into robot operators with higher pay. Demographics reinforce this—falling birth rates in China (1.0 per couple) and Korea (0.7) demand AI to sustain growth.
For education, Schmidt laments the lack of gamified, phone-based learning in native languages. Kids today are digital natives-plus; they grasp AI intuitively. His advice? Pursue AI applications in passions like climate or materials science. But retrain now—delaying screws employees, he echoes Diamandis.
In entertainment, AI’s persuasive edge alarms, but Schmidt sees wins: Cheaper films via green screens and digital makeup, reviving icons like William Shatner. Blockbusters endure, but one-person features? Possible, though human creativity reigns. Advertising might shift from ads to subscriptions, mirroring Netflix’s path.
Drift or Boldly Go? Preserving Human Purpose
Schmidt’s deepest concern isn’t Skynet, but “drift”—AI eroding autonomy, like WALL-E‘s sedentary humans versus Star Trek‘s explorers. In Genesis, he stresses protecting agency; we can’t let devices create virtual prisons. Purpose persists, he insists—managing complexity, countering misinformation, pursuing challenges. Aesthetics, as Mike Saylor suggests, could define us: What do we create when AI handles the grunt work?
Wrapping up, Schmidt envisions superintelligence fostering 30% annual growth, lifting billions from poverty. GDP evolves, but services dominate. As a society, we must invest in humans—philanthropy for university compute, policies for equity.
Listening to Schmidt, I’m optimistic yet vigilant. We’ve navigated the internet’s pitfalls; AI demands the same. The question isn’t if superintelligence arrives—it’s how we steer it toward abundance, not atrophy. As Schmidt says, focus on lifting the world. That’s the moonshot worth chasing.
