When my parents got married back in the 1980s, the standard wedding gift was a dinner service or a canteen of cutlery, something from Wedgwood or Royal Doulton or Arthur Price, and a serious set of the kind a couple would keep for forty years and pass on to their children, which cost a few hundred pounds. The average UK house in 1985, by way of comparison, sold for about £34,000. If you do the math, that means it took roughly 70 to 85 dinner sets to buy a house.
Today, you can walk into Pottery Barn and buy a beautiful 12-piece dinner set for around $300, while the median US home in 2025 sold for $426,800. That works out at something like 1,400 dinner sets per house, which is to say that houses are now 15-20x more expensive relative to dinner sets than they were when my parents got married. The plates are essentially the same. The houses are essentially the same. The ratio between them has been totally transformed, and it isn't just plates.
The pattern is everywhere
A 1980 colour television cost somewhere between $400 and $600, which is around $1,800 in today's money. Today, $500 will buy you a 65-inch 4K QLED with more pixels than the human eye can resolve at a normal viewing distance. The US Bureau of Labor Statistics, when it adjusts its data for quality, finds that television prices have fallen roughly 6.5% per year, every year, since 1950, which means that on a constant-quality basis the 1980 equivalent of a $650 TV would cost about $6 today.
An IBM PC XT in 1985 cost $4,395, which is around $10,700 in today's money for a machine with a 10-megabyte hard drive and two floppy drives. The phone in your pocket has roughly ten million times more storage and a million times more computing power than that XT, and it costs less than a tenth of the price in real terms.
Clothing, toys, software, appliances, furniture; the same pattern shows up everywhere. Anything that can be manufactured at scale, shipped on a container ship or digitally delivered to you and sold into a competitive market has been getting cheaper, in real and often nominal terms, for my entire adult life.
There is a famous chart from the American Enterprise Institute, sometimes called the "Chart of the Century," which plots the price change of major consumer goods and services from 2000 to 2024 against an overall inflation rate of roughly 74%. The picture splits cleanly into two clusters. The first is everything that has gone up faster than wages, led by hospital services at +220%, college tuition at +178%, medical care at +130%, childcare at +115%, food at +82%, and housing at +80%. The second cluster is everything that has gone down in absolute terms over the same period, which includes televisions, software, toys, cellphone service and clothing.
Goods are getting cheaper. Services are getting more expensive. The line between the two is almost perfectly the line between what you can put on a container ship/email and ship from anywhere in the world and what requires a human, in the room, doing it for you.
Why this happens
There is a name for this, and I didn't actually know it until today, when I started researching this. It is called Baumol's cost disease, after the economist William Baumol who described it in the 1960s, and it is one of those ideas that, once you see it, you cannot unsee.
The argument runs like this. Some sectors of the economy enjoy enormous productivity gains over time, in the sense that a factory worker today produces ten times what they did in 1980, and a software engineer can ship infrastructure that would have taken a team of fifty a generation ago. Wages in those sectors rise, because more productive workers are simply worth more.
But there are other sectors of the economy that don't get those productivity gains, and never will. It still takes one teacher to teach a classroom. It still takes four musicians nine minutes to play Beethoven's String Quartet in C minor, exactly as it did in 1820. It still takes a surgeon several hours to perform a hip replacement, and however good the technology assisting them becomes, the surgeon can only do so many in a day.
And yet those workers, the teachers and surgeons and musicians and nurses and carers, also need to earn a living, and they compete for talent against the productive sectors of the economy. Their wages have to keep pace with everyone else's, even though their productivity hasn't moved. The price of their services therefore rises and rises, decade after decade, while the price of everything that can be made more efficiently keeps falling. That is why a college degree costs three times what it did in 2000 while a TV costs a fifth of what it did. Same economy, opposite trajectories, one mechanism quietly running underneath.
The second piece of the puzzle is competition. When a manufacturer figures out how to make a product more cheaply, they have a choice between keeping prices the same and banking the margin, or dropping prices and capturing market share. Most of them try to bank the margin first, but sooner or later a competitor or a new entrant decides to forgo the margin and chase market share instead, because money seeks risk-adjusted returns, and an underpriced product earning a fair return is more attractive than a fat-margin product that is about to be undercut. This is the engine that Schumpeter called creative destruction, and it is, more than anything else, the reason goods get cheaper over time.
What AI does to this picture
The question I have been sitting with this week is what happens when AI makes my own business meaningfully more competitive, not by 10%, but by something like a step-change.
I can already see it coming. Faster product design, tighter supply chains, a leaner SG&A line because the work that used to require five team members can be done by one with the right set of tools. My costs go down on both lines of the income statement, COGS as well as SG&A. My first instinct, naturally, is to think great, more profit. But Schumpeter and a hundred years of competitive markets tell me that more profit is only the temporary state of affairs. The minute my unit economics improve, somebody, whether that's me or my competitor or some new entrant who hasn't yet shown up, will trade margin for share, and the savings will end up in the customer's pocket rather than mine. My customer will demand it, in any case, because they are in a competitive market, and they will have to pass savings on too.
If you multiply this across the entire economy, AI doesn't just make one business more efficient. It makes most of the goods-producing economy, and a meaningful slice of the services economy, more productive all at once. The whole cost curve drops.
This is why I don't think AI is going to make us all rich, in the way the headlines like to suggest. Money chases returns, competition compresses margins, and consumers capture the surplus, with the result that AI's most likely first-order effect is not a small group of people getting fabulously wealthy but everything, everywhere, getting cheaper. Which, if you squint, has the same end effect as a pay rise, in the sense that you can buy more with the same money, but the distribution is completely different. It is a pay rise for the entire population, paid in the form of falling prices.
A standard-of-living shift
A century ago the idea of having a private flushing toilet in every home was a luxury rather than a right, and a hundred years before that the idea of light at the flick of a switch, available at any time in any room, was practically unthinkable. Today we take both for granted, to the point that most of us would consider their absence a violation of basic human dignity.
So I find myself wondering what people in fifty or a hundred years will consider a basic standard of living that today would seem absurd. A personalised tutor for every child, available 24/7, who knows their entire learning history and adapts to the way they learn. A medical AI that diagnoses you long before symptoms appear. A bespoke physical product, designed for your exact dimensions, manufactured and delivered the same week. All three are technically feasible today, and all three are luxury goods. In two decades, with deflation compounding the way it has for televisions, all three will be effectively free.
That is the optimistic version of the AI story, and I think it is rather more likely than the dystopian one.
Where the money goes
But there is a catch buried in all of this, and it is where Baumol comes back, and where I think the most interesting investing and career questions of the next decade actually live.
If AI makes goods cheaper, and makes the easily-automatable parts of services cheaper too, then by definition the things that don't get cheaper become relatively more valuable. The irreducibly human stuff. A meal cooked for you by a chef who knows your wife's allergies and your kid's birthday. A live concert. A great therapist. A wedding photographer who actually saw the moment. A coach who pushes you. A dinner party with people you love.
We are heading into a world where things will be effectively free and where experiences, particularly experiences mediated by skilled humans, will be the genuine luxury. The early signs are already in the data. Restaurant prices have outpaced grocery prices by a wide margin. Live event prices have outpaced recorded music (Taylor Swift vs Spotify). The fastest-growing category of high-end consumer spending isn't cars or watches anymore, it's travel and dining.
That, in the end, is what my parents' cutlery is telling me. They got a beautiful set of plates as a wedding gift in the 1980s because plates were genuinely expensive then, and a meaningful inheritance to pass on to the next generation. If you got married next year, the equivalent gesture probably wouldn't be plates at all. It might be sending the couple to a meal at a restaurant they couldn't otherwise afford, with people who matter to them, on a night they will remember for the rest of their lives.
The plates are commodities now. The night out isn't.
That is the trade AI is making for us, whether we asked for it or not. And I for one think it is bloody marvellous.
Sources: Nationwide House Price Index; National Association of Realtors Q3 2025 metro home prices; AEI "Chart of the Century" (Mark J. Perry); BLS CPI data; NBER working paper "Measuring Moore's Law" (2018); William Baumol, "Performing Arts: The Economic Dilemma" (1966).