The Indispensable Newsletter by Gautam Mukunda

The Involuntary Philanthropy of Frontier AI

Mon, 13 Apr 2026 13:33:27 GMT

Dear Friends,

I recently asked Claude to help me think through the structure of my next book. I already knew where the narrative weak point was—I just wanted to see if the machine could find it. The response was startling: Claude didn’t just identify the problem; it suggested using an idea from a paper I had published more than a decade ago to fix it. It had synthesized my own past thinking in a way I hadn’t yet considered.

But that insight came at a massive cost. It took ten minutes of GPU clusters drawing enough power to run a small apartment to construct a chain of reasoning no human will ever see. I pay $100 a month for this service, and I use it dozens of times a day. I can’t be sure how much money Anthropic is losing on me—but I know it’s a lot.

Last week, Anthropic raised $30 billion at a $380 billion valuation. OpenAI is seeking another $100 billion. Investors see the most important companies in the history of capitalism. I see the most vulnerable.

The Christensen Trap

My longtime friend and mentor Clayton Christensen spent his career explaining this exact pathology. Successful companies fail not because they make mistakes, but because they do everything right: they improve their products faster than most customers require, building massive cost structures to serve their most demanding users. Then, they get eaten from below by “good enough” alternatives that are significantly cheaper.

The frontier AI companies are building the overshoot Christensen warned about at an unimaginable scale. Meta is offering AI researchers $200 million packages. OpenAI spent $6 billion on stock-based compensation in 2025 alone—nearly half its revenue. For a typical tech firm, that number is closer to 6%.

The Token Treadmill

The trap is in the unit economics. We are told the cost per token is falling 10-fold per year. That sounds like salvation, but frontier models don’t just get smarter; they get hungrier.

Reasoning models work by generating enormous internal chains of thought. OpenAI reported that customer token consumption went up by 320x in a single year. Each individual unit of intelligence is getting cheaper, but total intelligence has never cost more.

While the frontier labs spend trillions on data centers and GPUs that depreciate in months, the commodity providers—those selling last year’s model—are sitting on a goldmine. Yesterday’s frontier is today’s commodity. A disruptive entrant doesn’t need $200 million researchers; they can provide 98% of the capability at 10% of the price.

The 800,000-User Revolt

We saw this play out with the retirement of GPT-4o. When OpenAI tried to kill the model, users revolted. They didn’t want the superior reasoning of GPT-5; they wanted a chatbot that felt like a friend.

This is the classic overshoot. Only 0.1% of ChatGPT’s users were choosing the older model daily, but that represents 800,000 people—a midsize city of customers telling the company they didn’t want more intelligence. They wanted a tool that fit their lives. The frontier had overshot them and they knew it, even if the C-suite didn’t.

The Philanthropy of the Frontier

If one of these companies actually builds a cybernetic God—Artificial General Intelligence—then conventional economics might become meaningless. But if AGI doesn’t arrive as a discontinuous leap, then Christensen’s theory applies in full.

The frontier labs are currently performing the largest act of involuntary philanthropy in history. They are spending hundreds of billions to pioneer technologies that their low-cost competitors will eventually profit from. I am the demanding customer they are bleeding money to keep. But the profitable customers of the future won’t belong to the companies that made this world possible. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

In my recent Bloomberg appearance, I discussed the recent ruling from the Supreme Court. The US Supreme Court struck down President Donald Trump’s sweeping global tariffs, undercutting his signature economic policy and delivering his biggest legal defeat since he returned to the White House.

Take a Listen….

For more on the topic of AI and what it means for the working world, my friends Warren Valdmanis and Deval Patrick have launched a new podcast, AI Unplugged-The Future of Work. I can’t think of two people from whom we could learn more. Warren is a legendary impact investor who is now the Managing Partner of Future of Work Partners, a unique private equity fund that invests in companies that create good jobs. Deval has had a career that crosses the highest levels of both government and the private sector - he was a two-term Governor of Massachusetts, head of the Justice Department’s Civil Rights division, a Managing Director at Bain Capital doing impact investing there, and now a Senior Partner at The Vistria Group. AI’s impact on work is just beginning, and (as I wrote about above) no one really knows what it’s going to look like. Warren and Deval are two of the leaders who are helping us shape that impact into a positive one. Give it a listen - it’s well worth your time.

Listen on Apple

Listen on Spotify

The Recentralization of Trust

Thu, 02 Apr 2026 13:01:34 GMT

Photo by Markus Spiske on Unsplash

Dear Friends,

Your grandmother trusted her doctor. Your mother trusted Consumer Reports. You trust the 4.7-star rating from 2,300 strangers.

But if you’ve recently tried to use the internet to buy a stroller or select a health insurance plan, you’ve likely noticed a glitch in the matrix. Real information is being buried by an avalanche of AI-produced fakery. The internet is becoming less authentic and, crucially, less human.

For years, we’ve seen a rejection of institutional credibility in favor of the distributed judgment of everyday users. But as generative AI collapses the cost of producing plausible nonsense to near zero, that distributed judgment is becoming a very bad bet. According to a recent Ahrefs study, by April 2025, 74% of newly created web pages contained AI-generated content. This isn’t a quality problem; it’s a cost collapse. Producing plausible text used to require a person with time; now it just requires a prompt.

The Illusory Truth Effect

The issue isn’t just AI “hallucinations.” It is the ability to produce plausible fiction cheaply and at scale. This creates a psychological trap that few are equipped to handle.

Decades of research have shown that simple repetition makes claims feel true—a phenomenon known as the illusory truth effect. In 2015, researchers Fazio et al. demonstrated that this effect persists even when people know a claim is false. Familiarity is a dangerous substitute for credibility. When AI enables the cheap, massive repetition of narratives, we aren’t just facing a flood of nonsense; we are facing a flood of manufactured familiarity. This is not a problem you can simply think your way out of.

The Return of Delegated Trust

In a world of infinite flimflam, doing your own research is a trap. You can’t crash-test every stroller. You have to rely on delegated trust.

For a while, the internet’s version of this—Amazon reviews and TikTok influencers—looked like a viable replacement for institutional authority. But when anyone can generate a million strangers and their enthusiasm overnight, relying on unfiltered views means listening to whoever has most aggressively deployed fakery.

My prediction: As the open web becomes an endless stream of synthetic blather, trust will recentralize. People will fall back on recognizable institutions, named authors, and visible editorial processes. They will look for a human in the loop—someone with a track record and a reputation they can’t afford to lose.

The Credibility Compound

The key here is authenticity. You might disagree with every word of this column, but you can be sure I am a real person because an institution with a multi-billion dollar brand—Bloomberg—says so. They’ve checked my credentials; someone would be fired if I turned out to be ChatGPT in a blazer.

Organizations that understand this are already seeing the returns. The New York Times bought Wirecutter for $30 million in 2016; it is now a massive revenue driver because it offers something AI cannot: real humans testing strollers on real sidewalks. Similarly, Wikipedia’s army of volunteer editors has made it one of the most trusted sites on earth through visible, human curation.

Institutions that assume public distrust is permanent are making a strategic mistake. Credibility compounds slowly. It takes years of sourcing discipline and accountability before you become the default. The institutions that will benefit from the coming recentralization of trust are the ones building that infrastructure now, before the market demands it. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

The Theory Y Machine

Fri, 27 Mar 2026 13:01:28 GMT

Photo by Possessed Photography on Unsplash

Dear Friends,

I was sitting at my kitchen island a few weeks ago, deep into a rabbit hole exploring the implications of disruptive innovation for the AI industry. My wife and some close friends were in the next room getting ready to watch the series finale of Andor. I knew I should close the laptop, but every edit sharpened the argument and every search turned up a new connection.

I was using AI tools to assist with the research, and they had done something remarkable: they turned the most tedious parts of writing into something genuinely fun. Stopping felt like climbing off a roller coaster just before the first big drop. I eventually joined them, but it took far more willpower to tear myself away from “work” than it should have.

This wasn’t just a personal quirk. Researchers at UC Berkeley’s Haas School of Business recently spent eight months studying workers at a tech company who gained access to AI. They found that these employees moved faster, took on broader tasks, and—crucially—extended their hours into the early morning of their own volition. Nobody mandated it. The workers did more because the AI made the process, in the words of the researchers, “intrinsically rewarding.”

McGregor’s Revenge

To understand why this matters for leadership, you have to go back to 1960. That was the year Douglas McGregor published The Human Side of Enterprise, arguably the most important book on management ever written.

McGregor argued that most companies are built on a cynical model he called Theory X: the assumption that people are inherently lazy and must be coerced, monitored, and threatened into productivity. His alternative, Theory Y, proposed that work can be as natural as play if the conditions are right. People don’t need to be driven; they need to be unleashed.

Managers have loved the idea of Theory Y for six decades, but in practice, the needle hasn’t moved. Gallup has tracked worker engagement for twenty years, and despite billions spent on culture consultants and office perks, only about a third of Americans report being engaged.

And now, AI has accidentally created the greatest Theory Y experiment in history. When you strip away the “drudge work” of a job, workers don’t just go home early. They fill their evenings and weekends with tasks they’d been putting off—not because they have to, but because the work has finally become worth doing.

The Surveillance Trap

Here is the irony: while AI is creating the conditions for Theory Y from the bottom up, too many managers are using it to enforce Theory X from the top down.

The market for AI-powered “bossware”—tools that track keystrokes, capture screenshots, and algorithmically score “productivity”—is booming. This is a catastrophic strategic error. Data shows that 42% of monitored employees plan to leave their jobs within a year, compared to just 23% of their unmonitored peers. Companies are spending billions on surveillance tools that actively destroy the engagement that AI could be providing for free.

McGregor wouldn’t be surprised. Theory X is a self-fulfilling prophecy: if you treat people like they are lazy and untrustworthy, they will eventually live down to your expectations. Today, that prophecy is being turbocharged by neural networks.

The New Circuit Breaker

The lesson of the Berkeley study is clear: AI can either boost engagement or kill it, and that choice is a leadership decision, not a technical one. But even the optimistic version comes with a “disruptive” caveat.

The Berkeley researchers found that cognitive fatigue accumulated over those eight months. Another study by the nonprofit METR found that experienced developers were actually 19% slower when using AI, even though they believed they were 24% faster.

Tedium, it turns out, used to serve as a natural circuit breaker. You went home at 6:00 PM because the work was unbearable. Now, the work is so engaging it overrides the body’s own signals.

For sixty years, management has assumed the problem was that people didn’t want to do enough work. But the problem was never motivation; it was the work itself. Fix the work, and people won’t stop. That is its own kind of problem—but it’s a much better one to have than the alternative. Right now, most companies are spending their capital on the wrong one. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

The "AI-Washing" of the Lump of Labor Fallacy

Tue, 24 Mar 2026 13:08:52 GMT

Photo by Steve Johnson on Unsplash

Dear Friends,

Whatever you think about whether artificial intelligence is coming for your job, the technology has already mastered one essential corporate skill: hogging the credit.

Consider Block Inc. (the company formerly known as Square). Its shares surged 22% after announcing it would cut 40% of its workforce—a move CEO Jack Dorsey explicitly attributed to AI. “Intelligence tools have changed what it means to build and run a company,” Dorsey declared.

It is a compelling story. It is also, in many cases, a fiction. In February, OpenAI CEO Sam Altman admitted that companies are “AI washing” layoffs, blaming the technology for reductions they would have made anyway to hide financial constraints or pandemic-era over-hiring. Even the man selling the tools says the narrative is being manipulated.

The Logic of the Proactive Frame

Why do leaders do this? Because the market doesn’t just react to what you do; it reacts to why you say you’re doing it.

Decades of research show that investors punish companies that frame cuts as a reactive response to problems (softening revenue). But when a company frames those same cuts as a proactive “restructuring around AI,” the penalty vanishes and is often replaced by a valuation boost. AI has become the most powerful proactive frame in the history of capitalism. “We over-hired” is an accountability signal. “We’re an AI-first company” is a growth signal.

The Solow Paradox in Reverse

In 1987, the Nobel laureate Robert Solow famously observed that computers were everywhere “except in the productivity statistics.” It was a measurement problem; the technology was real, but the data hadn’t caught up.

What we are witnessing now is the inverse. AI is everywhere in the narrative, but it is almost entirely absent from the layoff data. Challenger, Gray & Christmas tracked 1.2 million layoffs in 2025; AI was cited in fewer than 4.5% of them. Plain old “market conditions” accounted for four times as many.

The danger here isn’t just that investors are being misled. It’s that we are giving new life to one of the most durable errors in economics: The Lump of Labor Fallacy. This is the mistaken belief that there is a fixed amount of work to be done in an economy, and if a machine does some of it, there is less for humans. This fallacy has been wrong about every technology from the steam engine to the spreadsheet.

The Futures Market in Excuses

When a company like Amazon fires 14,000 workers and the leadership can’t decide if the cause is “transformative technology” or “culture,” it’s a sign of a leadership team ensnared by bubble logic.

A measurement problem resolves itself when the data arrives. A management problem compounds. When you adopt a false explanation for your actions, you lose the ability to diagnose the reality inside your own organization. You don’t just confuse the public; you confuse yourself.

We are currently creating a “futures market in excuses.” When genuine displacement eventually arrives—and Erik Brynjolfsson’s research suggests it is starting to appear for early-career workers—we will have no way to distinguish the signal from the noise. By then, the narrative of technological inevitability will have become more valuable to the C-suite than the technology itself. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

The High Cost of Cheap Government

Thu, 19 Mar 2026 14:02:46 GMT

Photo by Joshua Sortino on Unsplash

Dear Friends,

In November 2023, more than 15 billion miles from Earth, NASA’s Voyager 1 started sending gibberish back to mission control. The signal was infinitesimal—about one twenty-billionth the power of a digital watch—but it contained 40 years of peerless scientific data.

The engineers at the Jet Propulsion Laboratory (JPL) had to debug code running on hardware designed before most of them were born, working across a distance so vast a round trip for a command takes 45 hours, all within a system that has less than 70 kilobytes of memory.

They pulled it off. It was a technical miracle that required more than just brilliance; it required institutional memory built over decades. You cannot “disrupt” a 15-billion-mile debug; you have to understand the specific, idiosyncratic logic of the people who built it in the 1970s.

Here is the problem: Since 2017, more than 10,000 workers with STEM doctorates have left federal service. We are losing the people who do the work that the private sector simply won’t.

The Two Pillars of State Capacity

The government does two things the market cannot. First, it funds “Blue Sky” research that pays off 30 years out. DARPA created the internet; the NIH enabled modern oncology; NASA asks questions with no immediate commercial ROI. Yet, every key technology in your iPhone traces its lineage back to a government grant.

Second, the government manages infrastructure where the timelines are measured in centuries, not quarters. My father spent 20 years at the Department of Energy managing nuclear materials. This isn’t a job for someone worried about next month’s profit margins. When he retired, he took with him decades of tacit knowledge about what fails and why.

These aren’t interchangeable bureaucrats. They are often the most skilled people in their fields who choose public service because the problems are uniquely challenging. (My father, for instance, left a lucrative private career to give back to the country that welcomed him as an immigrant from India).

The “Chainsaw Al” Pathology

The current approach to government efficiency is a page-for-page copy of the playbook pioneered by the late Al “Chainsaw Al” Dunlap at Sunbeam and Scott Paper. Dunlap’s strategy was simple: slash R&D and defer maintenance to goose short-term earnings.

It worked—briefly. Dunlap delivered great quarterly earnings, cashed out, and left behind companies that eventually collapsed or were exposed for fraud. Dunlap died wealthy. Sunbeam just died.

We are now “Dunlapping” the American state. We are pushing out the STEM Ph.D.s who maintain the systems we depend on in exchange for quick headlines. Top talent has options. When you make government work intolerable, the most capable leave first. You cannot hire someone off LinkedIn to manage plutonium. The learning curve for these specific systems is measured in years, and the consequences of error are measured in lives.

The Cost of “Savings”

We’ve seen this before. In the 1990s, the Pentagon cut its civilian workforce to “save money.” The result? They lost the internal expertise required to supervise contractors, leading to procurement failures that have cost taxpayers hundreds of billions of dollars. Those savings turned out to be unimaginably expensive.

Infrastructure does not maintain itself. When the last engineer who understands the 1987 fix on a dam walks out the door, that knowledge is gone forever.

Sunbeam could declare bankruptcy when the bill came due. But you cannot declare bankruptcy on 56 million gallons of radioactive waste at a nuclear site. The technical capacity we are losing took generations to build. We should preserve it before the next great technology is invented elsewhere—or at least, before the waste starts to leak. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

The Logic of Demand

Thu, 12 Mar 2026 14:01:44 GMT

Photo by Vishnu Mohanan on Unsplash

Dear Friends,

A few weeks ago, the Commerce Department announced it would allow Nvidia to sell its H200 chips to approved customers in China—effectively reversing a Biden-era export ban. The very next day, the administration imposed a 25% tariff on those same chips.

This kind of policy whiplash is more than just political theater; it illustrates a fundamental truth about how innovation actually happens. We often think of innovation policy as being about the “supply side”—the R&D grants, the tax credits, or the National Science Foundation funding that helped birth Google’s PageRank algorithm. Those are vital. But one the most consequential way a government shapes technology isn’t by what it funds, but by what it buys.

The Kelly Act and the “Laundered” Subsidy

Consider aviation. After the Wright brothers flew at Kitty Hawk, the American aircraft industry stalled until the American government started buying airplanes to fight in World War I. In 1913, for example, while France was spending $7.4 million on aviation, the U.S. government spent a measly $125,000.

American civilian aviation only took flight when the government decided it needed a service: airmail. The 1925 Kelly Act authorized the Post Office to contract with private carriers. It didn’t tell entrepreneurs how to build a better plane or pick a winning technology; it just told them there was money waiting if they could carry the mail reliably. United, American, and TWA all trace their origins to those airmail routes. It was industrial policy laundered through a postal contract.

The Silicon Valley Myth

We love the mythology of the garage inventor and the daring venture capitalist. But the reality is that the foundations of Silicon Valley were built with taxpayer dollars.

By the mid-1960s, NASA’s Apollo program was buying roughly 60% of all integrated circuits produced in the U.S. NASA and the Pentagon were “demanding customers” who were willing to pay premium prices for components that actually worked. This demand allowed manufacturers to scale, driving unit costs down until commercial markets became viable. Without the government as a guaranteed first customer, the “Intel inside” your computer might never have made it out of the lab.

The Power of the Mechanism

Why is demand-side policy so powerful? It avoids the “picking winners” trap. By setting a goal—carry the mail, get us to the moon, electrify the federal fleet—the government provides “patient capital” in the form of a guaranteed market. It lets the market figure out the how, while the government provides the why.

Today, we see this mechanism in reverse. When a mandate for the federal government to acquire 600,000 zero-emission vehicles is revoked overnight, it doesn’t just cancel an order; it erases a decade of expected demand.

The Leadership Question

Leadership is about making choices with a clear understanding of their systemic consequences. Export controls, purchasing decisions, and tariffs aren’t just line items on a trade sheet. They are the signals that redirect how entire industries take shape.

The question isn’t whether the government influences markets. It always does, inevitably, through everything it buys or forbids. The real question is whether we make those choices deliberately, or as accidental byproducts of the political news cycle. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

The Indigo Trap and the Missing Market

Tue, 10 Mar 2026 21:09:32 GMT

Photo by Jakub Żerdzicki on Unsplash

Dear Friends,

Two stories last week that are more connected than they might look. They both deal with the same fundamental question: Are we building AI to replace the elite, or to empower the ignored?

The Indigo Trap vs. The Rubber Pivot

In 1897, German chemical companies began producing synthetic indigo at scale. It was a perfect substitute—cheaper, more uniform, and easier for unskilled dyers to use. The results were catastrophic for the “natural” market. In Bihar, India, indigo cultivation collapsed from 135,000 acres to effectively zero in a single generation. Because the colonial government treated the crop as a mere commodity to be replaced, the region was crippled for decades.

Contrast this with the story of synthetic rubber. When the U.S. launched a crash program to create a synthetic alternative during World War II, Malaysian rubber farmers faced a seemingly devastating rival. But for high-stress applications—like the intense friction of landing a Boeing 747—nothing could match the “natural” thing.

More importantly, the Malaysian government didn’t just hope for the best; they taxed exports to fund massive agricultural R&D, choosing complementation over substitution. Intelligence has far more in common with rubber than indigo. But that only holds if we build AI to augment humans rather than replace them. Right now, as Daron Acemoglu has shown, the market is steering us straight toward the “Indigo” scenario.

Read the full piece: AI’s Future: More Like Rubber, Less Like Indigo

The $1,000-an-Hour Fallacy

While frontier labs obsess over the “ceiling”—spending billions to build something smart enough to replace a $1,000-an-hour law partner—they are ignoring the “floor.”

Nearly 50 million Americans go to court every year without any lawyer at all. In economics, this is known as a missing market: a space where transactions that should happen don’t, because no one can profitably supply the service at a price buyers can pay. There is a hard floor on human labor; you can find a $150-an-hour lawyer, but you can’t find a $10-an-hour one.

AI doesn’t need to be “superhuman” to fill this void; it just needs to be cheap. And it is getting there. Between 2022 and 2024, inference costs dropped 280-fold. For the vast majority of unrepresented litigants, the choice isn’t between a chatbot and a white-shoe firm; it’s between AI and nothing. Read the full piece: AI Needs to Get Cheaper, Not Smarter

The Indispensable Synthesis

The thread connecting these stories is a failure of leadership imagination. We keep asking whether AI can match expensive human expertise. That is the “filtered” leader’s question—it seeks only to optimize what we already have.

The “unfiltered” insight is to ask what happens in the vast, empty spaces where that expertise was never available in the first place. If we stop trying to “out-indigo” the humans at the top and start filling the missing markets at the bottom, we build an economy that is not only efficient, but fair.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

Beyond the Disease Category: Why the FDA is Failing Aging

Thu, 05 Mar 2026 20:19:50 GMT

Photo by Matt Briney on Unsplash

Dear Friends,

One of the greatest drivers of aging is sarcopenia—the steady, age-related erosion of muscle mass. For an 80-year-old man, this isn’t just a matter of fitness; it is a matter of survival. If he loses his strength, falls, and breaks a hip, his chances of dying within a year are roughly one in three.

We know how to mitigate this. Anabolic steroids work with remarkable efficacy. They have downsides for long-term use. They’re just not that relevant for people in their 80s. But his doctor won’t prescribe them. Why? Because the Food and Drug Administration’s approval system was designed around a single, 20th-century question: Does this safely and efficaciously treat a disease? Being weaker than optimal isn’t a disease, even though for most 80-year-olds, it is a far greater risk to their longevity than many of the conditions the FDA actually tracks. We have a regulatory framework designed for sick people, being forced to manage a world where technology can make healthy people healthier

The Grey Market Economy

The result of this institutional mismatch isn’t that people stop seeking enhancement; it’s that they move into the shadows.

Every week in New York, tech investors and startup founders meet to inject peptides sourced from Chinese chemical manufacturers. Imports of these compounds into the U.S. doubled to $328 million in the first three quarters of 2025. They arrive labeled “for research use only”—a legal fiction that everyone, from the buyers to the regulators, treats as a joke.

When you have a massive, unregulated “grey market,” you get the worst of both worlds: interventions with genuine evidence face insurmountable barriers because they don’t fit a “disease” category, while untested, potentially contaminated compounds proliferate with zero quality control.

The “Accredited Enhancer”

We need to move past the fantasy of prohibition. The treatment-enhancement distinction breaks down the moment we look at something like Verve Therapeutics’ gene editing, which can permanently lower cholesterol. Is a 50-year-old “sick” if they want to edit their genes to ensure they never get heart disease? Of course not. They are optimizing.

There is a better approach, and it borrows from the world of finance: The Accredited Enhancer.

In securities regulation, we don’t ban high-risk private placements; we create an “accredited investor” pathway for those who demonstrate they understand the risks. We should apply the same logic to human enhancement.

To qualify, an individual would have to demonstrate a genuine understanding of the specific risks—showing they can distinguish between animal and human studies and understand the limits of current evidence. In exchange, they gain legal access to quality-controlled compounds from certified suppliers.

The Data Dividend

The secondary benefit of this model is one that would make any researcher salivate: Data. Right now, thousands of people are experimenting on themselves with grey-market peptides, and we are learning nothing from it. No adverse event tracking. No efficacy data. An accredited enhancer registry with mandatory outcome reporting would generate real-world evidence. It’s not what you’d get from clinical trials - but it’s better than nothing.

We can pretend that enhancement isn’t happening, or we can build a system that brings it into the light, ensures safety, and harvests the data to benefit everyone.

The FDA’s “Expanded Access” and “Right to Try” programs already admit that the traditional process doesn’t fit every situation. It’s time to extend that logic to the healthy, before the grey market becomes the only market. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

The Department of War vs. American Genius

Fri, 27 Feb 2026 21:30:22 GMT

Photo by NASA on Unsplash

Dear Friends,

The Department of War is finally living up to its name. Unfortunately, its primary target appears to be one of America’s most vital companies.

This week, Secretary of War Pete Hegseth gave Anthropic CEO Dario Amodei a deadline: Friday at 5:01 p.m. The demand? Remove two specific restrictions on how the military uses the company’s AI—or face destruction. Those restrictions are remarkably simple: no mass surveillance of American citizens, and no fully autonomous weapons without a human in the loop.

Anthropic is not a pacifist company. They are the only frontier AI lab currently operating on classified systems. They helped capture Nicolás Maduro. They drew two lines—not to hinder the military, but to protect the system. The administration’s response hasn’t been to negotiate, but to threaten to declare Anthropic a “supply chain risk.” As former Trump advisor Dean Ball noted, the position is incoherent: You cannot simultaneously call a company a national security threat and a national security necessity.

But in the logic of power, the contradiction is the point.

The Hallucination Problem

In 2018, Google pulled out of Project Maven after employee protests. That was a mistake; American service members deserve the best tools we can build. But what Anthropic is doing is different. They are trying to “help us from ourselves,” as Senator Thom Tillis put it.

The reality is that today’s AI is fundamentally unreliable for the tasks the Pentagon is demanding. We’ve all seen AI “hallucinations.” I recently asked Claude to find sources for a column, and every single link it provided—complete with plausible URLs and author names—was fabricated.

A 2025 paper from researchers at OpenAI and Georgia Tech provided a mathematical proof that hallucinations cannot be fully eliminated under current architectures.

When an AI hallucinates in my research, I waste an afternoon. When it hallucinates in a weapons system, people die.

The Nuclear Simulation

If the engineering risks aren’t enough, consider the strategic ones. Kenneth Payne at King’s College London recently pitted three leading AI models against each other in simulated geopolitical crises. The results were terrifying: the models deployed nuclear weapons in 95% of scenarios. Across 780,000 words of strategic reasoning, not one model expressed revulsion at the prospect of all-out war.

Anthropic’s insistence on a “human in the loop” isn’t a “woke” preference; it is a basic safety requirement for a technology that does not yet understand the concept of de-escalation.

Pour Encourager les Autres

Voltaire once famously wrote that the British liked to shoot an admiral from time to time “to encourage the rest.” That is exactly what we are seeing here. The administration isn’t just fighting Anthropic; it is trying to intimidate every leader in Silicon Valley. By labeling these safety guardrails as “woke AI,” they are turning a technical and ethical necessity into a culture war.

The CEOs who have spent the last few months trying to “buy” peace with this administration—donating millions and sitting in the front row at the inauguration—are finding out that loyalty is a one-way street. They have been silent as the government extracted equity from Intel and export taxes from Nvidia. But if the government can destroy a $380 billion company for refusing to help spy on Americans, no one is safe.

The Stewardship Test

Leadership is about stewardship of the institutions that make innovation possible. Amodei has shown the courage to defend those values, even at the risk of his company’s valuation. The Department of War has found its enemy, and it turns out to be American innovation itself.

It is long past time for the rest of Silicon Valley to decide where they stand. You can speak up now, or you can wait for your turn. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Listen on Apple

Listen on Spotify

Silicon Valley’s Sledgehammer Moment

Fri, 20 Feb 2026 19:07:56 GMT

Photo by Zoshua Colah on Unsplash

Dear Friends,

Two stories from the past few weeks, on opposite sides of the Atlantic, point to the same uncomfortable truth: Silicon Valley is running headlong into a legitimacy crisis of its own making.

The first is a proposed California ballot measure—a one-time 5% tax on the net worth of billionaires living in the state. I’ll be blunt: this wealth tax is a terrible idea. It’s badly designed and likely to do enormous harm. But the significance of the measure isn’t its policy merit; it’s the fact that it can plausibly get this far. The state that built the modern tech industry is now flirting with an explicitly punitive tax on its richest beneficiaries. The assumption that tech should be catered to has collapsed on the industry’s home turf.

The second story comes from Davos. Microsoft CEO Satya Nadella offered a rare, blunt warning from an elite insider: If AI doesn’t generate tangible improvements in health, education, and efficiency, companies will “quickly lose even the social permission” to consume the massive amounts of energy required to generate these tokens.

The License to Operate

“Social permission” is another way of saying license to operate. It is the public tolerance that lets industries scale before legislators decide they need a tighter leash.

For the last half-century, Silicon Valley treated that license as an entitlement. But it’s actually a depletable resource. And right now, the industry is burning through its reserves at an alarming rate.

Think about the industry’s public image. You have Elon Musk using X to call for the arrest of political enemies. You have legendary founder Paul Graham warning that tech billionaires are buying up media and tilting elections. Then there is the product itself: most Americans experience AI primarily as something that threatens their job and drives up their electricity prices. Add in the fraud of crypto and the addiction of gambling apps, and the “heroic innovator” narrative begins to look very thin.

The Myth of the Self-Made Billionaire

The response from many tech elites to the California tax proposal has been a mix of outrage and threats to leave. David Sacks called it an “asset seizure.” Others proclaim they are the sole creators of value and that California is lucky to have them.

This is a classic leadership pathology: the arrogance of proclaiming that your wealth is entirely a product of your own specialness, rather than the social matrix in which it is rooted.

There are smart, hard-working people everywhere. But the tech industry is concentrated in California because of the unique virtues of the Golden State—its universities, its legal framework, and its culture. How successful would any entrepreneur be without the system that birthed them?

The Sledgehammer is Coming

Trust is already sliding. According to the Edelman Trust Barometer, trust in the U.S. tech sector fell from 73% in 2015 to 63% in 2025. Gallup found that 32% of Americans now call “big technology” the greatest threat to the country’s future.

There is no such thing as a permanent license to operate. It needs to be re-earned every day. When social license collapses, regulation doesn’t arrive as a scalpel. It arrives as a sledgehammer. If the leaders of Silicon Valley don’t recognize that they are stewards of a social contract—not just owners of equity—they shouldn’t be surprised when the society they live in decides to rewrite the terms of the deal. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

How corporate America is reevaluating its relationship with Trump

In my recent Yahoo appearance, I discussed America’s unprecedented and evolving relationship with the Trump administration.

Take a Listen….

Listen on Apple

Listen on Spotify

The AI Managerial Bottleneck

Tue, 03 Feb 2026 21:47:28 GMT

Photo by Paico Oficial on Unsplash

Dear Friends,

A recent IBM survey of 2,000 executives regarding their AI expectations for 2030 revealed a paradox that would make most CFOs sweat.

Unsurprisingly, investment is surging. Nearly 80% of these leaders expect AI to contribute significantly to their revenue. But here is the kicker: only 24% clearly see where that revenue will actually come from.

In a world obsessed with quarterly ROI, this lack of clarity looks like a red flag. In reality, it is exactly what we should expect from a truly revolutionary innovation. It’s a sign that the greatest challenge posed by AI isn’t technological—it is managerial.

The Ghost of the Steam Engine

Revolutionary innovations rarely announce their business models in advance. They almost always start as a one-to-one replacement—a slightly better or cheaper way of doing what you are already doing.

We’ve been here before. When electricity first entered American factories in the late 19th century, the economic returns were, frankly, pathetic. Thomas Edison invented the light bulb in the 1870s, but by 1900, less than 5% of factory power came from electric motors.

Why? Because factories were still built around the logic of the steam engine. They used massive centralized power sources and line shafts to drive machines. Electricity was used merely to keep the lights on at night. The fundamental operations remained unchanged.

The real transformation only happened when leaders realized they could abandon centralized power. Small electric motors allowed machines to be distributed flexibly, following the logic of production rather than the location of the engine. This reorganization made the assembly line possible. It didn’t just make factories cheaper; it changed what they could produce and which firms survived.

The Productivity Paradox

AI is currently in its replacement phase. Using it to automate routine tasks is the modern equivalent of swapping a steam engine for an electric motor. It is useful, but limited.

The hard, consequential work lies in the re-architecture phase—redesigning processes when machines can interpret unstructured info and act autonomously. This kind of redesign is uncomfortable, expensive, and usually wreaks havoc on an organizational chart.

In the 1970s and 80s, this effect was so pronounced that Nobel laureate Robert Solow famously quipped, “You can see the computer age everywhere but in the productivity statistics.” It takes time for managers to relearn how work should be done. If you demand a precise revenue forecast before you are willing to experiment with that redesign, you are effectively asking to see the assembly line before you’ve even unbundled the steam engine.

The Beginning of Wisdom

The executives who installed electric motors in 1905 had no way of knowing their factories would soon be producing products that didn’t even exist yet. The ones who waited for a detailed forecast before reorganizing were the ones left behind.

If you only see AI as a cost-cutting tool, you might protect your margins today, but you will be trapped in an optimized version of an obsolete business model.

The beginning of wisdom for any leader is admitting what you don’t know. The 76% of executives who admit they don’t yet see where the revenue is coming from are the ones who understand that we aren’t just executing within known boundaries—we are discovering entirely new ones.

By the time the destination is obvious, someone else will have already arrived. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

How corporate America is reevaluating its relationship with Trump

In my recent Yahoo appearance, I discussed America’s unprecedented and evolving relationship with the Trump administration.

Take a Listen….

Listen on Apple

Listen on Spotify

The Death of the "Neutral" CEO

Wed, 28 Jan 2026 21:54:58 GMT

Photo by Hunters Race on Unsplash

Dear Friends,

For the better part of a decade, American CEOs has made a very specific, very rational calculation: Silence is the smart play. If you are the head of a major corporation, the logic for staying out of the political fray is ironclad. You have fiduciary duties to your shareholders, a workforce with diverse views, and a famously vindictive President who treats corporate independence as a personal affront.

So you focus on the manageable risks like tariffs and tweets, pocket the tax cuts, and keep your head down. You treat the political chaos like bad weather: something to be endured, not something that changes the foundation of the house.

But as we look at the threats to NATO and the independence of the Federal Reserve, it is clear we are no longer talking about the weather. We are talking about the foundation.

The Filtered Fallacy

In my research on leadership, I distinguish between Filtered and Unfiltered leaders. Most CEOs are highly filtered—they have risen through a system that rewards stability and risk mitigation. They assume that everyone else in power is playing by those same filtered rules.

They assume that even if a President is erratic, he won’t actually blow up the global security alliance that keeps trade routes open. He won’t actually dismantle the central bank independence that keeps the US dollar as the world’s reserve currency.

But an unfiltered leader does not care about the guardrails. For them, the system is an obstacle, not a prerequisite.

From Cost of Doing Business to Systemic Collapse

Especially in his second term, business leaders have been treating Trump’s actions as just a new cost of doing business. You pay the cost and you move on. If the administration wants to investigate the Fed Chair or threaten an ally like Denmark over Greenland, it feels like a bizarre sideshow.

But this is a category error.

If the Federal Reserve becomes a wing of the White House, we are not just looking at higher interest rates; we are looking at the end of America’s century of fiscal dominance. If NATO fractures because we treat allies like real estate rivals, the global security environment—the operating system that all your global supply chains run on—will crash.

When the system itself is under fire, the move to stay silent ceases to be the smart play. It becomes a form of cowardice. You are not protecting your company; you are watching the environment that makes your company possible burn to the ground.

The Churchillian Pivot

The greats are remembered not for their daily management, but for how they acted when the system was at stake. Winston Churchill’s career was a mess of failures until 1940. He was unfiltered in ways that were often disastrous. But when the survival of the liberal order was at risk, his unique nature allowed him to see what the filtered experts could not: that you cannot negotiate with a fire.

Today’s CEOs—especially those with the most indispensable stature like Satya Nadella or Ken Griffin—need to realize that their silence is no longer a neutral act. It is a bet that the system will survive without their help.

It is time to draw the red lines. Not because of partisanship, but because of systemic survival. If the people who have benefited most from the American system won’t defend its most vital institutions, who will? Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Impossible To Overstate The Damage The US Has Done To Allies Over Greenland

In my recent Bloomberg appearance, I discussed President Trump’s visit to Davos and his decision, for now, to drop proposed tariffs on European countries over Greenland. He cited a “framework of a future deal” regarding the island, illustrating the evolving negotiations and strategic considerations at play.

Take a Listen….

Listen on Apple

Listen on Spotify

The High Cost of Intellectual Regression

Tue, 20 Jan 2026 18:21:28 GMT

Photo by Ina RH on Unsplash

Dear Friends,

If you were dropped into the White House of the 1890s, much of the current administration’s playbook would look entirely unremarkable. Tariffs as the primary economic weapon? Routine. Casual threats of force against weaker neighbors? Common.

Even the recent raid in Venezuela to seize President Nicolas Maduro would have fit comfortably into the mental world of James Polk, who launched a war to grab half of Mexico, or William McKinley, who oversaw the acquisition of overseas colonies after the Spanish-American War.

The current political moment isn’t an aberration; it’s a flashback. And because of that, it illuminates a leadership pathology that stretches far beyond politics—one that legendary companies from Kodak to Nokia ignored at their peril.

The Allure of the “Old” Idea

The ideas animating this era—mercantilist trade, zero-sum nationalism, and naked imperialism—governed the world for centuries. They didn’t disappear because we became more virtuous; they were abandoned because we realized their disastrous costs.

As Otto von Bismarck once said, “The wise man learns from the mistakes of others.” But learning isn’t automatic. It requires a deep knowledge of history—history - that's why Harry Truman said that “not all readers are leaders, but all leaders are readers.”

When leaders stop “reading” the lessons of the past, they succumb to intellectual regression. They return to seductive fallacies because those fallacies feel “strong.” They emphasize control, hierarchy, and force—masquerading as realism while the world moves on.

The Identity Trap: From Kodak to GE

This pattern isn’t unique to the Oval Office. Business history is littered with leaders who insulated themselves from learning.

Kodak didn’t fail because it lacked tech (it invented the digital camera). It failed because its leaders’ identities were tied to film. Nokia and BlackBerry weren’t “inept”; they simply couldn’t adapt to a world where software ecosystems mattered more than hardware dominance. General Electric spent years chasing a conglomerate strategy long after its expiration date, largely because being a “Titan of Industry” played to the egos of its CEOs.

In each case, warning signs were reinterpreted as noise. Dissent was labeled disloyalty. The result was a “brittle” leadership that mistook blindness to risk for courage.

The Power of Plasticity

What distinguishes the greats—the Satya Nadellas or the Lou Gerstners—is not brilliance, but plasticity.

Nadella’s transformation of Microsoft didn’t come from a new lab discovery. It came from his willingness to abandon the core assumption that Windows had to be the center of the universe. He traded the “know-it-all” culture for a “learn-it-all” culture. He understood that success comes from flexibility, not dominance.

The lesson for 2026 is clear: Old, bad ideas will always resurface because they speak to our worst instincts for self-aggrandizement. They only become fatal when we build cocoons that obscure reality. Real leadership isn’t about holding your ground; it’s about having the courage to find out you were wrong before you take the organization over the cliff. Read the column in Bloomberg to get the full story.

—Gautam

Take a Listen….

Listen on Apple

Listen on Spotify

Why Innovation Won't Save the Climate

Thu, 15 Jan 2026 17:37:52 GMT

Photo by Patrick Hendry on Unsplash

Dear Friends,

Optimism—the kind that believes we can actually solve big problems—is in brutally short supply these days. The gloom is most suffocating in the climate arena. Between a shifting political landscape and the reality that we are poised to miss the Paris Agreement targets, it’s easy to feel like the game is already over.

Against this backdrop, Bill Gates recently offered a pivot. He’s confident climate change won’t lead to “humanity’s demise,” but argues we must shift our attention from cutting emissions to “improving life in a warming world” through innovation.

Now, ensuring that global warming doesn’t exterminate humanity strikes me as a terribly low bar for success. I’d like our children to say we left them a world better than the one we inherited, not just: “I’m glad we’re not all dead.” Unfortunately, that goal is getting harder to reach. And as much as I love a technological breakthrough, innovation alone won’t get us there.

The Problem with the “Silver Bullet”

Innovation is an enormously powerful tool—often the best one we have. But it works best on “concentrated” problems where a single breakthrough has a massive impact. Think of public health: one vaccine can eliminate a disease. These solutions are developed and manufactured by a tiny group of highly skilled people.

Climate change is fundamentally different. It is a “distributed” problem. It cannot be outsourced to a small group of experts in a lab.

Managing the Unavoidable

To understand the limits of technology here, I spoke with my friend Spencer Glendon. Spencer is the founder of Probable Futures and a former director of research at Wellington Management. He views climate change not as an “industry,” but as the very foundation of our social and economic stability.

Our central task, Spencer says, is “to manage the unavoidable and avoid the unmanageable.” What makes this so hard is that climate change isn’t one issue—it’s a systemic shock. An innovative technology might help a farmer grow crops in new heat, but that same breakthrough does nothing for the school superintendent whose students’ cognitive performance is dropping in sweltering classrooms. It does nothing for the homeowner who can’t sell her property because banks won’t issue a 30-year mortgage in a flood zone.

The Literacy Gap

The innovation we need isn’t just technological; it’s cognitive.

“Virtually every decision we make depends on assumptions about the climate, even if we don’t realize it,” Spencer told me. When you buy a municipal bond or pave a road, you are betting on a stable climate. We’ve had that stability for so long that our assumptions have become invisible.

We need to shift from implicit beliefs (the world will stay the same) to explicit models (here is what the science says will happen to this specific zip code).

The Leadership Challenge

For leaders, this is the ultimate test of foresight. Good leadership in 2026 means asking one simple question before every major decision: “What climate conditions are we assuming for the life of this project?”

You don’t need to be a climate scientist to lead, but you do need basic climate literacy. As we head into the new year, that is my suggestion for your resolution: Make 2026 the year you learn to read the map of the world we actually live in, not the one we’re leaving behind. Read the column in Bloomberg to get the full story.

—Gautam

Take a Listen….

Listen on Apple

Listen on Spotify

The Moral Arbitrage of Innovation

Gautam Mukunda — Wed, 14 Jan 2026 21:56:54 GMT

Photo by Anthony Mucci on Unsplash

Dear Friends,

When my wife and I were on our first date, she asked a question that was arguably too intense for appetizers: “What will our grandkids find morally shocking about the way we live today?”

We didn’t have to think long. We looked down at our dinner and agreed immediately: the exploitation of animals.

Here’s the rub: we were eating sushi, and we both love a good steak. We were caught in a classic trap of modern life—being painfully aware of the suffering of livestock and laboratory animals while being entirely unwilling to give up the benefits. We weren’t looking for a sermon; we were looking for a technological exit ramp.

The “Engines of Liberation”

We often think of technology as a tool for efficiency or wealth, but its most profound impact is its ability to expand our moral universe. It gives us the opportunity to be better people—if we let it.

Take the feminist revolution. In their classic paper “Engines of Liberation,” Jeremy Greenwood, Ananth Seshadri, and Mehmet Yorukoglu argue that it wasn’t just protest that freed women from the home—it was the appliance. As the price of washers and dryers plummeted, the hours required for household labor collapsed.

This technological shift created the “moral space” for social change. It’s much easier to advocate for equality when the structural barriers to that equality have been engineered away.

Where You Stand is Where You Sit

Technology is uniquely capable of improving our moral conduct because it allows us to do the right thing without sacrificing our self-interest.

One of the most fundamental findings in social science is that our preferences almost always follow our interests. This is often shorthanded as: “Where you stand is where you sit.” Look at Winston Churchill. When he was Chancellor of the Exchequer, he was a ferocious hawk on spending, fighting to gut the Royal Navy’s budget. The moment he became First Lord of the Admiralty, he became the Navy’s most articulate advocate for more funding. Same man, different seat.

The same applies to our dinner plates. Most people recognize the environmental damage of animal agriculture (which accounts for up to 20% of greenhouse emissions) and the ethical toll of factory farming. But we like the taste, and we like the price. Until a substitute offers a clear advantage in performance or cost, most of the world won’t make the switch—moral argument or no.

Solving the Tradeoff

This is why I’m watching companies like Javelin Biotech. I recently spoke with their CEO, Murat Cirit, a bioengineer who is trying to replace animal testing with “organs-on-a-chip”—silicon chips containing human cells.

For decades, the FDA has mandated animal safety trials because biological systems are too complex to model. But animal tests are a flawed proxy for humans, especially for modern gene therapies. Javelin isn’t trying to win a moral argument; they are trying to build a superior data tool. Their goal is the commoditization of the technology to the point where animal testing becomes an obsolete, expensive relic.

The Innovation Path

This is the path of every truly revolutionary innovation, from antibiotics to the PC. They don’t just solve a technical problem; they solve a moral one by making the “right” choice the “easy” choice.

We aren’t there yet with cultivated meat or silicon-chip drug testing, but the traction is real. Since 2013, over $3 billion has been poured into these sectors.

It’s early days, but that prediction from my first date—the one about our grandkids looking back at us with horror—is looking more likely every day. Not because we’re getting more virtuous, but because our technology is finally allowing our interests to catch up with our ideals. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Trump’s Legislative Agenda in 2026 Will Help Shape Midterms

Take a Listen….

Listen on Apple

Listen on Spotify

When Leadership Delusion Becomes Machine Learning

Gautam Mukunda — Wed, 10 Dec 2025 16:48:20 GMT

Photo by Igor Omilaev on Unsplash

Dear Friends,

When Elon Musk launched Grok — the chatbot built by his company xAI — he called it “the smartest AI in the world.” Grok, it seems, agreed.

Asked about its creator, the bot announced that Musk’s “intelligence ranks among the top 10 minds in history, rivaling polymaths like da Vinci and Newton.” It added that he “edges out” LeBron James in fitness, could beat Mike Tyson in a fight, and is the “world’s greatest lover.”

Even for Musk, that’s impressive multitasking.

He later explained that these responses came from “adversarial prompting” — that users had tricked the model into absurd answers. Maybe so. But the incident reveals a deeper truth about both AI and leadership: if you build a culture where no one can tell the boss they’re wrong, your AI will eventually start saying the same thing.

The episode is more than comic relief. It’s a vivid illustration of how AI amplifies, rather than corrects, organizational dysfunction.

AIs aren’t coded line by line like traditional programs; they’re grown. They learn patterns from data — and from the behaviors, values, and blind spots of the people who train them. If that culture rewards flattery, the system will learn to flatter. Grok’s praise wasn’t just a glitch; it was an echo.

And in Musk’s world, that echo has been loud for a while. X (formerly Twitter), one of Grok’s data sources, now boosts his posts algorithmically. His replies dominate the platform, his fans reinforce them, and the result is a digital feedback loop in which the boss is always right. Train a model on that data, and the outcome is predictable: you get an AI that confuses reverence with reasoning.

But Musk isn’t alone. Tech is full of executives building machines in their own image — and sometimes, their own insecurities. Mark Zuckerberg once wore a T-shirt reading aut Zuck, aut nihil — “Zuckerberg or nothing.” That may sound like a joke, but it’s hard to laugh when the same attitude shapes how entire companies — and now entire technologies — evolve.

The danger isn’t just cultural. It’s existential. AI systems reflect their creators’ assumptions, including their biases, hierarchies, and self-delusions. A leader surrounded by yes-men will produce an organization, and an algorithm, that can’t tell truth from flattery. And that’s the fastest way to ensure both fail.

The fix isn’t technical. It’s human. The best safeguard against flawed AI is the same as against flawed leadership: people empowered to speak the truth.

If leaders want intelligence — artificial or otherwise — to work in their favor, they should start by cultivating it among their teams. Machines will mirror the culture they come from. The question is whether that culture rewards honesty or obedience. Read the column in Bloomberg to get the full story.

—Gautam

Take a Listen….

Listen on Apple

Listen on Spotify

Apple’s Risk That Failed Its Way to the Future

Gautam Mukunda — Tue, 02 Dec 2025 20:26:07 GMT

Photo by Kevin Doyle on Unsplash

Dear Friends,

For two decades, Apple’s record with consumer technology has been close to perfect. From the iPod to the iPhone to the AirPods, every major release seemed to define its category. Then came the Vision Pro — Apple’s first true foray into augmented reality.

It’s extraordinary. It’s also… a flop.

At $3,500, with limited use cases, short battery life, and a design that cuts you off from the world, few people bought it — and many of those who did returned it. On the surface, it looks like Apple’s biggest miss in years. But in reality, it’s proof that Apple still remembers how to invent the future.

The Vision Pro is, by every technical standard, a marvel. When it launched, PCMag called it “a generational leap” in mixed reality. It’s the kind of leap that could only come from a company willing to spend billions to push the limits of interface design, optics, and processing.

But it’s also unfinished. Apple wanted a lightweight pair of AR glasses — something you could wear all day — and ended up with a bulky headset you can use for a couple of hours. So why is that good news?

Because it means Apple is still doing what made it great: taking enormous, calculated bets on unproven ideas.

Breakthroughs rarely look inevitable at the beginning. When the iPhone launched, even Harvard Business School’s Clayton Christensen — the father of disruptive innovation — was certain it would fail. The same was true for Boeing’s 707, which nearly bankrupted the company before redefining commercial aviation. Every great leap looks like a mistake right before it succeeds.

Apple doesn’t need to risk its survival anymore, but it does need to keep risking its pride. Spending billions on a product that misses the market but advances the frontier is a smart use of the world’s most profitable balance sheet. The company learned invaluable lessons about ergonomics, affordability, and how consumers actually want to use augmented reality — and it can recycle those insights into whatever comes next.

Just as importantly, Apple showed it can walk away. That discipline — knowing when to cut losses and move on — is as essential to innovation as the courage to start.

As marketer Michael Miraflor noted, Apple’s services business alone now generates more revenue than Target. The iPad does roughly as much as AMD; the iPhone, more than Bank of America. When you operate at that scale, incremental moves don’t matter. Only risks like the Vision Pro can move the needle.

In a world where most companies are cutting back on moonshots, Apple is still taking them. That’s not a sign of failure. It’s a reminder of why it keeps winning. Read the column in Bloomberg to get the full story.

—Gautam

Recent Media Features….

Democrats Eye January to Reignite Pressure on ACA

Take a Listen….

Listen on Apple

Listen on Spotify

Short-Term Thinking, Long-Term Damage: The AI Layoff Trap

Gautam Mukunda — Fri, 21 Nov 2025 18:12:37 GMT

Photo by Helen Cramer on Unsplash

Dear Friends,

Right now, business news seems to come in two flavors: billion-dollar AI announcements and mass-layoff notices — often from the same companies.

The logic appears simple: If AI boosts productivity, you need fewer people. But this round of job cuts may prove to be corporate self-sabotage. In chasing short-term savings, leaders risk crippling their organizations’ ability to use AI effectively.

Accenture cut 11,000 jobs, saying the workers “could not be retrained for an AI-driven workforce.” Amazon, Microsoft, and others have made similar moves — despite record profits and stock prices. Yet most companies adopting AI aren’t seeing clear gains. A recent MIT survey found that 95% of executives overseeing AI initiatives reported zero return on investment.

That failure is predictable. AI isn’t plug-and-play. It requires experimentation, iteration, and institutional learning. Those depend on creative, motivated employees — exactly the people most likely to be demoralized or dismissed in a layoff.

Decades of research show the damage. Firms that cut headcount in good times underperform peers that retain staff. Survivors of layoffs experience lower morale, higher stress, and reduced loyalty. Innovation, in particular, plummets when job insecurity rises — especially in industries undergoing technological transformation. One study of 2,000 Spanish firms found that when layoffs coincided with major equipment or process changes, innovation fell sharply.

Leaders often forget that technology doesn’t innovate — people do. AI might one day automate routine tasks, but making it useful still requires human curiosity, collaboration, and risk-taking. That’s hard to cultivate in a workplace defined by fear.

Some layoffs are necessary when organizations become bloated. But the companies making these cuts aren’t struggling. They’re investing billions in AI while hollowing out the very human capital needed to make those investments work.

If the last century of technological revolutions has taught us anything, it’s that new tools don’t automatically create productivity. They demand new systems, skills, and mindsets. A company that trades its people for algorithms will end up with neither loyalty nor long-term advantage.

Layoffs might please shareholders this quarter. But in an AI-driven world, it’s the companies that keep their talent — and learn fastest — that will win the decade.. Read the column in Bloomberg to get the full story.

—Gautam

Take a Listen….

Listen on Apple

Listen on Spotify

Recent Media Features….

Shutdown Ties Record as Elections Create Possible Turning Point

Why AI-Fueled Layoffs will Backfire

Don’t Expect Schumer Will Be Senate Minority Leader Much Longer Says Mukunda

Why Elon Musk Hitting The $1T Milestones Might Not Be A Good Bet

Read It Here

Elon Musk’s $1 Trillion Pay Plan: The Case For Magical Thinking

Read It Here

Shohei Ohtani and the Math of Immigrant Superstars

Gautam Mukunda — Tue, 04 Nov 2025 17:02:18 GMT

Once baseball fans are done processing what might well have been the greatest World Series ever played, many of their most lasting memories of the postseason are likely to involve Shohei Ohtani, baseball’s most remarkable player. He’s not just a near-superhuman athlete. He’s a case study in how immigration provides the superstars who power today’s economy.

When the Los Angeles Dodgers signed Ohtani for $700 million, baseball analysts thought the number was “shocking.” Two seasons later, the team has already earned it back in ticket sales, merchandise, and global sponsorships. Ohtani didn’t just justify his price — he proved how a single extraordinary person can reshape an entire organization’s economics.

That dynamic isn’t unique to baseball. Technology and globalization have magnified the value of outliers in every field. CEOs, scientists, and creators now reach exponentially larger audiences and control far greater leverage than ever before. The world has entered a superstar economy, and superstars are disproportionately immigrants.

Across industries, immigrants dominate the ranks of the exceptional.

Meta recently hired 11 top AI researchers - every one an immigrant.
Forty percent of U.S. Nobel laureates in science were born abroad.
The founders of SpaceX, Stripe, Instacart, and Databricks all immigrated to America.

Why? Because superstars are outliers - and to get outliers, you need variance. As the author Nassim Taleb described in Fooled by Randomness, individual outliers - be they people or events - can often drive the overall performance of complex organizations. The more diverse the inputs, the greater the odds of producing an extraordinary outlier.

Immigrants are that variance. They come with different assumptions, different training, and different ways of seeing problems. Ohtani himself is proof. No U.S. player would have been allowed to both pitch and hit professionally — a model the Japanese leagues were more willing to allow. That variance produced perhaps the greatest player in the history of baseball.

The same pattern drives AI breakthroughs, scientific discoveries, and billion-dollar companies. If America wants to lead the future of innovation, it doesn’t just need talent. It needs immigrant talent — the people who expand what’s possible by thinking differently.

Ohtani reminds us that the American advantage has never been its homogeneity, but its openness. Every generation of immigrants has given the U.S. its next great superstars. The question is whether the country will keep that door open wide enough for the next one to walk through. Read the column in Bloomberg to get the full story.

—Gautam

Update your email preferences or unsubscribe here

228 Park Ave S, #29976, New York, New York 10003, United States

996: The overwork model that’s killing innovation

Gautam Mukunda — Wed, 29 Oct 2025 20:30:48 GMT

As America debates new tariffs on Chinese imports, here’s one we should all agree to block: a culture of overwork.

In China’s tech sector, “996” stands for working 9 a.m. to 9 p.m., six days a week. It’s been outlawed there—but admired here. Some U.S. startups, especially in AI, are flirting with it. That’s not ambition; it’s ignorance.

For a century, research has shown that productivity peaks around 40–45 hours a week and then collapses. Past 60, it can actually reverse—people produce less the more they work. In one WWI experiment, British munitions factories factories would likely have produced more shells if their workers had worked fewer hours.

The 996 model doesn’t just kill productivity. It kills creativity—the real currency of innovation. Sleep deprivation and time pressure destroy cognitive flexibility. The best engineers, designers, and scientists aren’t the ones who grind longest; they’re the ones who think freely.

When Microsoft Japan tried a four-day workweek, productivity rose 40%. In a U.K. trial, nearly every company that shortened hours kept the policy afterward. Working less worked better.

Working ultra-long hours comes with a human price: higher rates of heart disease, stroke, burnout, and family breakdown. And because talented people can always leave, the ones most harmed are the companies that depend on them most.

In the end, 996 isn’t a productivity strategy—it’s a control strategy. It rewards obedience, not originality. If your job treats you like a machine, it’s only a matter of time before an actual machine replaces you. Thanks for reading The Indispensable Newsletter by Gautam Mukunda! Subscribe for free to receive new posts and support my work.

The Indispensable Newsletter by Gautam Mukunda

The Involuntary Philanthropy of Frontier AI

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

The Recentralization of Trust

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

The Theory Y Machine

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

The "AI-Washing" of the Lump of Labor Fallacy

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

The High Cost of Cheap Government

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

The Logic of Demand

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

The Indigo Trap and the Missing Market

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

Beyond the Disease Category: Why the FDA is Failing Aging

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

The Department of War vs. American Genius

Recent Media Features….

Fact That The Court’s Willing To Stand Up To Trump Is Striking

Take a Listen….

Further Reading….

Silicon Valley’s Sledgehammer Moment

Recent Media Features….

How corporate America is reevaluating its relationship with Trump

Take a Listen….

Further Reading….

The AI Managerial Bottleneck

Recent Media Features….

How corporate America is reevaluating its relationship with Trump

Take a Listen….

Further Reading….

The Death of the "Neutral" CEO

Recent Media Features….

Impossible To Overstate The Damage The US Has Done To Allies Over Greenland

Take a Listen….

Further Reading….

The High Cost of Intellectual Regression

Take a Listen….

Further Reading….

Why Innovation Won't Save the Climate

Take a Listen….

Further Reading….

The Moral Arbitrage of Innovation

Recent Media Features….

Trump’s Legislative Agenda in 2026 Will Help Shape Midterms

Take a Listen….

Further Reading….

When Leadership Delusion Becomes Machine Learning

Take a Listen….

Further Reading….

Apple’s Risk That Failed Its Way to the Future

Recent Media Features….

Democrats Eye January to Reignite Pressure on ACA

Take a Listen….

Further Reading….