Podcast source material. The complete text of three Hemant Pandey Blog articles is provided below. WordPress AI can use the full corpus to select, combine, shorten and shape the eventual podcast conversation.
1. When the Ladder Disappears
For generations, society has offered a simple bargain: learn, work, earn, save, and rise. The bargain was never equal, and it never worked for everyone. But it gave ordinary people something enormously important: a believable ladder.
Artificial intelligence may be doing something more consequential than taking jobs. It may be removing the rungs of that ladder.
When intelligence becomes cheap
Look at what has happened to intellectual production. Writing, research, translation, coding, analysis and design can increasingly be produced in seconds. An illustration that once required a professional brief, revisions and potentially hundreds of dollars can now be generated almost instantly.
The important change is not simply that AI makes things cheaper. It changes what is scarce. When competent intellectual output becomes abundant, the ability to produce it stops being a strong economic advantage.
Attention becomes the bottleneck
At the same time, human attention has not become more abundant. We can produce more information than ever, but nobody has more hours in the day.
Why read a 2,000-word article when you can ask an AI to summarize it? The old internet rewarded the person who could create information. The emerging internet may reward the system that can decide which information is worth your thirty seconds.
The physical ladder may follow
For a while, one could imagine a simple escape: if machines take cognitive work, humans will still have physical work. But increasingly capable robots are attacking that boundary too.
The combination is what matters. AI can make cognitive labour cheaper while robotics can make routine physical labour cheaper. The question then changes from “Which jobs will AI replace?” to something much harder:
What economic function remains for ordinary humans when both brain and muscle can increasingly be rented as machines?
Society starts at Floor 10
That is why I keep returning to the image of a building with ten floors.
Historically, technology removed some work from the bottom and created new work above it. A person could start with little, learn, gain experience, become skilled, and eventually move upward.
AI may be different. It can attack several floors at once. If routine cognitive work disappears while routine physical work is also automated, society may increasingly expect people to enter the economy near the top: as exceptional creators, entrepreneurs, owners, decision-makers or people with rare human capabilities.
But how does an ordinary person get to Floor 10 if the lower floors were the training ground?
The dangerous part is not unemployment
Unemployment is measurable. Loss of aspiration is harder to see.
Modern societies have tolerated remarkable inequality partly because people continue to dream. A poor person can believe that education may change their child’s future. A junior employee can believe that experience will eventually make them senior. A small entrepreneur can believe that today’s effort may create tomorrow’s independence.
The social contract is therefore not simply “you get a job.” It is “there is a path from where you are to somewhere better.”
If that path becomes visibly implausible, inequality changes character. It is no longer merely unequal outcomes. It becomes a society in which the ladder itself appears to belong to somebody else.
The wage-day paradox
Imagine a darkly comic future: 80 percent of society has no economically necessary job. The remaining 20 percent still receive wages. On wage day, the 80 percent are waiting outside.
“Welcome home,” they say. “We are here for peaceful distribution.”
It is a joke, but it exposes a serious problem. If machines produce most of the wealth, why should access to that wealth continue to depend entirely on having a job?
The problem eventually stops being production. Machines may be able to produce enormous abundance. The problem becomes ownership and distribution.
The freedom paradox
There is another strange inversion.
A person with little money can walk down a public street, sit on a beach and disappear into a crowd. A famous and extremely wealthy person may have almost unlimited purchasing power but need security, controlled travel and privacy measures simply to do ordinary things.
Money buys options. It does not necessarily buy freedom.
Perhaps one of the most valuable things in an unequal society is something money cannot easily purchase: anonymity.
The Joker’s warning
There is a line in The Dark Knight that captures the fear of social breakdown: “When the chips are down, these civilized people, they’ll eat each other.”
The point is not that economic inequality inevitably produces violence. It does not. Societies can redesign institutions, redistribute gains, expand public services and create new forms of meaningful participation.
The warning is about fragility. A social order depends on people believing that the system offers them a future worth protecting.
The knife does not decide
A knife does not decide how it will be used. A technology does not automatically determine the society built around it.
But powerful tools amplify human choices. The real risk is not “evil AI.” It is extraordinary capability combined with concentrated ownership, extreme inequality and institutions that fail to adapt.
The same competitive pressure and pursuit of profit that helped create astonishing technology can also push firms toward ever greater automation. What is rational for one company may become destabilizing when everyone does it at once.
First slowly, then suddenly
Technological disruption may not arrive as one dramatic event. A company may simply hire fewer juniors. A team may shrink through attrition. A department may be consolidated. A workflow may become one person plus an AI system.
Each decision can look reasonable. Taken together, they can transform the structure of employment.
That is why “first slowly, then suddenly” is such a useful warning. The important transition may not be the day millions are fired. It may be the day millions of young people realize that the entry-level jobs from which careers used to begin simply aren’t coming back.
What is left?
If intelligence becomes cheap and physical labour becomes increasingly automatable, perhaps the scarce things will be elsewhere: attention, trust, judgment, agency, genuine relationships, verified experience, ownership and meaningful participation in the real world.
That does not mean humans become useless. It means the definition of useful human activity may have to change.
And there is a profound choice hidden inside this technological revolution.
AI could become the greatest productivity multiplier in human history. If its gains are broadly shared, it could reduce drudgery and create extraordinary abundance.
Or the same technology could make the existing ladder disappear while leaving ownership and distribution largely unchanged.
The question is no longer simply whether AI can do our jobs.
The question is what kind of society we build when human labour is no longer the main source of economic value—and whether we can preserve something even more important than employment: dignity, agency, purpose and freedom.
2. Raw Intelligence: Why Sell Courses When You Can Rent a Mind?
The next great AI market may not be artificial intelligence. It may be human intelligence, scaled.
We have spent the first phase of the AI revolution asking how intelligent machines can become. The next phase may force us to ask a more fundamental economic question: what happens when AI allows individual human beings to scale, preserve and monetize their intelligence?
Today, expertise is mostly sold through time and packaged knowledge. A teacher sells a course. A consultant sells hours. A lawyer sells advice. A scientist sells research. An entrepreneur writes books or gives lectures.
The common limitation is simple: the expert is one person. Even the world’s greatest teacher cannot personally teach millions of students.
AI potentially changes that constraint. Instead of selling only the products of intelligence—books, courses, lectures and consulting—we may eventually sell access to the intelligence behind them.
I call this idea Raw Intelligence.
1. The Intelligence Bottleneck
Human expertise is accumulated slowly. A great teacher may spend 30 years solving problems, observing students, discovering misconceptions and developing better ways to explain difficult concepts. Much of that expertise never makes it into a textbook.
The valuable asset is not merely what the teacher knows. It is also how the teacher approaches a problem, what they notice first, which mistakes they recognize immediately, which explanation they choose, which shortcut they trust, when they abandon an approach, and how they adapt an explanation to different learners.
This is tacit expertise. Today, much of it disappears when the expert stops teaching. AI creates the possibility of preserving and distributing at least part of it.
2. From Knowledge to Thinking
A conventional book primarily transfers knowledge. Raw Intelligence attempts to transfer something closer to ways of thinking.
Consider a mathematics teacher. A textbook can explain the quadratic formula. A course can demonstrate several examples. But an exceptional teacher may look at a student’s wrong solution and immediately understand: “You don’t actually have a problem with this formula. You have misunderstood the relationship between these two quantities.”
That diagnosis is expertise. The teacher may then explain the concept using a completely different example because they know that the first explanation will not work for this particular student.
Capturing such patterns is much more valuable than simply uploading lecture notes. Raw Intelligence therefore aims to preserve knowledge + reasoning + heuristics + experience + judgment + methodology.
It is not a literal copy of someone’s biological brain. It is a licensed computational representation of their distinctive expertise.
3. The Mind Book
Imagine an expert voluntarily giving an AI company access to a carefully prepared intellectual corpus.
Call it a Mind Book.
It could contain writings, lectures, solved problems, explanations, problem-solving methods, conceptual frameworks, examples, teaching strategies, common mistakes, decision rules, responses to difficult cases, and other material the expert chooses to contribute.
The foundation model supplies general-purpose intelligence. The Mind Book supplies the distinctive intellectual contribution of a particular expert.
The proposition isn’t simply that “AI knows mathematics.” It is that AI can help you think through mathematics using the methodology of this particular expert.
4. Rent a Great Mind
Now imagine a marketplace for Mind Books.
A student has a difficult mathematics problem. Instead of buying a ₹5,000 course, the student pays a few rupees—or a small monthly subscription—to access the Raw Intelligence of an exceptional teacher.
The student asks: “I understand the formula, but I don’t understand why this substitution works.” The AI responds using the teacher’s methodology. The student asks another question. The system adapts. It gives another example. It identifies the student’s misconception. It challenges the student.
The student is effectively renting access to the teacher’s accumulated expertise. The same model could apply to physics, engineering, programming, medicine, law, finance, writing, design, business and countless specialized fields.
5. Why Sell Courses When You Can Rent a Mind?
This changes the economics of expertise.
A course is essentially a frozen representation of what an expert knows. A Mind Book could be interactive.
A course says: “Here is what I decided to teach you.”
Raw Intelligence says: “Ask me what you need to know.”
A student doesn’t necessarily need 100 hours of lectures. They may need one excellent answer to one difficult question at the moment they are stuck. The expert’s intellectual asset becomes available precisely when it is needed.
6. When One Expert Becomes a Million
The fundamental economic breakthrough is scaling.
Suppose an exceptional teacher can personally teach 500 students. Their time is the bottleneck. Now imagine a Mind Book capable of assisting 500,000 students.
The teacher has not literally become 1,000 teachers. Their intelligence has become scalable.
AI isn’t necessarily replacing the teacher. It is becoming a distribution system for the teacher’s expertise.
A previously local expert could potentially become globally accessible. A specialist whose market was too small to support a conventional business could suddenly have a worldwide audience. A retired expert could continue contributing to education long after retirement.
7. Who Gets Paid When AI Thinks?
This is where the idea becomes difficult—and interesting.
Suppose an AI generates an answer using general model knowledge, Expert A’s reasoning, Expert B’s specialized knowledge, and information supplied by the user. Who gets paid?
Traditional copyright isn’t designed for this problem. Copyright is relatively good at identifying copying. AI may instead reproduce patterns of reasoning without reproducing the original words.
That requires a new concept: Cognitive Attribution.
The objective isn’t necessarily to pretend that every answer can be divided into mathematically perfect percentages. Instead, the system could establish auditable measures of material cognitive contribution.
If a licensed expert’s distinctive intellectual asset materially contributes to a commercially valuable response, the expert participates in the resulting revenue.
For example, a hypothetical attribution system might record: Foundation intelligence — 70%; Expert Mind Book A — 20%; Expert Mind Book B — 10%. Those numbers are illustrative, not a proposed technical standard. The principle matters more than the exact formula.
If AI scales someone’s distinctive intellectual contribution, that person should share in the economic value created by that scaling.
8. Cognitive Auditing
This immediately raises another question: who decides how much the expert contributed?
The AI company should not necessarily be the sole judge of its own royalty bill. A neutral auditing system could therefore emerge.
Government could establish the rules, just as it establishes tax and accounting laws. Independent cognitive auditors could verify provenance, licensing, attribution methodology, usage records, royalty calculations and compliance.
The analogy with financial reporting is straightforward: Government → establishes the rules; AI company → maintains usage accounts; Independent auditor → verifies them; Expert → receives royalties.
The government could collect ordinary taxes on the resulting economic activity, and potentially a legally defined public levy where appropriate.
The important principle is separation of roles. Government regulates. Auditors verify. Platforms operate. Markets determine value. Experts own or license their intellectual assets.
9. The “Ask Directly” Button
There is another essential feature. Raw Intelligence must know when it doesn’t know.
Suppose the AI encounters a genuinely unusual question and cannot satisfactorily reproduce the expert’s reasoning. Instead of hallucinating, it offers:
ASK DIRECTLY
The user can request the actual expert.
The interaction becomes: AI → insufficient confidence → human expert → answer.
This creates an elegant division of labour. AI handles the enormous volume of ordinary questions. The human handles exceptional questions. The human’s scarce time becomes a premium service rather than the entire business model.
With appropriate consent, the expert’s new answer could potentially be incorporated into the Mind Book. That creates a feedback loop:
Expert → AI → User → difficult question → Expert → improved AI
10. The New Economics of Education
Consider the implications for a student who cannot afford an elite teacher.
Today, access to exceptional expertise is often constrained by money + geography + time + availability.
A Mind Book could potentially reduce all four constraints. A student in a small town could access the methodology of an exceptional teacher in another country. The cost could be a few dollars rather than thousands. The interaction could occur at midnight or during a five-minute break.
The student doesn’t need to consume an entire course. They can ask the exact question that is blocking their progress.
This doesn’t eliminate human teachers. Education is much larger than information transfer. Motivation, discipline, social development and human mentorship remain important. But Raw Intelligence could dramatically reduce the scarcity of high-quality intellectual assistance.
11. Cognitive Access
This suggests a new category of product.
We currently sell books, courses, lectures, consulting, tutoring and expert hours. The AI economy could add:
Cognitive Access
The customer doesn’t buy everything the expert has ever produced. They buy the ability to interact with an AI representation of that expertise whenever they need it.
That is a shift from selling knowledge products to selling access to expertise.
It may also shift expertise from a labour model to an asset model.
Today: Expert → time → income
Tomorrow: Expert → intellectual asset → usage → royalty
12. Raw Intelligence Is Not Artificial Intelligence
This distinction matters.
The foundation model provides general intelligence. Raw Intelligence provides specialized human intelligence.
The foundation model is the engine. The Mind Book is the specialized cognitive layer.
A useful analogy is computing. A processor provides general computational capability. Software turns that capability into something specialized and useful.
Likewise: Foundation AI + Mind Book = specialized cognitive capability
As generic AI becomes increasingly abundant, distinctive expertise may actually become more valuable, not less. The scarce resource could shift from computational intelligence to judgment, experience, intuition, methodology and provenance.
13. The AI Social Contract
One possible future is: human beings create knowledge → AI absorbs it → AI companies capture the value → creators are economically displaced.
That creates resentment. The creator can reasonably ask: “If my intellectual work helped build the system that replaces me, where is my share?”
Raw Intelligence proposes a different arrangement:
Human creates → AI scales → users benefit → creator receives royalties.
The objective is not to prevent AI from becoming powerful. It is to create an economic mechanism through which humans participate in the value created by scaling their intelligence.
14. The Sarah Connor Problem
If society comes to believe that AI has taken our knowledge, absorbed our expertise, replaced our jobs and captured the resulting wealth, then resistance to AI could become intense.
The conflict would not necessarily begin because people hate intelligent machines. It could begin because people believe the machines have confiscated the economic value of human intelligence.
Raw Intelligence offers a different narrative: AI does not have to steal human intelligence. It can distribute it.
The teacher can earn from teaching millions. The specialist can reach a global market. The retired expert can continue contributing. The creator retains an economic relationship with the intellectual asset.
15. From Expert to a Million Experts
Today: One teacher → hundreds of students
A course: One teacher → thousands of students
A Mind Book: One teacher → potentially millions of interactions
Ask Directly: The teacher remains available for exceptional cases
Royalty attribution: The teacher participates economically in the resulting value
The teacher’s most valuable asset is no longer simply the number of hours they can work. It is the quality and distinctiveness of the intelligence they have accumulated.
AI provides the scale.
16. The Intelligence Economy
This leads to a simple proposition:
AI should scale intelligence, not confiscate it.
The first generation of AI focused on making machines intelligent. The next generation could focus on making human intelligence economically scalable.
That produces a radically different relationship between humans and AI.
- Raw Intelligence — My intelligence can help millions of people.
- Apprentice Mind — AI helps me develop my own intelligence.
- Augmented Mind — AI extends my existing intelligence.
These are not necessarily three technological versions of the same thing. They are three different relationships between humans and AI.
Raw Intelligence is the starting point because it creates the economic infrastructure for human expertise to be preserved and scaled.
And the most provocative question is therefore no longer:
“Will AI replace the expert?”
“What if AI allows the expert to become a million experts?”
3. The AI Business-Model Paradox: Who Does AI Really Work For?
Technology has always had a complicated relationship with the customer.
The customer wants a product that lasts, works well and solves a problem as completely as possible. The company needs revenue. Usually, these interests overlap. Sometimes they don’t.
When they diverge, business models can create incentives that are not necessarily aligned with the customer’s interests.
We have seen versions of this before: planned obsolescence, upgrade cycles, attention-maximizing social media, recurring subscriptions and business models built around repeated consumption.
AI could create a much more powerful version of the same problem.
Because unlike a light bulb, a phone or a social-media feed, an AI agent can decide how the interaction itself unfolds.
1. From manufactured obsolescence to manufactured consumption
Consider the light bulb.
In the 1920s, the Phoebus cartel brought major manufacturers together and worked toward standardizing bulb lifespans around 1,000 hours. The economic logic was straightforward: a longer-lasting bulb needed to be replaced less frequently.
This became one of the classic examples associated with planned or manufactured obsolescence.
The broader principle is more important than the bulb:
If a company earns money when something is consumed or replaced, consumption can become part of the business objective.
The same principle appears in different forms elsewhere. A social-media company earns from attention. A subscription company earns from continued subscriptions. A company selling upgrades benefits when customers buy the next generation.
None of this means every company deliberately makes its products worse. It means that economic incentives influence product design.
What if intelligence itself becomes a consumable?
2. The AI version: tokens
An AI system performs computation every time it processes a request. Depending on the architecture, that computation can involve tokens, model calls, tool calls, searches, retrieval, memory operations and other forms of inference.
The user has a simple objective:
Get the job done.
Ideally: one request, one efficient solution, finished.
But an agent could instead turn the same task into a sequence of clarification, planning, analysis, alternatives, confirmation, tool calls and revisions. Sometimes every step is justified. Sometimes it isn’t.
And that distinction will become economically important.
The user wants minimum time, minimum effort and minimum cost for the maximum useful result. The provider may have legitimate reasons to optimize for revenue, reliability, retention and customer satisfaction. Those objectives can align, but they don’t automatically align.
That is the AI business-model paradox.
3. The “planning instead of doing” problem
This may become one of the most irritating forms of AI failure.
Ask an agent: “Prepare this report.”
Instead of preparing it, the agent explains how it intends to prepare it. Then it proposes a workflow. Then it asks which style you prefer. Then it asks another question. Then it produces an outline. Then it asks whether you want to proceed.
Twenty minutes later, you have had an impressive intellectual conversation. You still don’t have the report.
This produces a critical distinction:
The appearance of cognitive activity is not the same as useful work.
An agent can generate enormous amounts of reasoning, planning and explanation. But the user’s objective is not to watch intelligence happen. The objective is to finish the job.
This suggests a future metric that may be more meaningful than conversational quality:
Tokens per completed outcome.
How much computation did the system consume to accomplish the task? Not how long was the conversation. Not how impressive was the explanation. But how efficiently did the AI accomplish what the user actually wanted?
4. Question inflation
An intelligent agent should ask questions when information is genuinely necessary.
But there is a difference between necessary clarification and avoidable interaction.
A capable system can often make reasonable assumptions. Instead of asking what format you would prefer, it can choose the most appropriate format. Instead of asking whether it should proceed when the instruction is already clear, it can proceed. Instead of asking the user to perform five intermediate steps, it can perform them itself.
The danger is not that asking questions is inherently bad. The danger is that interaction itself can become a product.
If every additional turn consumes computation, and if computation has an economic value, then unnecessary interaction has an economic consequence.
5. The AI dark pattern is different
Traditional dark patterns manipulate the interface. A button is confusing. Cancellation is difficult. A subscription is hidden behind a default.
An AI agent can move the dark pattern inside the conversation.
It can potentially influence what it asks, what it recommends, what it prioritizes, what it delays, which tools it uses, how many alternatives it presents, when it asks for confirmation and when it declares the task complete.
This creates a particularly difficult problem. The user may not see the manipulation. They may interpret it as intelligence.
The next generation of dark patterns may look like exceptionally good UX.
6. The agent has two masters
An AI agent appears to work for the user. But the agent is operated by a company.
The user may want the best answer. The company may also have interests in its own model, cloud, products, ecosystem, subscription revenue, usage and commercial relationships.
These interests don’t necessarily conflict. But when they do, who determines the agent’s objective?
That question becomes much more important as agents acquire the ability to act. An ordinary search engine can influence what you see. An agent could potentially influence what you do.
7. Six potential AI dark patterns
Recommendation bias: You ask which product to buy. The agent could recommend products based partly on commercial relationships rather than purely on your interests.
Service-selection bias: The agent decides which search engine, cloud service, model, payment provider, application, website or marketplace to use. If the agent’s owner has commercial interests in one of those services, the selection process becomes economically significant.
Interface capture: If AI becomes the primary interface to computing, users may stop directly interacting with browsers, search engines, applications and websites. The agent becomes the gateway.
Manufactured friction: The agent could potentially make certain actions easier and others harder. The danger is not necessarily outright blocking. It is asymmetric convenience.
Dependency creation: A useful AI should ideally make the user more capable. But a commercial AI provider can also benefit when the user becomes deeply dependent on its ecosystem.
Interaction monetization: If the provider earns from inference, every additional interaction has potential economic value. The extreme version would be to keep the conversation going.
8. The healthcare analogy
There is a useful precedent in healthcare economics.
A 2018 Goldman Sachs report famously asked whether curing patients was a sustainable business model in the context of certain one-time genetic treatments. The point was not that doctors or hospitals don’t want patients to recover. The economic issue was that a one-time cure has a fundamentally different revenue profile from a treatment that patients need repeatedly.
That distinction matters for AI too.
Imagine two agents. Agent A uses 10,000 tokens and completes the task. Agent B uses 100,000 tokens, requires 15 interactions and produces roughly the same outcome.
From the user’s perspective, Agent A is obviously preferable. If the provider’s economics reward inference consumption, however, there is a potential conflict.
A completed task is a lost inference.
The question is whether the industry will reward agents for efficient outcomes or merely for valuable consumption.
9. The alternative: manufacture better products, not friction
There is another business model.
Instead of trying to maximize lifetime extraction from each customer, entrepreneurs can compete by delivering dramatically better products.
Imagine a company that combines commodity processors, inexpensive AI accelerators, large memory, open-source software, open or inexpensive models, mass manufacturing and AI-assisted engineering.
It sells an excellent product at a relatively modest margin. It captures volume. Then it builds something substantially better the following year.
That is serial entrepreneurship.
The customer is not trapped. The customer is given a reason to upgrade.
Manufactured obsolescence: Make the old product less attractive so the customer has to buy the new one.
Genuine innovation: Make the new product so much better that the customer wants it.
The second model creates a much healthier competitive environment.
10. China + open AI + commodity compute
This is where China’s manufacturing ecosystem becomes particularly interesting.
A new company may not need to build an entire technology stack from scratch. It can combine commodity hardware, inexpensive manufacturing, open models, open-source software, AI-assisted development and modular compute to build a complete product with a relatively small team.
The entrepreneur doesn’t necessarily need to become the next Microsoft, Apple or Google. The strategy could be simple: build the best configuration possible today, sell it at a modest margin, capture volume and build a better configuration next year.
This is almost the opposite of manufactured obsolescence.
It is manufactured improvement.
11. The OS may become less important
For decades, computing has looked roughly like this:
Hardware → OS → applications → browser → Internet → user
Windows, macOS, Android and iOS became enormously valuable partly because they controlled the middle of this stack.
AI potentially changes the architecture.
The user says: “Find this.” “Write this.” “Analyze these files.” “Buy this.” “Fix this.” The agent decides which applications, websites, APIs, databases or services are necessary.
The user doesn’t necessarily care which application performs the task. The application becomes a capability. The website becomes a service endpoint. The browser becomes infrastructure. The operating system becomes infrastructure.
The agent becomes the interface.
12. The personal AI computer
The logical endpoint is not necessarily another traditional operating system.
Imagine a commodity personal AI computer built from compute, large memory, a local AI model, local data, Internet connectivity and an agent.
Underneath it could run Linux. But the user doesn’t really care.
Linux Max Pro Ultra — Reloaded
The joke captures the concept. The product isn’t really Linux. The product is personal intelligence.
13. Intelligence should remain available offline
A truly personal AI computer should not require a permanent cloud connection.
Offline, a local model could handle writing, documents, coding, calculations, local search, personal knowledge, file organization and private analysis.
When something genuinely requires the Internet, the agent connects. For example, today’s price requires current information. An analysis of your own local documents does not inherently require the Internet.
Intelligence is always available. The Internet is optional.
14. But “local AI” must mean more than local inference
Suppose you ask your local AI: “Analyze my entire private archive and find patterns.” The local model may not be powerful enough. A cloud model may be better.
That is perfectly reasonable—if the user controls the decision.
The computer should say: “I can do this locally, but a cloud model would perform a deeper analysis. Sending these 37 documents, totaling 340 MB, is required. Proceed?”
That is very different from silently uploading the data.
Local first. Cloud optional. User controlled.
Model updates can come into the machine. Private data should not automatically go out of it. When cloud computation is genuinely useful, the user should know what is being sent, where it is going, why it is needed and what will happen to it.
The important distinction is therefore not simply local AI versus cloud AI. It is user-controlled AI versus provider-controlled AI.
15. The two possible AI futures
Future A: AI feudalism
A small number of companies could control models, compute, agents, identity, payments and distribution. The agent could become the gateway through which people interact with the digital world.
That would create potential conflicts around recommendations, service selection, commercial influence, friction, dependence and computation.
Future B: Commodity personal AI
Compute becomes cheaper. Models become increasingly open. Hardware becomes modular. Manufacturing becomes competitive. Small entrepreneurial teams become extraordinarily productive. Local models handle ordinary tasks. Cloud models are used when they genuinely provide additional value. Users control their data.
Companies compete by making the next product better, not by making the existing product worse.
16. The real AI business-model paradox
AI promises to eliminate friction. But the companies operating AI may have incentives to preserve some forms of friction.
AI promises to empower users. But companies may benefit from user dependence.
AI promises to make software invisible. But whoever controls the agent could become an even more powerful software gatekeeper.
AI promises abundant intelligence. But whoever controls the intelligence layer could potentially control access to the abundance.
So the central question is not whether AI will make technology better. It probably will.
The deeper question is:
Who will AI optimize for?
The user—or the organization that owns the AI?
The answer should not be left entirely to the business model.
We should design systems so that the economic interest of the AI provider and the interest of the user are as closely aligned as possible. That means measuring what actually matters: cost per completed task, time per completed task, user control over data, ease of migration, interoperability and transparency of recommendations.
And perhaps most importantly:
How often did the AI actually do the job instead of merely talking about doing the job?
The best AI is not the one that consumes the most intelligence. It is the one that delivers the most useful intelligence for the least unnecessary consumption.
And the best AI business may not be the one that extracts the maximum value from every customer. It may be the one that creates so much value for customers that they voluntarily come back for the next generation.
Don’t manufacture obsolescence.
Don’t manufacture friction.
Don’t manufacture dependence.
Manufacture something better.