| season | rocks moved | arms | lent against |
|---|---|---|---|
| spring | 412 | 1,236 | 1,236 |
| summer | 398 | 1,194 | 1,194 |
| autumn | 431 | 1,293 | 1,293 |
| winter | 377 | 1,131 | 1,131 |
| season | rocks moved | arms | lent against |
|---|---|---|---|
| this afternoon | 0 | 0 | — |
Intel Corporation Santa Clara, California
To Busicom Corp., Tokyo
Re Return of development fee, MCS-4 (4004) calculator chip set
For Rights to sell the chip set for non-calculator uses
$60,000
Dated May 1971
Homebrew Computer Club Newsletter 1976
An Open Letter to Hobbyists
…As the majority of hobbyists must be aware, most of you steal your software. Hardware must be paid for, but software is something to share. Who cares if the people who worked on it get paid?
Is this fair? … Who can afford to do professional work for nothing? What hobbyist can put 3-man years into programming, finding all bugs, documenting his product and distribute for free?
Bill Gates
General Partner,
Micro-Soft
3 February 1976
Vague but exciting…
CERN DD/OC March 1989
Information Management: A Proposal
This proposal concerns the management of general information about accelerators and experiments at CERN. It discusses the problems of loss of information about complex evolving systems and derives a solution based on a distributed hypertext system.
From: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds) Newsgroups: comp.os.minix Subject: What would you like to see most in minix? Date: 25 Aug 91 20:57:08 GMT Hello everybody out there using minix - I'm doing a (free) operating system (just a hobby, won't be big and professional like gnu) for 386(486) AT clones. This has been brewing since april, and is starting to get ready. I'd like any feedback on things people like/dislike in minix, as my OS resembles it somewhat [...] PS. Yes - it's free of any minix code, and it has a multi-threaded fs. It is NOT protable (uses 386 task switching etc), and it probably never will support anything other than AT-harddisks, as that's all I have :-(.
Each dates the other.
Name one, and ask whether anyone has drawn an exchange rate for it yet.
reading
—
web sites found
Netcraft Web Server Survey, monthly since Aug 1995
reading
—
index points
the only number on screen
Nasdaq Composite
Nasdaq Composite recorded closes, left scale, plus the dated peak and trough. Straight segments between points; no interpolation is claimed. Web sites found Netcraft Web Server Survey, monthly hostnames, right scale. Active sites same survey, sites with content of their own, from June 2000; totals marked derived in the source are computed from the Apache count and share Netcraft published. The two scales share a time axis and nothing else; the right one is redrawn when it changes.
Five surges, each drawn in real years after its big bang and not stretched to align. Big bangs from Perez (2002) Table 2.1; bubbles, turning points and golden ages from her Table 5.1 as she has updated it. Deployment is drawn to the next big bang. The fifth turning point is her “2001–???”, drawn to 2026, when this was written; its second dark band is the 2004 to 2007 bubble in mortgage debt, which she counts as the same surge’s second crash.
Drag the pin, or use the arrow keys.
A stylised surge: two periods, four phases, one turning point, sixty years or so. The faint second curve is a sixth surge, if you think one has begun. Perez’s placement is from her 2024 and 2025 statements, put at 2026, when this was written.
Start somewhere else. A world that moves rocks.
In this world moving rocks is the whole economy. It takes three people to move one, and the people who lend money to rock-movers keep a book with a column for arms. Lend against arms, get paid in rocks. It is a good book. It has not been wrong in a century.
Someone builds a lever.
She is not trying to move rocks for money. She wants to see whether a long stick and a small stone will do what she suspects. Move as many as you like, and watch the book.
It does not record her, and not because the lenders are fools. The book has no column for stick. However many rocks she moves, they count as zero arms.
She is not undervalued. She is unvalued. There is no rate at which sticks convert to arms, so there is nothing for anyone to be wrong about.
Two halves of a map.
She moves rocks all afternoon. The book records none of them. One half of the map shows what is possible. The lever-builders hold it, and they count on it constantly: rocks moved, sticks that snapped, the length that works. The other half shows what is priced, and the lenders hold that. The halves are never in the same room.
September 1920
Someone builds a transmitter.
Frank Conrad, an engineer at Westinghouse, had been playing records into the air from the second floor of his garage since 1916, for other amateurs, for no reason. Nobody’s book had a column for it. Then a Pittsburgh department store put a receiver in its West Basement, picked up one of his concerts, and ran an advertisement: sets on sale here, ten dollars and up.
Conrad’s boss saw the advertisement and understood that transmissions sold receivers. Westinghouse applied for a broadcasting licence, and on 2 November 1920 its new station read out the presidential election returns. KDKA is known as the first commercial radio station in the United States. It is still on the air in Pittsburgh.
Broadcasting was not born when someone realised you could transmit music. It was born when a shop noticed that transmitting music made people buy radios. Four years of a hobby and the exchange rate was written by a retailer, in a newspaper, at ten dollars and up.
May 1971
The event Carlota Perez calls the ‘big bang of the fifth technological revolution’ started with a refund.
Intel built the 4004 as a contract job: a chip to run a desk calculator for Busicom, a Japanese manufacturer, who owned the rights. In May 1971 Intel paid Busicom back its $60,000 development fee to get the right to sell the chip for anything other than calculators.
That is the price the market put on the microprocessor at birth. Sixty thousand dollars, and the other party was glad to take it.
The book had a column for calculators.
5 March 1975, and 3 February 1976
The Homebrew Computer Club meets for the first time in a garage in Menlo Park. Steve Wozniak will say for the rest of his life that the Apple I came out of these meetings, and he passed the schematics around for nothing.
Eleven months later a twenty-year-old member writes the club an angry letter. Most of them are running his BASIC without paying for it. Who can afford to do professional work for nothing?
Read it as an instrument being invented in public, in one page: software is a thing with a price. The club hated it. Both sides were describing the future accurately.
March 1989
At CERN, Tim Berners-Lee hands his manager a document called Information Management: A Proposal. Mike Sendall writes across the top of it, in pen: Vague but exciting…
He could not evaluate it. He let it proceed anyway. That is what good behaviour looks like when half the map is missing. Not foresight. Tolerance.
25 August 1991
A student in Helsinki posts to a newsgroup.
None of this was hidden. The refund is in Intel’s own history. The letter was printed. The memo is in CERN’s archive. The post is public to this day. The half of the map that showed what was possible was never secret. It was just not denominated in anything the other half could read.
Before the next story can be told, the reader needs the thing the era actually used to see with.
Advertising on the web was sold the way advertising had been sold for a century: cost per thousand impressions. Eyeballs. It was not a new metric invented by hype. It was the old ruler, carried over whole from print and broadcast, and pointed at a medium that could have measured far more.
A meter has appeared. It has one number on it. For the next few minutes it is the only quantity we will look at.
27 February 1995
Newsweek runs an essay titled “The Internet? Bah!”
After two decades online, I’m perplexed. It’s not that I haven’t had a gas of a good time on the Internet. I’ve met great people and even caught a hacker or two. But today, I’m uneasy about this most trendy and oversold community. Clifford Stoll
Stoll was not an outsider. He had been online twenty years and had tracked a real intruder across the network for a year. He had his hands on the thing. It did not help.
9 August 1995
Netscape goes public at twenty-eight dollars a share, doubled from fourteen at the last minute. Then demand is so heavy that no orderly market can be made, and for nearly two hours the stock does not trade at all.
The first print is seventy-one. It closes at 58.25, valuing a sixteen-month-old company with no profits at about 2.1 billion dollars.
Two hours with no price is the instrument failing to read. Everyone agreed the thing was valuable. Nobody could say what the number was.
June 1998
An economist writes, in a piece called “Why Most Economists’ Predictions Are Wrong”:
By 2005 or so, it will become clear that the Internet’s impact on the economy has been no greater than the fax machine’s. Paul Krugman
This is usually quoted to make him look foolish. He was not being foolish, and he later said plainly that he was wrong. He was reading the productivity statistics, and the productivity statistics genuinely were not showing it.
The same thing had happened with electricity. Factories bought motors and bolted them onto the driveshafts that steam had required. Thirty years of almost no measured gain, until plants were rebuilt around the new power source. The instrument was not broken. It was pointed at the old shape of the world.
1999 to 2001
Roughly a hundred million miles of optical fiber go into the ground worldwide, at a cost of about thirty-five billion dollars. Most of it is never switched on. Contemporary estimates of the share actually lit and carrying traffic run as low as two to five percent.
Those figures were disputed at the time and should be disputed here. The industry association put backbone fiber at around twenty-five percent lit. Carriers argued that counting dark fiber as idle was misleading, since lighting a strand costs far more than laying it.
Either way: capital did something enormously wasteful, and in doing so paid in advance for bandwidth that other people would spend a decade later.
10 March 2000
The Nasdaq Composite closes at 5,048.62.
By the only measurement anyone has, this is the most the future has ever been worth.
9 October 2002
1,114.11.
A fall of 77.9 percent over 647 trading days. The index will not see its March 2000 close again until April 2015, a little over fifteen years later.
The meter is not lying now and was not lying before. It measured what it was built to measure, faithfully, the whole way down.
Here is a different instrument.
It has been publishing since August 1995. The Netcraft Web Server Survey sends one request a month to every web site it can find and prints how many answered. The first survey found 18,957. In the month the first meter peaked, it found 13,106,190. In the month the first meter broke, 35,114,328.
It was not run by a bank or an analyst. Mike Prettejohn ran it from his house in Bath, on a 64 kilobit line he describes as home, creche, office and datacentre, with no outside money. The Apache team found it within hours. It was public the entire time you were reading the other one.
Nobody could say what one web site was worth in index points. Two economies ran side by side on this one chart, and only one of them was denominated in dollars.
Now look closer. It did dip.
June to October 2002
The count reached 38,807,788 in June 2002 and then fell four months in a row, to 35,114,328 in October. The same month the red line bottomed. Netcraft wrote down what had fallen:
Many of the sites that fell by the wayside were parked sites at domain registries and template produced sites at mass hosting companies which retreated from an advertising supported business model. Three companies, Verisign, register.com, and homestead.com, collectively lost over 3 Million such sites during the period. Netcraft, December 2002
The crash exposed a problem: most “websites” weren’t actually sites—they were empty, mass-produced placeholders parked by registrars. So Netcraft drew a second line on the chart: sites with real, original content. They started tracking these separately in June 2000, just after the peak, when register.com alone held over 1.4 million parked shells. Even through the worst of the bust, 2002, true content sites grew—from 14.1 to 16.6 million. Up 17 percent, while everything else was falling.
Only the sites already turned into “eyeballs” disappeared in the crash; the rest kept going.
February 2002
An exchange rate arrives.
Google’s AdWords had launched in October 2000, seven months after the peak, and it launched on impressions. The press release lists the rates: “CPMs from $15 or 1.5 cents an impression.” They were still selling eyeballs.
In February 2002 AdWords Select replaced that with an auction on clicks, ranked by bid and by whether anyone actually wanted the thing. The click was not new. Every web server had logged it since 1995, in a column called the referer. The market had been reading the column next to it. Google’s profit that year was fourteen times the year before.
A rate is not a metric. One person can invent a metric in an afternoon. A rate needs a second party, a market, and enough volume to hold. That is why it took seven years, and why the cracked meter is reading again.
2001 to 2005
The count keeps going, and the right-hand axis has to be redrawn to keep it on the page.
Ten months after the peak, on 15 January 2001, as the money was leaving, a non-commercial project with no business model began the most-used reference work in history. Wikipedia is on this chart as a mark and nowhere else, because no instrument on this page can read it. Its builders were never funded, so they never stopped.
By August 2005, ten years after the first survey, Netcraft found 70,392,567 sites, nearly a third of them with content of their own. The first meter, reading again through its crack, ended that year at 2,205: less than half of what it said at the top.
It has happened four times before.
Perez dates five revolutions from a single visible event each, and every one of them ran into a speculative frenzy and a crash between fifteen and thirty years in. Canal mania. Railway mania. The London-funded build-up of Argentina and the American West. The roaring twenties. The internet.
Then, each time, a deployment period in which the potential installed at a loss was finally spent. The fifth surge is the one that breaks the pattern. Its turning point has run for a quarter of a century, and she says plainly that she does not yet know why.
Before you trust the atlas, hear its author.
There is a certain amount of circularity in all these observations, because the phases have been dated taking the occurrence of crises into account. Carlota Perez, Technological Revolutions and Financial Capital, 2002
She calls the model “a stylised description” and “a heuristic device.” The phases are dated partly from the crises they are then said to explain. The curve you have been walking is a way of seeing, not a law.
Now the instrument is yours.
Perez says artificial intelligence is not a sixth revolution. It is the third leap inside the fifth: the microprocessor in the seventies, the internet in the nineties, AI in the twenties. Most people, asked cold, put it at the foot of a new curve.
The information revolution itself is still spreading but without directionality. There is an almost “juvenile” concentration in games or space travel or social media and now a not-so-juvenile focus on military technology. Carlota Perez, asked in 2025 where the surge stands
Drag the pin. Where do you think AI is? Then see where she puts it.
So the question is not what nobody is measuring.
Take the numbers you count yourself with. Some already have an exchange rate. Some do not. The second kind is where you are unpriced, and it is the only place a next surge can start.
Perez would add one thing, and it should be the last word. Where the surge goes from there is not up to the tinkerers. They supply the potential. Direction is a political choice, and she thinks it is one the state has to make.