When Intelligence Gets Cheaper, Decisions Become More Costly
Updated: 12 hours ago
You can sink three-pointers one after another in an empty gym.
But in the final two seconds of a game, the score, clock, defense, fatigue, teammates, and crowd pressure all bear down on you at once. The issue is no longer whether you can get the ball into the basket.
It is whether you can recognize if taking the shot is the right decision.
That is exactly where business is heading in the AI era.
Knowledge is not disappearing. Expertise is not losing its value. But access to information, generating options, and preparing an initial analysis are becoming cheaper rapidly. Competitive advantage is shifting from simply knowing more to turning what you know into the right decision, at the right time and in the right context.
Will companies become NBA teams?
Joe Hudson, who works with OpenAI research teams and technology
executives, predicts that AI will make companies increasingly resemble NBA teams: flatter structures, smaller rosters, and greater economic leverage behind each employee’s decisions.
In Hudson’s analogy, the valuable player is not the one who memorizes the most plays or makes the most three-pointers in practice. It is the one who makes the right choice in the final two seconds, stays calm under pressure, and reads teammates, opponents, and the mood of the game.
When AI increases a person’s productive power, it does not amplify only good decisions. It also amplifies hesitation, misplaced priorities, and friction within teams. Hudson therefore groups the distinguishing capabilities of the future around judgment, difficult conversations, turning failure into learning, and maintaining clarity under pressure. He explains this approach in his June 2026 essay ‘The New Inner Game’.
The point of the NBA analogy is not that everyone must be a superstar.
More than which player has the ball, what matters is the team knowing why it is moving the ball in that direction.
In finance, the equivalent is concrete.
A system can calculate cash flow, generate currency scenarios, measure debt repayment capacity, or model company value. These are the equivalent of making three-pointers.
But identifying the truly critical assumption, distinguishing temporary from structural risk, understanding how an intervention affects another part of the business, and recognizing what management can still change are about reading the game.
Two accounts absent from the balance sheet: Pace and spin
Hudson monitors how teams work through two concepts: pace and spin.
Pace describes how quickly an idea meets reality, is tested, and becomes new knowledge.
Spin describes movement without progress: hesitation, repetition, delays driven by fear, and energy lost to friction.
In Hudson’s definition for his own team, pace concerns how fast work advances and how much available capacity is used, while spin means effort people believe is wasted. He describes how these indicators make team problems visible in his own explanation.
Translating these concepts into financial language produces an interesting picture.
Financial pace is the speed at which a signal moves from corporate data to a scenario, from a scenario to a challenged decision, and from a decision to a measurable result.
Financial spin is the energy spent repeatedly reconciling files, preparing the same data again, defending assumptions of unclear origin, waiting for approval, and producing reports that ultimately change no decisions.
A finance team can therefore work very fast while maintaining a slow decision pace.
Reports may arrive on time, spreadsheets may be updated, and meetings may take place. But if new information does not change the scenario, the scenario does not create alternatives, and alternatives do not reach a timely decision, the wheel is turning while the company stands still.
Some spin comes from systems; some comes from human behavior.
If an employee fears delivering bad news promptly, questioning an established assumption, or challenging a project backed by their manager, critical information can remain outside the model. Amy Edmondson’s study of 51 manufacturing teams found that psychological safety was associated with learning behavior, which mediated team performance. Safety here means neither low standards nor constant agreement; it means errors, doubts, and objections can become visible in time. The study’s findings are available here.
In finance, late bad news is often more expensive than the bad news itself.
Sometimes success comes not from being smarter, but from not celebrating too soon.
A similar lesson comes from how AI models operate.
Poolside CEO Eiso Kant explains that improvements in the Laguna S model did not come only from greater raw capacity. According to the research team, distinguishing behaviors included more verification, fewer assumptions taken for granted, not declaring success prematurely, backtracking when needed, and continuing to search for a solution. In the interview where Kant discusses these observations, sustained performance cannot be explained by model size alone.
We should not confuse people and machines here.
AI does not feel anxiety, show courage, or feel ashamed of failure. But a model can be developed to accept its first answer, verify it, backtrack, or try another route.
People and models intersect not because they share emotions, but because both need a working discipline beyond raw capacity.
The most knowledgeable person does not always make the best decision.
The largest model does not always produce the most reliable result.
Sometimes the difference is not how quickly you find an answer, but what you do with it once you have found it.
Ainsideme’s place in the game
For Ainsideme, the central task is not producing more financial output, but increasing decision pace by reducing financial spin.
This is what unites our five digital applications, from cash flow and risk to valuation, trade and project finance, export transactions, and cross-border structures:
Not merely processing data, but turning it into a financially meaningful question.
Not giving a single result, but making its underlying assumptions and alternatives visible.
Not treating the first answer as final, but testing whether it remains valid under changing conditions.
A cash forecast can balance mathematically while misreading collection behavior.
A valuation model can work perfectly in technical terms while missing the economic driver that supports the value.
A financing alternative can look cheaper while making the company more vulnerable when exchange rates, maturities, and debt-service effects are assessed together.
The real question is therefore not the model’s score.
It is how accurately that model can read the game when the company takes the court.
The strongest finance teams of the future will not be those producing the most spreadsheets or the fastest answers.
They will be those achieving a higher decision pace with less spin, refusing to mistake the first result for the final result, and reading the game again at critical moments.
Because when intelligence is distributed more evenly across the court, the same thing still creates an advantage:
Not just knowing where to pass the ball, but being able to explain why you passed it there.
Comments