Will AI Make Your Team Less Useful, or Free It for Work It Never Had Time For?

History suggests AI may expand the work a small business can take on, not leave its people idle. Here is how Northeast Indiana owners can test that idea.

Will AI Make Your Team Less Useful, or Free It for Work It Never Had Time For?

Probably not: history suggests AI will more likely expand what a small business can take on than leave its people idle, if the owner chooses how that capacity gets used.

That is the central argument of a Sunday episode of The AI Daily Brief, in which host NLW walks through a lengthy essay by Venkatesh Rao about fears that artificial intelligence will make human minds redundant. This post turns that argument into questions a Northeast Indiana owner can actually test against the work the team already does every week.

Key Takeaways

  • Much of the fear about AI concerns relevance, not only hacking or lost jobs, and the episode argues that history offers a more hopeful precedent.
  • Rao calls the modern version of the old zero-sum worldview cognitive mercantilism, and he maps its four beliefs onto intelligence.
  • Demand has historically grown as goods became cheaper, and the same pattern could apply to the demand for intelligence.
  • Freed capacity only matters if it turns into real new work, such as requests your team could not answer before.
  • Under a draft-and-approve setup, an AI teammate handles the reading and drafting while a person keeps the final decision.
  • The destination may be good, but the transition still costs time and trust, so managing it carefully matters more than resisting it.

Why does AI feel like a threat to relevance?

Most people who worry about artificial intelligence are not only concerned about security breaches or lost jobs, although both fears are real and widely discussed in every industry. The deeper worry concerns relevance, because when an ability that people long considered uniquely human becomes abundant and inexpensive, many wonder what their own skills are still worth.

For a small business, that fear often appears as a manager who wonders whether an experienced estimator still has a meaningful role, or as a receptionist who suspects the front desk is about to disappear entirely. Those worries are rarely stated openly, yet they shape how willing people are to experiment with new tools, and they can quietly slow adoption long before any written policy exists.

NLW argues that this anxiety has a historical precedent. He suggests that four centuries of economic change point toward a far more hopeful outcome than the headlines imply, and that precedent deserves careful testing rather than acceptance on faith.

What is cognitive mercantilism, and why does it matter?

Cognitive mercantilism is Rao's label for the belief that intelligence, much like gold in the early modern era, is scarce, finite, and safe only in the hands of a few credentialed holders. Rao builds the label on the historian Joyce Appleby's book The Relentless Revolution: A History of Capitalism, which describes the economic worldview that existed before capitalism took hold in England. That older worldview treated wealth as land, precious metals, and the labor of largely impoverished people.

Under that older system, the central goal of statecraft was to control territory and accumulate bullion, while ordinary people were regarded as inputs to production rather than as customers or inventors in their own right. The surplus they generated flowed upward to the elites who commanded their labor, and nobody imagined that their demand could drive a larger economy.

Rao argues that today's fears about AI repeat that structure with remarkable fidelity. He lists four beliefs that defined the old order and maps each one onto intelligence. The first treats knowledge as a fixed stock, so national and corporate strategy becomes a race to attract the best minds and to own whatever they produce.

The second holds that ordinary people cannot be trusted with powerful tools, and that intelligence is safe only with those already certified to hold it. The third focuses almost entirely on controlling supply while ignoring demand from everyone outside the elite. The fourth assumes most people are good for little beyond routine labor, so automating their work simply creates what Rao calls "useless mouths to feed."

The same pattern appears in much smaller settings. A business that keeps its procedures inside one employee's head is hoarding intelligence in miniature, because that knowledge leaves the company whenever that person does. Owners who treat know-how as bullion tend to guard it rather than share it, and that guarding often limits what the whole team can accomplish together.

What did the old zero-sum view get wrong about demand?

The old view got demand badly wrong, and that error matters for a business today. Mercantilist thinkers assumed that the economy held a fixed amount of wealth and a fixed appetite for goods, so one nation's gain necessarily came at another's expense. Capitalism disproved that assumption gradually, and the disproof began on farms.

Agricultural productivity shrank the share of people who had to farm from a large majority to a small minority in advanced economies. That shift did not leave the rest of the population idle or useless. It produced abundant food, and it freed people to take up trades, services, and occupations that the earlier economy could never have supported at any scale.

Rao makes the point personally, and the personal version is memorable. He admits he "can barely keep a houseplant alive," yet he can eat a far wider variety of food than an early modern peasant could have imagined. His argument is that most people will likewise no longer need to produce intelligence in order to consume it well and benefit from it.

The useful question for an owner is therefore not how much intelligence a team can produce on its own. The useful question is how much intelligence the team can put to work once that intelligence becomes cheap enough to reach ordinary tasks. That reframing moves attention away from headcount and toward unmet demand, which is where real growth opportunities usually sit.

Could demand for intelligence grow the way food demand did?

Yes, according to the argument, because demand for intelligence has no obvious ceiling once its cost falls. NLW revisits his May episode, titled The New Jobs AI Will Create, in which he argued that human demand is far more elastic than job-loss forecasts tend to assume.

Those forecasts quietly presume that people already demand as much as they ever could, a view that NLW calls completely ahistorical. He sorts the ways latent demand gets unlocked into six dimensions: price, access, complexity, continuity, personalization, and relational trust, and each one deserves a closer look.

Each dimension describes a customer who wanted something but could not obtain it at an affordable price, through a convenient channel, or in a form they could actually use. Price means the customer wanted the service but it cost too much. Access means they could not reach it at all. Complexity means the system was too confusing for them to navigate without frustration.

Continuity means the customer would benefit from help all the time but receives it only occasionally. Personalization means a generic version exists, yet something made specifically for them is what they truly value. Relational trust means they want the experience to feel more human, more responsive, and more dependable than the alternatives they have tried.

For a small business, that list reads like a catalog of unmet customer needs that most owners already recognize. Consider a heating and cooling company in the Fort Wayne area whose office cannot answer after-hours questions, or a family clinic in Auburn whose intake forms confuse patients and send them to competitors. Each situation points to complexity or continuity demand that a cheaper, more available form of intelligence could serve.

Whether that demand becomes real revenue depends on the business itself, its pricing, and how well it serves the people who ask. Unmet demand becomes an opportunity only when the owner can respond quickly and accurately, in a way that preserves the customer's trust.

Why would people work harder when intelligence gets cheaper?

Rao points to early modern Britain, where tea, sugar, tobacco, and colorful Indian cottons created new desires among ordinary households that had never before been able to afford such goods. As those goods spread through the economy, annual working days rose from about 250 to 300 before the Industrial Revolution arrived. Much of that increase was voluntary, because people wanted what the extra income could buy them.

Productivity produced surplus income, surplus income sustained a growing demand for consumer goods, and that demand encouraged further gains in productivity. Rao describes that cycle as virtuous, and he suggests that AI could bring a similar pattern to cognitive work. People may willingly expend more effort once a much larger range of possibilities becomes within reach.

NLW adds an important caveat that keeps the argument honest. His own idea of an infinite backlog explains the anxiety of working more, because it describes a pile of work with no natural stopping point. It does not fully explain the excitement that makes people want to keep building long after the workday should have ended.

Both feelings can be true at the same time, and an owner who recognizes both is better positioned to manage them. That means deciding in advance how much of the new capacity goes to new work, and how much goes to rest, training, and the quality of the work already in hand. Capacity without boundaries tends to become exhaustion.

What else was in the November 1 episode?

Beyond the main essay, NLW spends considerable time on the infinite backlog, an idea he developed in earlier work and revisits here. In his account, office work has always contained more tasks than any person could finish, and what counts as done has been a negotiation between that backlog and natural limits such as the hours in a day.

AI systems that can run tireless parallel copies of a worker break that limit in a new way. The backlog becomes reachable, and the only remaining constraint is the imagination required to decide what belongs on it. That framing helps explain why some owners feel busier after adopting AI, even when the tools genuinely save time on existing tasks.

Rao extends the argument further by suggesting that AI is not only a supply technology. Many products and services already exist mainly because the market demands them, with humans serving as secondary consumers. The systems that AI helps build, along with the systems built around them, can create new pathways and needs for intelligence that did not exist before.

Rao also asks whether many people avoid deep exploration because they lack the capacity to explore, or because they lack the surplus time and justification to explore. He answers that the second explanation is more likely. Finally, NLW closes with a caution: even a good transition carries real costs, and managing those costs is a different task from trying to stop the change altogether.

What does an AI teammate do with this idea?

An AI teammate is one practical way to turn cheaper intelligence into new capacity, but only when the owner sets the rules before the work begins. Hey Button runs on OpenClaw, an open automation framework, with models such as Claude and GPT answering through it. Each teammate is configured around real workflows, and the owner chooses one of three access levels: read-only, draft-and-approve, or full access.

Under draft-and-approve, the teammate can read incoming requests, organize the relevant details, and draft a reply for review. Nothing goes out until a person approves it. That default fits the episode's logic well, because the owner keeps the judgment while the teammate absorbs the reading and drafting work that would otherwise sit in the backlog.

Consider an illustrative case. A prospect asks for a quote on a spring cleanup. The teammate organizes the request, identifies the missing details such as the address and photos, and drafts a reply asking for them. A person reviews that draft before anything reaches the customer, which keeps the customer relationship in human hands.

A model switch does not change the teammate's instructions, the business facts it was configured with, or the access boundaries the owner set. Customer messages, spending, and sensitive changes stay behind the approval rules the owner chooses. A reader who wants to see which work should stay with people can continue with what a person can do that an AI teammate cannot, then review what to hand to AI and what stays with the owner.

What should a small business do first?

Start with the customer need rather than the technology itself. Pick one workflow where demand goes unmet because the team is too busy, such as a backlog of estimates, slow replies after business hours, or intake forms that patients abandon halfway through. Then write down what the AI may read, what it may draft, and what always returns to you for approval.

Next, run that workflow on real examples before it touches a customer. Track whether new inquiries appear and whether requests that used to go unanswered now get handled in a timely way. Watch whether your people spend the freed time on judgment, relationships, or craft that only they can provide, because that is where the transition either succeeds or quietly fails.

This approach is not right for every business. A shop with steady, well-served demand may reasonably choose to use freed time for quality or rest rather than for growth. The essay describes possibilities rather than guarantees, so an owner should judge the claim against their own numbers rather than against a historical trend.

A short intake conversation is a practical first step. Start the Hey Button questionnaire and describe the task you would hand off first. Setup and ongoing support are quoted around your workflows and integrations, and any software subscriptions or usage charges are identified before you commit, with no surprise line items.

Sources

  • The AI Daily Brief, "AI Won't Make Our Minds Useless According to 400 Years of History," November 1, 2026 (episode, show notes, and transcript): https://aidailybrief.ai/e/2026-11-01
  • Venkatesh Rao, "Cognitive Mercantilism" (essay discussed in the episode).
  • Joyce Appleby, The Relentless Revolution: A History of Capitalism (book cited in the essay).
  • NLW, "The New Jobs AI Will Create," The AI Daily Brief, May 2026 (earlier episode cited in the episode).

Frequently Asked Questions

Not necessarily. Agriculture and early industry both moved people toward judgment, relationships, and craft as routine production got cheaper. Identify which skills your team uses to earn customer trust and protect time for those tasks first.
It is the assumption that a fixed amount of work exists, so every task a machine completes is one fewer task for a person. NLW argues this ignores demand elasticity, because cheaper goods and services often create new demand that nobody had planned for.
Track what customers ask for before and after a change, not only the hours saved. Look for new inquiries, requests that now get answered, and jobs your team used to decline for lack of time. Rising numbers mean freed capacity is turning into demand.
By default, no. Under draft-and-approve, the teammate reads requests, organizes details, and drafts replies, but a person approves each message before it goes out. Broader access can come later, once real examples show the rules still fit.
The Malthusian worry holds that population growth will outrun food supply, leaving most people in subsistence. The essay compares that fear to AI doom talk, which imagines intelligence outrunning human capacity to absorb it. Both assume demand is fixed.
Will AI make the skills on my team worthless?
Not necessarily. Agriculture and early industry both moved people toward judgment, relationships, and craft as routine production got cheaper. Identify which skills your team uses to earn customer trust and protect time for those tasks first.
What is the lump-of-labor fallacy?
It is the assumption that a fixed amount of work exists, so every task a machine completes is one fewer task for a person. NLW argues this ignores demand elasticity, because cheaper goods and services often create new demand that nobody had planned for.
How can I tell whether freed time creates real new work?
Track what customers ask for before and after a change, not only the hours saved. Look for new inquiries, requests that now get answered, and jobs your team used to decline for lack of time. Rising numbers mean freed capacity is turning into demand.
Should an AI teammate send messages without approval?
By default, no. Under draft-and-approve, the teammate reads requests, organizes details, and drafts replies, but a person approves each message before it goes out. Broader access can come later, once real examples show the rules still fit.
What is the Malthusian worry, and why does it come up here?
The Malthusian worry holds that population growth will outrun food supply, leaving most people in subsistence. The essay compares that fear to AI doom talk, which imagines intelligence outrunning human capacity to absorb it. Both assume demand is fixed.

Hey Button

Want a teammate that handles this for you?

Answer a few questions about your business and Hey Button will show you what an AI teammate would take off your plate, with you approving anything that goes out.

Start the questionnaireSee what teammates do
Lucas M. Button

Written by Lucas M. Button

Founder, Hey Button

Lucas builds AI teammates for small businesses across Northeast Indiana and writes about what works, what doesn't, and what to hand off first. More about Lucas