The Unbiasing Pendulum of Entrepreneurship
"The moment you start believing your own BS, that’s when things start to crumble", Anonymous.
Over the past few months, and most recently at a CEE tech conference, I met dozens of founders I deeply admire for their passion and resilience. Yet beneath the optimism, a quiet unease lingered. Many expressed a shared disorientation: a feeling of drifting without fixed reference points, propelled by the inertia of the AI wave we all seem trapped in, heading toward a shore we cannot (yet) see.
I was reminded of a 2012 interview where Jeff Bezos warned against “chasing the hot new thing,” comparing it to trying to catch a wave already breaking. Instead, he advised founders to “position themselves and wait for the wave to come to them” by focusing relentlessly on the customer.
But in today’s hyper-accelerated AI landscape, that advice feels almost impossible to follow. In fact, the same Bezos described AI in 2024 as a “horizontal enabling layer,” claiming, “There is not a single application that you can think of that is not going to be made better by AI.”
The Returns
For decades, tech entrepreneurship has been biased toward growth. It’s our universal metric of success, measured in revenue, users, or valuation. The larger the dollar value and the shorter the time span, the greater the growth rate. Founders are praised for achieving it; investors are rewarded for spotting it early.
We tolerate early inefficiency, both moral and financial, because we believe progress will redeem it later. OpenAI generated around $4.3 billion in revenue in the first half of 2025, about 16% more than it generated all of last year, burning $2.5 billion in the process.
Anthropic revenues have grown rapidly this summer, jumping from $4 billion to $5 billion in just a few weeks, hoping to reach $9 billion by the end of the year (2025). At the same time the company agreed to pay $1.5bn to settle a class action lawsuit filed by authors who said the company stole their work to train its AI models.
Or consider the CEO who declares, “I’ll fire anyone who doesn’t use AI,” or the enforcement of a 996 work culture. These aren’t aberrations, they are expressions of a belief that temporary distortions are acceptable in the pursuit of progress. As Gordon Gekko, beautifully played by Michael Douglas in Wall Street (1987), famously put it:
“Greed is right. Greed works. Greed clarifies, cuts through, and captures the essence of the evolutionary spirit. Greed, in all of its forms.”
A recent UN Trade and Development (UNCTAD) report project the global AI market will soar from $189 billion in 2023 to $4.8 trillion by 2033, a 25-fold increase in just a decade. Such projections feed the entrepreneurs’ conviction that growth is both inevitable and desirable.
The Costs
Our reductionist obsession with growth blinds us to a fundamental economic variable: cost. Behind every exponential curve lies an invisible cost curve rising just as fast. The energy consumed to train ever-larger models, the legal and ethical fallout of unlicensed data use, the psychological toll of constant acceleration. These are all part of the same equation, just written on the other side of the pendulum.
Cost is often treated as an afterthought, something to be optimized “once we scale.” This mindset shapes not only how we build, market, and sell, but also how we lead. It encourages us to defer reflection, to kick the can down the road and rationalize inefficiency as “fuel for growth”.
Aristotle, in Oeconomica, offered a timeless warning:
“There is a consideration which is common to all branches of economy and which calls for the most careful attention, especially in personal economy, namely, that the expenditure must not exceed the income.”
Somewhere along the way, technology entrepreneurship lost touch with this basic (and ancient) principle. We built a trust economy, a system where projected future value is traded as present capital. We monetize hope (and hype) by pricing the promise (valuation), not the realized utility (value) produced by our labor (intellectual or otherwise).
In other words, we’re betting that LLM costs will collapse like semiconductor prices once did, and that AI demand will explode as the Internet’s did. Neither trend is ours to command, yet history tempts us to believe it’s inevitable because from renewable energy to internet infrastructure and electric vehicles, the pattern seems to repeat.
Many startups, following this “Bezosian wisdom,” use that argument to justify waiting. And if things happen to unfold in their favor, they claim victory, wearing the laurels of success, even if they were merely beneficiaries of a hopeful tide. Meanwhile, they perform their little theatrical ritual, pretending to deliver productivity gains without looking closely at costs. When they do, they shrug it off: “That’s someone else’s problem.” The cost of maintaining this act compounds over time, culminating in the erosion of customer trust. And if we really want to take history lessons, we shouldn’t be selective: in every technological wave, correction arrives sooner or later; usually before either demand or cost catches up.
Once more, Jeff Bezos on AI (2025):
“This is kind of an industrial bubble. It could even be good, because when the dust settles and you see who are the winners, society benefits from those inventions... and that’s what’s going to happen here.”
Economically, this echoes Nicholas Georgescu-Roegen’s The Entropy Law and the Economic Process: every act of production consumes energy and creates disorder. Systems that ignore cost imbalances must, eventually, return to equilibrium, often painfully. In startup terms, that equilibrium arrives when capital dries up, demand stalls, or trust evaporates.
The Cost of Returns
Today, some contrarian investors are starting to pay more attention to fundamentals - a welcome sign of maturity. They are beginning to recognize that: (1) AI-driven productivity is asymmetrically distributed, and (2) it comes at a cost.
I believe we are starting to shift gears, realizing that efficiency is the new metric in the age of AI. Not only growth, and not even productivity (understood as raw output). By optimizing operating costs, realized productivity moves closer to potential productivity. When this happens, the value created by an AI product (the outcome) aligns with that of the startup building it (the valuation). This isn’t to suggest that valuations must fall; rather, that the real value created through AI products can finally rise to meet, or even surpass, those valuations.
This is what I call the unbiasing pendulum of entrepreneurship: the recognition that growth without cost is an illusion. It’s a pendulum because collective opinion is swinging to the opposite extreme of what was once the unquestioned norm. It’s unbiasing, not because it overcorrects, but because in its natural motion, it passes through both poles (growth and cost) giving the entire system its balance and weight.
The unbiasing process means restoring cost as a fundamental variable in value creation because there are no absolute returns, only relative ones defined by the price we pay to achieve them.
But we should be cautious. As Hannah Arendt warned in The Human Condition, when efficiency becomes an end in itself, we risk forgetting why we sought to innovate in the first place. That brings to mind two final things: the Knobe effect and the parable of the Mexican fisherman.
The Knobe effect describes how moral evaluation shapes our perception of intentionality. When a side effect is morally good, we tend to judge it as unintentional; when it’s bad, we judge it as deliberate. Joshua Knobe attributed this to moral reasoning; others have linked it to psychological bias or the implicit trade-offs between profit and principle. The takeaway: our moral framing influences how we interpret actions, especially those justified by “efficiency.”
Nicholas Georgescu-Roegen, the economist mentioned earlier, argued that the true product of the economic process (subject to entropy) is the pleasure of living. Without purpose and pleasure, he claimed, we cannot speak meaningfully of economy. If the answer to “why efficiency?” isn’t grounded in ethics and human well-being, then perhaps we’re doing it wrong. And that brings us to the parable of the Mexican fisherman:
A boat docked in a tiny Mexican village. An American tourist complimented the Mexican fisherman on the quality of his fish and asked how long it took him to catch them.
“Not very long,” answered the Mexican.
“But then, why didn’t you stay out longer and catch more?” asked the American. The Mexican explained that his small catch was sufficient to meet his needs and those of his family. The American asked, “But what do you do with the rest of your time?”
“I sleep late, fish a little, play with my children, and take a siesta with my wife. In the evenings, I go into the village to see my friends, have a few drinks, play the guitar, and sing a few songs ... I have a full life.”
The American interrupted, “I have an MBA from Harvard and I can help you! You should start by fishing longer every day. You can then sell the extra fish you catch. With the extra revenue, you can buy a bigger boat.
“And after that?” asked the Mexican.
”With the extra money the larger boat will bring, you can buy a second one and a third one and so on until you have an entire fleet of trawlers. Instead of selling your fish to a middle man, you can then negotiate directly with the processing plants and maybe even open your own plant. You can then leave this little village and move to Mexico City, Los Angeles, or even New York City! From there you can direct your huge new enterprise.”
“How long would that take?” asked the Mexican.
“Twenty, perhaps 25 years,” replied the American.
“And after that?” the Mexican asked.
“Afterwards? That’s when it gets really interesting,” answered the American, laughing. “When your business gets really big, you can start selling stocks and make millions!”
“Millions? Really? And after that?”
“After that you’ll be able to retire, live in a tiny village near the coast, sleep late, play with your children, catch a few fish, take a siesta with your wife and spend your evenings drinking and enjoying your friends.”
The unbiasing pendulum of entrepreneurship can be a reminder for some of us that progress without reflection risks becoming motion without meaning and perhaps the real measure of innovation is not how fast we move, but how consciously we account for the cost of the journey.