Picture a family in a mid-sized American city on a Saturday afternoon. They carry home a new 55-inch television that cost less than a nice dinner out did a few years ago. That same week, a letter from the landlord arrives. The rent is going up again, and the increase alone is larger than the price of the TV. On Monday, a routine visit to urgent care produces a bill that dwarfs both.
Nothing about this family is unusual. They are living inside two price curves at once. One curve keeps falling, and it carries screens, phones, software, solar panels, and now artificial intelligence. The other keeps rising, and it carries the bills that decide whether a household feels secure: rent, health care, college, and child care. Understanding why those curves split apart is the first step toward bending the expensive one down.
The Curve That Keeps Falling
In 1936, an aircraft engineer named Theodore Paul Wright published a paper in the Journal of the Aeronautical Sciences called “Factors Affecting the Cost of Airplanes.” He had noticed a steady pattern. Each time the total number of airplanes built doubled, the labor needed for each new plane dropped by a predictable share. Cost, he found, follows experience rather than the calendar.
That observation became known as Wright’s Law, and it describes almost every manufactured product that is built at scale. Workers learn shortcuts. Engineers redesign parts. Suppliers find cheaper materials.
Solar, Batteries, and Screens
Solar power is the textbook case. A watt of solar module capacity cost about $96 in 1976, measured in today’s dollars. By 2024 it cost about $0.26. Lithium-ion battery packs tell a similar story. BloombergNEF’s annual survey found average pack prices of roughly $790 per kilowatt-hour in 2013 and about $115 in 2024.
Consumer electronics ride the same curve. According to Bureau of Labor Statistics price data compiled by economist Mark Perry of the American Enterprise Institute, the price of televisions fell about 98 percent between January 2000 and late 2025. Computer software fell about 75 percent. Toys fell about 74 percent, and cell phone service fell about 43 percent.
AI Joins the Club
Artificial intelligence is now on the fastest version of this curve yet. The Stanford AI Index 2025 reports that the cost of querying a model performing at the level of GPT-3.5 fell more than 280-fold between November 2022 and October 2024. A task that once cost real money per request now costs a fraction of a cent.
That matters because AI is not a product sitting on a shelf. It is a tool that can be applied to other work. If cheap intelligence can reach the expensive side of the economy, it could do something TVs never could. It could lower the price of services, not just gadgets.
The Curve That Keeps Rising
Now look at the other half of Perry’s chart, which he has called the “chart of the century.” Over the same stretch from 2000 to late 2025, overall consumer prices rose about 92 percent. Hospital services rose about 275 percent. College tuition and fees rose about 196 percent. Child care and nursery school rose about 185 percent. Medical care overall rose about 129 percent, and housing rose about 111 percent.
Average hourly earnings rose about 131 percent over the same period. That number explains a great deal. Wages beat general inflation, so the typical hour of work buys more of most things than it did in 2000. Yet hospital care, tuition, and child care rose faster than wages. For those bills, an hour of work now buys less than it did a generation ago.
Why These Bills Hurt More
The categories that got cheaper are mostly things people buy occasionally. A family replaces a TV every several years. The categories that got expensive are paid every month or come in large, unavoidable lumps. Rent arrives on the first. Child care is often the second-largest bill a young family pays. A single hospital stay can erase a year of savings.
So averages can mislead. A household that rents, has small children, and carries a health condition lives almost entirely on the rising curve. Cheaper electronics are pleasant. They do not cover the gap.
What Happened to Wages
Wage stagnation is real, but the details matter. The Bureau of Labor Statistics tracks median usual weekly earnings for full-time workers, adjusted for inflation. In constant 1982 to 1984 dollars, that figure was $332 in 1979, $334 in 2000, and $374 in 2025. That is growth of about 13 percent over more than four decades, and almost all of it arrived after 2014.
Economists debate how to read this. Some point out that average hourly earnings, total compensation including benefits, and different inflation measures show stronger gains. Others note that rising health insurance premiums have absorbed much of what employers spend on workers, so take-home pay grew slowly even when total compensation grew faster. Both views can be true. Either way, the typical worker gained modestly, while the price of the biggest essentials rose sharply.
Why the Expensive Curve Will Not Bend on Its Own
The rising categories share a few traits. They are services rather than goods. They cannot be imported. They depend on land or on licensed people. And they often face limited competition. Each trait points to a specific cause.
These causes do not belong to one political party. They are structural, which means they respond to structural fixes.
Baumol’s Cost Disease
In the 1960s, economists William Baumol and William Bowen studied live performance. A string quartet in 1800 needed four musicians and about the same amount of time as a string quartet today. Productivity did not rise. Yet musicians must be paid wages that compete with factories and offices where productivity did rise. So the price of a concert climbs.
Teaching, nursing, and child care work the same way. One caregiver can safely watch only so many toddlers. A nurse can attend only so many patients per shift. When wages rise across the economy, these labor-heavy services get more expensive even if nothing about them changes. That is not a failure. It is arithmetic.
Land, Zoning, and Gatekeeping
Housing has a different problem. A home is a building plus the land under it, and land in high-demand cities is scarce by definition. Zoning rules that ban apartments on most residential lots, require large minimum lot sizes, or mandate parking make that scarcity worse. When cities limit what can be built, rising demand turns into rising rent.
Health care and higher education add licensing and gatekeeping. Licensing protects quality, but it also limits who may perform a task and how many new providers can enter. Hospital mergers have reduced competition in many regions. Accreditation and prestige make it hard for cheaper colleges to win students. Each barrier may have a reason. Together they keep prices high.
Can AI Bend the Expensive Curve?
Here is the honest answer. In some places, yes. In others, only if policy changes first. The difference depends on whether the cost comes from labor and paperwork, which software can shrink, or from land and permission, which software cannot create.
Hype in either direction does damage. Promising that AI will fix rent creates cynicism. Dismissing it entirely leaves real savings on the table.
Where AI Plausibly Helps
Several costs on the rising curve are made of paperwork and routine cognitive work, which is exactly what cheap AI does well.
- Tutoring. A capable AI tutor can give one-on-one practice to students whose families could never afford a private tutor, easing one source of the Baumol problem in education.
- Health care administration. Billing, coding, prior authorization, and documentation consume a large share of clinician time and hospital budgets. Automating that work could free staff for patients.
- Design and permitting. Software can check building plans against codes, speed reviews, and help standardize designs, cutting months from projects where delay itself costs money.
None of these replaces a nurse at a bedside or a teacher in a room. They shrink the overhead around those people, which is where much of the cost has quietly grown.
Where AI Cannot Help Without Policy
No algorithm can create land in a city where zoning forbids building. A faster permit review means little if the answer is still no. AI can draft a perfect apartment plan, but only a city council can make that plan legal on most lots.
Credentialing works the same way. If a task requires a specific license, cheaper software does not lower the price until rules allow trained people, supported by good tools, to do more of the work. In these areas, technology is ready before the law is. Savings arrive only when reform catches up.
Building Abundance That Reaches Every Household
Michael Ten’s mission at IdeaRiff is abundance for all, and the two price curves show why that goal needs more than one tool. Cheaper technology handles half the job. Reforms in housing, health care, and education handle the other half. Without both, families keep buying cheaper TVs while falling behind on rent.
There is also a role for income floors. A basic income, delivered through trusted community channels such as credit unions and nonprofit networks, can give households stability while slower reforms take hold. Delivery through local institutions keeps money close to the people it serves. Ownership matters as well. When communities can own and use open tools, the savings from cheap AI flow to them rather than collecting with a few platforms.
What to Push For
Readers who want the expensive curve to bend can support concrete, nonpartisan steps that already have backing across the political spectrum.
- Zoning reform that legalizes apartments, duplexes, and accessory dwellings near jobs and transit.
- Faster, digital permitting with clear timelines.
- Scope-of-practice updates that let nurse practitioners, pharmacists, and other trained professionals do more.
- Price transparency and competition review in hospital markets.
- Open educational tools and lower-cost credential paths.
- Pilots of income floors run through credit unions and nonprofits.
None of these requires picking a side. Each simply removes a barrier between people and cheaper essentials.
What to Watch For
The best test is simple. Track the next annual update of the BLS price data. If AI is truly reaching the rising curve, hospital services, tuition, and child care should start growing more slowly than wages. If they do not, the savings are being captured somewhere along the way.
Watch local decisions too. City council votes on housing and state votes on licensing will shape your budget more than any new gadget. The TV will keep getting cheaper on its own. The rent will not.
This article is for educational purposes only and is not financial advice.



