1. The Intelligence Illusion
Humanity has accumulated intelligence for millennia. Books preserved it, universities organised it, the internet made it widely accessible, and Google made it searchable. Artificial intelligence has taken the next step: it can synthesise information, explain concepts, identify patterns and generate new possibilities through ordinary conversation.
Each technological advance has reduced the cost of accessing and processing knowledge. Yet the ability to make money has not become equally accessible. A person can now obtain a business plan, a marketing strategy, a trading explanation or a software prototype in minutes. What remains difficult is turning any of these into a reliable source of income. The plan may be coherent, the strategy may be familiar, and the prototype may work. None of these establishes that customers will pay, that costs can be controlled or that the activity can survive competition.
We are witnessing an extraordinary reduction in the cost of intelligence. We should not confuse it with an equivalent reduction in the difficulty of economic success.
Intelligence is cheap. The ability to convert it into money is invaluable.
2. A Brilliant Poem Can Be Worth Zero Dollars
A deeply moving poem may have enormous aesthetic value and earn nothing. A remarkable painting may remain unsold. A magician can astonish an audience without creating anything the audience can use outside the performance. Intellectual excellence, artistic merit and the capacity to generate wonder do not automatically create a paying market.
AI can reproduce this distinction at an unprecedented scale. It can generate essays, images, music, designs and explanations that appear impressive. But the fact that something has been created does not establish that anyone needs it, wants to buy it or considers its price justified.
Four separate questions are involved. Can something be produced? Does it have value to someone? Is there sufficient demand for it? Can the producer capture enough of that value to earn a return? We see an impressive demonstration and assume economic value must follow. Yet the distance between producing something remarkable and building a business around it can be enormous.
A magic trick works because the audience experiences wonder. A commercial enterprise must do something more: turn an experience, capability or product into an exchange that can be sustained.
3. The 160-IQ Petrol-Pump Worker
A person with an IQ of 160 can work at a petrol pump. Another person with an ordinary IQ can build a highly profitable business. These possibilities do not establish that intelligence is irrelevant; they establish that intelligence alone does not determine economic outcomes.
Opportunity, capital, temperament, social circumstances, practical skills, focus, judgment and execution all influence what a person can accomplish. Someone may possess extraordinary analytical ability yet struggle to select one objective and pursue it long enough to produce a result. Someone else may have less intellectual range but be exceptionally good at recognising a market opportunity, organising resources and serving customers consistently.
The economy does not pay a universal premium for intelligence. It rewards particular outcomes under particular conditions. When AI makes high-level intellectual assistance widely accessible, access to an explanation or business plan can no longer be assumed to provide a lasting competitive advantage.
4. Intellectual Buddhi Vilas: When Thinking Becomes Its Own Reward
For some highly divergent thinkers, the pleasure of discovering an idea can compete with the slower work of completing a project. One question leads to another, one connection opens ten more, and the mind continually finds new territory to explore. The intellectual activity is genuine, sometimes brilliant, but the original task remains unfinished.
Call this intellectual Buddhi Vilas: thinking that becomes rewarding in itself, independently of whether it produces a practical result. The appeal lies in discovery, connection and possibility. Completion may offer less immediate stimulation than beginning another investigation.
AI can intensify this pattern. Every answer generates further questions; every idea can become an essay, a framework or a new project. Exploration becomes almost frictionless, while implementation still requires decisions, repetition, compromise and the willingness to stop exploring.
This is not a universal consequence of high intelligence, nor is intellectual exploration worthless. The problem arises when the reward of thinking repeatedly displaces the work required to turn a selected idea into something useful. The economy does not pay for every possibility we discover. It pays, when it pays at all, for what we manage to bring into existence and deliver.
5. The Golden Goose: Why Profitable Knowledge Is Not All Public
Business families often accumulate practices over generations that never appear in textbooks or public interviews. These may include how to judge a supplier, negotiate credit, recognise a genuine bargain, assess a customer’s reliability, manage cash flow during a crisis or know when an apparently attractive opportunity should be rejected.
Some of this knowledge is deliberately protected as a trade secret. Some is transmitted through apprenticeship and observation. Some is tacit: people know how to do something without being able to describe every judgment involved. A successful business may disclose its products and general strategy while keeping the precise methods that sustain its advantage private.
If profits depend on a repeatable advantage, revealing that advantage may help competitors eliminate it. Not every profitable method is secret, and secrecy alone does not create success. But commercially useful knowledge is not necessarily distributed in proportion to its value.
AI can learn from published accounts of successful businesses without automatically acquiring the private practices, relationships and accumulated judgment behind them. Even when a successful method is visible, its underlying logic may not be identifiable from the result alone. Several different processes can produce the same outcome. A finished product is evidence of decisions, but it is not a complete record of why those decisions were made.
6. The Trading Paradox: Knowledge Is Abundant, Success Is Scarce
Thousands of books explain market behaviour, hundreds of courses teach strategies, and enormous quantities of financial information are publicly available. Yet access to this knowledge does not guarantee profitable performance.
SEBI’s study of individual equity derivatives traders illustrates the gap. Across FY2022–FY2024, 7.2% of individual traders were net profitable over the three-year period, while 1% earned more than ₹1 lakh in net profit after transaction costs. The latter figure is a specific profit threshold, not the proportion of all traders who made any profit. It concerns equity derivatives, not every form of investing or trading.
The finding does not establish why traders fail. But it demonstrates that the availability of information and strategies is not sufficient to make profitable outcomes common. A trader must identify a genuine edge, account for costs, size positions appropriately, manage risk and execute under uncertainty. A strategy that worked in historical data may fail in live markets; one that succeeds in one environment may stop working when conditions change.
AI can explain a strategy, analyse historical prices and help write testing software. None of those activities, by itself, proves that a durable trading advantage exists. The strategy must survive realistic costs, out-of-sample testing and the conditions under which it will actually be used.
7. Those Who Teach May Have a Different Business
The familiar saying “Those who can, do; those who can’t, teach” is an unfair generalisation about teachers, but it raises a useful question when applied to people selling direct routes to financial success.
A trader earns from market outcomes. A trading coach earns by teaching. A seller of an unsubstantiated promise earns by persuading people that a method will work. These are different businesses with different conditions for success. If someone possesses a reliable and scalable method of making money, why sell it to others rather than use it? There may be legitimate answers: the method may require more capital, teaching may scale better, or the person may simply prefer education. But where the seller’s main evidence of success is revenue from selling a method, the method’s actual results deserve independent scrutiny.
A course can sell well without producing profitable traders. A coaching programme can generate revenue without delivering every promised transformation. This is the logic behind snake oil: selling a promise more convincing than the evidence supporting it. Not every coach or educator is a fraud; the issue is whether claims match results.
AI introduces a related economic separation. An AI provider can earn money by selling access to intelligence even when a customer fails to turn that intelligence into income. The provider’s revenue and the user’s outcome are separate variables. Selling knowledge is one business. Using knowledge successfully is another.
8. The Fragrance of the Flower
The secret of making money may resemble the fragrance of a flower. You can dissect the flower, identify its components and study the conditions in which it grows. Yet the recipe for its fragrance may not be apparent from the dissection.
The analogy is not that fragrance has no chemical explanation, or that commercial success is inherently mysterious. It is that identifying the components does not necessarily reveal the effective combination, proportions and conditions that produce the whole result.
A successful business may combine timing, customer psychology, trust, pricing, relationships, tools, risk tolerance, experience and execution. Each element can be studied separately, but the commercial outcome may depend on interactions that are difficult to isolate or reproduce.
A fee card can be inspected for typography, layout, pricing options and visual hierarchy. General design principles can suggest improvements. But the visible card may not reveal the hidden commercial reasoning behind its structure: why options were chosen, what customer behaviour they are intended to influence, or which practical constraints shaped the design. The person who developed the system may possess information the artefact itself does not disclose. Without it, several plausible explanations may fit the same visible result.
The ingredients of success can be visible while the synthesis that makes them effective remains difficult to reproduce.
9. Fine-Tuning: Why the Best Optimisation May Not Be the Best Solution
AI is good at optimisation when the objective, constraints and relevant variables are sufficiently specified. But real life is not a perfect statistical model. Information is incomplete, circumstances change, and important considerations may be difficult to measure.
A business may deliberately hold extra inventory because a supplier is unreliable. A trader may accept lower expected returns to reduce the risk of ruin. A teacher may limit enrolments to preserve quality. These choices can appear inefficient against a narrow metric while making sense within the larger system.
The issue is not that optimisation is bad, or that AI cannot reason about trade-offs. It is that the objective being optimised may be incomplete. Maximising one measurable quantity can damage another that matters more in practice.
Fine-tuning configures a system for a particular person, task, environment and set of constraints. It may require changing parameters, but also changing the objective, relaxing a constraint or accepting an imperfection because the supposedly superior alternative is impractical. A fee-card design can be optimised for visual clarity and still fail commercially. The correct configuration depends on its purpose, customers, business positioning and pricing logic.
The challenge is not merely to optimise a system. It is to discover what to optimise, what not to optimise, and which compromises make the system work.
10. Subjective Experience: The Daily Grind
This daily grind is what we call subjective experience: trying, failing, adjusting, negotiating, recognising exceptions and learning which compromises work. It is the accumulated history of dealing with reality as it behaves rather than as a model predicts it should behave.
A businessperson may spend decades learning when to extend credit, which customers to trust, how to respond to a supplier’s failure, when to raise prices and when to reject an opportunity. Much of this knowledge is never written down. Some is difficult to articulate because it has become embedded in habits of attention and judgment. Its value becomes apparent when a particular situation demands a decision.
AI can learn from recorded experiences, analyse operational data and improve through feedback. It can help people discover patterns faster and make tacit assumptions more explicit. The limitation is not that AI can never acquire practical knowledge. It is that it cannot automatically reconstruct experiences that were never captured, observe conditions it cannot access or infer every relevant adjustment from the final result.
Experience also teaches what to ignore. A model may identify a statistically interesting pattern irrelevant to the decision at hand. An experienced operator may recognise that a signal is misleading in current circumstances, even when the reason is difficult to formalise.
The recipe for success may therefore exist not in a document but in the history of decisions, mistakes and corrections that teaches someone what to do next.
11. Economics as a Hard Test of Intelligence
A mathematical theorem can be assessed against a defined standard. An economic opportunity is often less clearly specified. The problem may be unknown, the objective incomplete, the relevant information hidden and the environment changing. Competitors can respond to success, eroding the advantage that produced it.
A generated business plan is not a business. A proposed trading strategy is not a proven edge. A functioning software product is not necessarily commercially viable. Each is a candidate solution that must encounter reality.
A serious test of AI’s economic capability would ask whether it can independently discover an opportunity, assemble a workable system, operate within a defined budget, attract genuine customers, deliver something useful, cover its costs and sustain positive returns over a meaningful period. Results would need independent verification, with human assistance and risks clearly accounted for.
It would be premature to claim that economics is universally harder than the Millennium Prize Problems. They test different capabilities. But open-ended economic success is demanding because the answer is not simply a statement to be verified. It is an outcome that must be produced and sustained in a world that does not stand still.
12. The Scarcity That Remains
As intelligence and information become abundant, the scarce capability may be the fine-tuning that turns them into economic results. It requires selecting the right problem, combining public knowledge with private and tacit experience, configuring tools for a particular situation, accepting necessary imperfections and executing consistently.
AI can assist at every stage. It can search, reason, analyse, design, simulate and help people learn from feedback. These capabilities can create substantial economic value. But access to intelligence does not guarantee a durable commercial advantage, just as access to trading knowledge does not guarantee profitable trading.
The distinction is not between intelligence and some mysterious force beyond intelligence. It is between having information and building a system that works under real constraints. Sometimes the missing ingredient is capital; sometimes distribution, relationships, timing, discipline or specialised knowledge. Often it is the interaction of several factors that cannot be reduced to a universal recipe.
Humanity has spent millennia accumulating intelligence. The internet made it searchable, and AI has made it easier to synthesise and apply. The next challenge is not simply to produce more intelligence. It is to understand how intelligence becomes effective action in particular circumstances.
Intelligence can generate possibilities. Subjective experience teaches us how reality resists them. Fine-tuning turns what we know into something that works—and economic value emerges only when that result meets a real need under real constraints.
Further reading
Intelligence Is Not Economic Value: The Missing Link Between Knowledge and Wealth
The Actionable Intelligence Problem: Does AI Need Superintelligence or Practical Wisdom?
The Thinking Calculator: Why AI Is a Tool, Not an Intellectual Proprietor