The Generative Institution

Planted 02026-08-08

Generate descendants.

How can an institution become more powerful after it disappears?

It sounds like a contradiction. Institutions are supposed to preserve things. They gather people, money, knowledge, machinery, rituals, and authority into one place. Their survival seems like the obvious measure of their success.

A company succeeds if the company survives.

A school succeeds if the school survives.

A movement succeeds if its organization survives.

And yet history repeatedly produces the opposite pattern.

A small community forms around an unusual concentration of talent and an idea that does not yet fit comfortably into the surrounding world. Its members invent methods that barely exist elsewhere. They teach one another. They develop a private language, a culture, a way of seeing.

Then something goes wrong.

The founder dies.

The company collapses.

The institution is attacked.

Employees rebel.

Technology changes.

Money runs out.

Management makes a disastrous decision.

The center breaks apart.

And the people begin to move.

Two and a half thousand years ago, something like this happened to the followers of Pythagoras.

In the twentieth century, it happened at Fairchild Semiconductor.

A few decades later, it happened at Robert Abel & Associates, one of the strange laboratories from which modern computer graphics emerged.

Their stories seem, at first, to belong to different worlds: ancient philosophy, silicon chips, television commercials and digital imagery.

But look closely at the people leaving each institution and another story becomes visible.

Perhaps some institutions do not exist primarily to survive.

Perhaps they exist to generate descendants.

Pythagoras entered southern Italy sometime in the sixth century BCE and gathered followers around him at Croton.

Today his name immediately summons a theorem. But the historical Pythagoras was not simply a mathematician standing beside a triangle. The community associated with him prescribed a way of life and became connected with ideas about number, music, cosmology, the soul, ritual, and moral discipline. What precisely Pythagoras himself taught is difficult to reconstruct, because later generations attributed an enormous range of achievements to him. Even Aristotle distinguished Pythagoras from later thinkers he cautiously called the “so-called Pythagoreans.”

It was not merely a classroom.

It was an attempt to construct a world.

For people living inside such a community, mathematics was not just a tool. Number appeared capable of revealing an order beneath appearances. Musical intervals could be investigated numerically. The cosmos itself might possess a structure that human beings could learn to perceive.

That is exactly the kind of conviction capable of producing an unusually intense intellectual culture. Members were not simply learning isolated propositions. They were learning to see through the same conceptual lens.

And concentrated communities like this are powerful for a simple reason: knowledge moves differently inside them.

A student does not merely receive conclusions from a teacher. He watches what problems the teacher considers important. He absorbs habits of attention. He learns which questions are worth asking, which explanations feel satisfying, which mistakes are embarrassing, which possibilities are exciting.

Much of what makes a tradition powerful is never formally written down.

It lives in people.

That creates a vulnerability.

If knowledge resides in books, destroy the organization and the books may remain.

If knowledge resides in people, destroy the organization and the knowledge begins walking away.

Ancient testimony preserved by Aristoxenus tells of an attack on a Pythagorean gathering in Croton around the middle of the fifth century BCE. A house associated with Milo was burned. Two men, Lysis and Archippus, were said to have survived.

Lysis did not remain in Croton.

He went first to Achaea.

Then he went to Thebes.

There, according to the tradition, the Pythagorean exile became the teacher of Epaminondas, the future Theban general, who reportedly regarded Lysis with such intimacy that he called him father.

A building burns in southern Italy.

A man walks away.

Years later, fragments of the culture that had occupied that building are being transmitted to a young man in another city.

The institution has lost its center.

The knowledge has acquired legs.

Other branches moved differently.

Philolaus of Croton developed an elaborate cosmology organized around what he called limiters and unlimiteds and gave number a fundamental role in making the cosmos intelligible. Eurytus appears in the tradition as his successor. Archytas of Tarentum, perhaps connected through that intellectual lineage, developed a sophisticated mathematical account of the world through proportion and became the first great mathematician whom we can securely place in the later Pythagorean tradition.

The names matter because the doctrine did not simply “spread.”

People carried different pieces of it to different places.

Lysis carried a way of life to Thebes.

Philolaus pushed the tradition toward cosmology.

Archytas pushed it toward mathematics and harmonics while becoming a major political figure at Tarentum.

And eventually these wandering lines crossed other intellectual networks. Ancient reports connect Philolaus’s circle with figures who appear in Plato’s Phaedo, while Plato’s own engagement with Pythagorean mathematics is especially associated with Archytas. Later still, thinkers in the Academy, Middle Platonism, and Neoplatonism repeatedly reconstructed Pythagoras for their own purposes.

The original Pythagorean community disappears.

Pythagoreanism does not.

It becomes less centralized, less controllable, and harder to define.

But it also becomes more portable.

And once portable, it can mutate.

The destruction of the center creates specialization at the edges.

In 1956, William Shockley returned to California after sharing the Nobel Prize for work that led to the transistor and assembled a formidable group of young scientists and engineers at Shockley Semiconductor Laboratory.

There was Robert Noyce.

Gordon Moore.

Jean Hoerni.

Jay Last.

Eugene Kleiner.

Victor Grinich.

Julius Blank.

Sheldon Roberts.

Eight people whose names, at the time, meant far less than Shockley’s.

Shockley possessed the reputation.

They possessed the future.

His management drove them away.

In 1957 the eight left together and created Fairchild Semiconductor. A young banker named Arthur Rock helped arrange the financing. He would later move west, become one of Silicon Valley’s early venture capitalists, and help finance companies including Intel and Apple.

Shockley considered the departing men traitors.

The phrase survived.

They became the “traitorous eight.”

From Shockley’s perspective, something valuable had escaped from the institution.

He was correct.

He simply misunderstood what would happen next.

Because the eight did not merely leave with technical knowledge.

They left with one another.

They already knew how the others thought. They knew who could solve which problems. They knew whom they trusted. They knew how quickly they could work together.

The social network itself had become a technology.

At Fairchild, Jean Hoerni developed the planar process, a manufacturing breakthrough that made semiconductor devices more reliable and suitable for mass production. Robert Noyce recognized that the same process could be used to connect multiple transistors into an integrated circuit on one piece of silicon. Gordon Moore watched the number of components that could be placed on those circuits rise and described the trend that became Moore’s Law.

Fairchild had gathered an extraordinary concentration of capability.

And then Fairchild began to leak.

Jay Last left first.

He created Amelco Semiconductor, and Jean Hoerni and Eugene Kleiner joined him.

Others left to create Signetics, an early independent integrated-circuit company.

Kleiner noticed another opportunity entirely. Semiconductor companies did not merely need chips; they needed machines to manufacture chips. While at Fairchild he encouraged a technician to create Electroglas, helping create the surrounding equipment ecosystem on which semiconductor manufacturing would depend.

Then Kleiner moved one level further outward.

He stopped merely building semiconductor companies and began helping finance new companies.

In 1972 he co-founded Kleiner Perkins.

Another Fairchild alumnus, Don Valentine, founded Sequoia Capital the same year.

The diaspora had discovered venture capital.

A semiconductor company produces an engineer.

The engineer becomes a founder.

The founder becomes an investor.

The investor finances hundreds of founders.

The original institution is no longer reproducing companies one generation at a time.

It has begun reproducing the mechanism of reproduction.

Then came the departures everyone remembers.

In 1968, Robert Noyce and Gordon Moore left Fairchild and founded Intel.

Their first hire was another Fairchild colleague, Andy Grove.

A year later Jerry Sanders, who had risen through Fairchild’s sales organization, joined seven other former Fairchild employees to create Advanced Micro Devices (AMD).

Others kept moving.

Fairchild people helped build National Semiconductor, LSI Logic, Altera, and a growing forest of companies whose founders, employees, investors, customers and suppliers repeatedly crossed paths.

Eventually people needed a word for them.

They called them the Fairchildren.

A 1986 semiconductor genealogy traced 126 companies directly back to Fairchild.

But even that diagram understates what happened, because people did not carry only semiconductor designs out of Fairchild.

They carried a style of organization.

They carried expectations about stock ownership.

They carried habits of technical argument.

They carried a belief that an engineer could become an entrepreneur.

They carried friendships.

They carried rivalries.

They carried lists of people they would hire when they got the money.

They carried the names of customers who might take their calls.

They carried unfinished ideas.

The institution was walking out through the front door one human being at a time.

Every departure weakened the center.

But every departure strengthened the ecosystem.

A company evaluates talent loss from inside the company. The departure of a brilliant engineer is a subtraction. Something valuable has left.

History can evaluate the same event differently.

Jay Last leaving Fairchild was subtraction for Fairchild.

Jay Last founding Amelco was multiplication for Silicon Valley.

Eugene Kleiner leaving was subtraction.

Kleiner Perkins was multiplication.

Noyce and Moore leaving was subtraction.

Intel was multiplication.

Jerry Sanders leaving was subtraction.

AMD was multiplication.

The institution loses a person.

The network gains a node.

And once those nodes begin creating nodes of their own, the original company is no longer the most important object in the story.

The genealogy is.

Fairchild never fully recovered from the exodus. It changed owners repeatedly and eventually disappeared as an independent name. Yet its descendants became vastly larger than Fairchild itself.

The parent disappeared into the success of its children.

A generation later, another strange institution appeared in Los Angeles.

Robert Abel & Associates didn’t manufacture semiconductors.

It made images.

In the 1970s and early 1980s, computer-generated imagery was still exotic. There was no mature visual-effects industry built around digital production. The software packages, production pipelines, artistic conventions and specialized labor markets that later artists would take for granted were still being invented.

If you wanted to make sophisticated computer imagery, you often had to invent the means of making it.

Robert Abel and his collaborators therefore needed unusual people.

Not simply artists.

Not simply programmers.

People capable of moving between worlds.

Among them was Bill Kovacs, a programmer who began experimenting with an Evans & Sutherland computer Abel had acquired. Kovacs eventually led Abel’s software group as it developed a repertoire of computer-graphics techniques.

Then Kovacs left.

In 1984 he joined Larry Barels and Mark Sylvester to found Wavefront Technologies, which developed commercial 3-D animation software (which, you may know the name of the software called Maya).

An internal production tool had become the seed of a software company.

That is the Fairchild pattern in another medium.

At Abel, Randy Roberts collaborated with technical director Tim McGovern on breakthrough computer imagery.

Then McGovern left.

He became an early member of Sony Pictures Imageworks and later won an Academy Award for the visual effects of Total Recall.

Roberts went to Rhythm & Hues.

Frank Vitz, another technical talent in Abel’s orbit, went on to work on the groundbreaking Spider-Man ride film for Universal.

Tom Barron went on to Image G.

Ray Feeney went on to Silicon Grail.

Sherry McKenna became a founder of Oddworld Inhabitants.

Allen DeBevoise built Creative Planet.

Abel alumni including Rob Legato, Charlie Gibson, John Nelson, Scott Farrar, and McGovern would go on to Academy-recognized visual-effects careers.

But perhaps the clearest example was John Hughes.

Hughes spent roughly a decade at Robert Abel & Associates. He worked on motion-control systems. He became a technical director. He learned inside a company where creative and technical problems constantly collided.

Then, as Abel’s corporate situation deteriorated, Hughes and a group of colleagues began planning an escape.

Pauline Ts’o.

Keith Goldfarb.

Cliff Boule.

Frank Wuts.

Charles Gibson.

They joined Hughes in founding Rhythm & Hues in 1987.

They did not begin with a giant studio.

They began with a tiny collection of equipment, their accumulated knowledge, and each other. Hughes later recalled that when Omnibus finally collapsed, employees came to work and found the doors locked. The new company suddenly had to become real.

What had been a department of expertise inside Abel was now an independent organism.

Rhythm & Hues would eventually become one of the major visual-effects houses in Hollywood.

Again: the people were the continuity.

And the branches continued crossing.

Richard Hollander, who had worked at Abel developing computerized motion-control systems, went on to co-found VIFX. VIFX later became associated with Fox and Blue Sky before being acquired by Rhythm & Hues. Hollander then led Rhythm & Hues’ film-effects operation and later left for Pixar.

Abel leads to Kovacs.

Kovacs leads to Wavefront.

Abel leads to Hughes and his colleagues.

They lead to Rhythm & Hues.

Abel leads to Hollander.

Hollander leads to VIFX, then back into Rhythm & Hues, then to Pixar.

Abel leads to McGovern.

McGovern leads to Sony Imageworks.

Abel leads to people who become supervisors, founders, software developers, directors and teachers elsewhere.

What appeared from inside the company as turnover becomes, from a distance, the construction of an industry.

The business disappeared.

The people did not.

The company had spent years concentrating knowledge inside human beings.

When the container broke, those human beings carried the capability outward.

What had once required one extraordinary company gradually became something an entire industry knew how to do.

The company lost its monopoly on the future because the future had arrived.

Once you see this pattern, it begins appearing everywhere.

At MAGI/SynthaVision, another pioneer of early computer imagery, Chris Wedge, Michael Ferraro, Carl Ludwig, Eugene Troubetzkoy, Alison Brown, and David Brown worked together before MAGI went out of business.

They regrouped.

In 1987 the six founded Blue Sky Studios.

Troubetzkoy, a theoretical physicist, and Ludwig, a former NASA engineer, developed the rendering technology that gave Blue Sky its distinctive look. Wedge became its creative leader. The descendants of a failed technical company eventually made Ice Age.

At Triple-I, Gary Demos and John Whitney Jr. were pushing computer imagery toward feature films. They became convinced that the company was unwilling to devote enough computing power to the opportunity.

So they left.

They founded Digital Productions and built their new company around precisely the bet their previous employer had resisted.

At Atari, programmers David Crane, Alan Miller, Bob Whitehead, and Larry Kaplan watched the games they created become enormously valuable while the programmers themselves received little recognition.

They asked Atari to treat game designers more like creative authors.

Atari refused.

So the four left in 1979 and founded Activision.

What escaped Atari was not simply four employees. It was the idea of the independent third-party videogame developer.

At Digital Equipment Corporation, engineer Edson de Castro wanted to pursue a new family of 16-bit computers.

DEC rejected the proposal.

De Castro left with colleagues and founded Data General.

Their first machine, the Nova, attacked the market that de Castro had wanted to pursue from inside DEC.

At Silicon Graphics, founder Jim Clark became convinced that cheaper systems would eventually undermine the expensive workstation business. He left SGI in 1994, met a young programmer named Marc Andreessen, recruited Eric Bina and other members of the Mosaic browser team, and created Netscape.

Knowledge that had accumulated around interactive graphics and high-performance computing jumped into the commercial internet.

At Thinking Machines, Brewster Kahle worked amid an extraordinary attempt to commercialize massively parallel computing.

Thinking Machines eventually collapsed.

Kahle carried another obsession outward: making enormous bodies of information searchable and accessible over networks. His work led through WAIS and ultimately to the Internet Archive.

A vanished supercomputer company became one ancestor of a system devoted to ensuring that other vanished things could still be remembered.

And then there was General Magic.

The company tried in the early 1990s to build a personal digital communicator before the surrounding technological ecosystem was ready.

The product failed.

The people did not.

Tony Fadell later became central to the iPod and iPhone and founded Nest.

Andy Rubin became one of the creators of Android.

Pierre Omidyar, while working at General Magic, began the project that became eBay.

Megan Smith later became a Google executive and U.S. Chief Technology Officer.

Other General Magic alumni flowed back into Apple and outward through Google and the wider technology industry. One former colleague, Joanna Hoffman, later described them as the generation that carried the abandoned vision into other companies and recreated it there.

The product failed.

The product’s future employees succeeded.

Again and again, the same strange thing happens.

The institution cannot hold the people who have learned how to see its future.

So the future leaves with them.

Pythagoras’s communities, Fairchild Semiconductor, Robert Abel & Associates, MAGI, Atari, DEC, General Magic, Thinking Machines: they were not alike in purpose.

One pursued a philosophical way of life.

One built semiconductors.

One made moving images.

One built videogames.

One built computers.

One tried to invent the smartphone too early.

But each occupied a similar position in the development of a new intellectual or technological world.

Each assembled scarce people before the surrounding ecosystem knew how to produce them reliably.

Each gave those people difficult problems before standardized solutions existed.

Each forced them to invent methods rather than merely apply methods.

Each created knowledge that could not easily be separated from the people practicing it.

And each, intentionally or not, became a school.

That may be the crucial property of what we could call a generative institution.

Its defining product is not necessarily the thing listed on the invoice.

Fairchild sold semiconductors.

Abel sold images.

Atari sold games.

Thinking Machines sold computers.

General Magic tried to sell communicators.

But underneath those visible products, another manufacturing process was occurring.

Fairchild was manufacturing semiconductor entrepreneurs.

Abel was manufacturing people who could build the computer-graphics industry.

Atari was manufacturing videogame auteurs.

Thinking Machines was manufacturing people accustomed to thinking at scales the rest of the computing world had barely encountered.

Pythagorean communities were manufacturing people capable of carrying a way of seeing into another city.

Their most consequential products could resign.

People can leave.

We normally treat this as an organizational weakness.

Perhaps it is also one of our most important mechanisms of progress.

Suppose Shockley had successfully kept Noyce, Moore, Hoerni, Last and Kleiner under his control.

There might have been no Fairchild in 1957.

Suppose Fairchild had somehow prevented Noyce, Moore, Sanders, Last, Kleiner and the others from leaving.

Fairchild itself might have become stronger.

Silicon Valley might have become weaker.

Suppose Robert Abel had been able to keep Bill Kovacs, John Hughes, Richard Hollander, Tim McGovern and the rest of that generation inside one company indefinitely.

Robert Abel & Associates might have accumulated extraordinary power.

But perhaps there would have been no Wavefront in the same form.

No Rhythm & Hues in the same form.

No VIFX in the same form.

The institution might have grown.

The ecosystem might have shrunk.

Because innovation does not merely require concentration.

It sometimes requires dispersal.

A dense institution allows ideas to combine.

A diaspora allows them to recombine.

At Fairchild, Hoerni could work beside Noyce, Moore, Last and Kleiner.

But when they separated, something new became possible.

One branch became Intel.

One became Amelco.

One became venture capital.

Another became AMD.

Another became National Semiconductor.

Each branch took part of the original environment and combined it with something Fairchild itself did not contain.

The same thing happened to Abel’s people.

Kovacs took production software and turned it toward commercial software.

Hughes took the hybrid culture of art and engineering and built another studio around it.

McGovern carried visual-effects knowledge into Sony.

Hollander carried engineering knowledge through multiple companies.

The original institution had been a bundle.

Diaspora unbundled it.

Then the pieces encountered new people.

And recombination began again.

Most descendants do not become Intel.

Most startups fail.

Most intellectual branches disappear.

Many Fairchildren vanished.

Many early computer-graphics companies merged, collapsed or were forgotten.

General Magic itself failed.

Thinking Machines went bankrupt.

Robert Abel & Associates disappeared.

Rhythm & Hues, one of Abel’s great descendants, would itself eventually go through bankruptcy after producing some of the most celebrated visual effects of its era.

The genealogy does not move upward in a clean line.

It branches.

Branches break.

Other branches cross.

Some die.

Some unexpectedly produce another generation.

There is another strange consequence of generative institutions.

Their descendants can become so successful that they obscure their own ancestry.

Millions of people know Intel.

Far fewer know Fairchild.

Billions of people have watched computer-generated imagery.

Far fewer know Robert Abel & Associates.

Millions have watched Ice Age without knowing that Blue Sky emerged from six people who had worked together at MAGI.

Billions use Android without knowing that Andy Rubin once sat inside General Magic, another company trying to build a version of the mobile future before the world was ready for it.

Millions have bought games from third-party studios without knowing that the very legitimacy of the independent console-game developer was once fought over by four programmers leaving Atari.

Students encounter ideas associated with Pythagoreanism without imagining Lysis escaping southern Italy and teaching in Thebes.

Success erases the conditions that made success possible.

Once an innovation becomes infrastructure, it stops looking like an innovation.

Once a descendant becomes enormous, the ancestor begins to look small.

The river becomes more visible than the spring.

We look backward and search for winners.

Which company won?

Which philosopher was correct?

Which institution survived?

What about

What did this institution cause to exist after it was gone?

That changes the map.

Failed companies now appear enormously successful.

A persecuted philosophical community may appear historically resilient.

A studio that vanished decades ago may turn out to occupy a crucial position in the genealogy of an entire medium.

The unit of analysis shifts.

Not the institution.

The lineage.

A generative institution is therefore a particular kind of place.

It creates an unusually dense environment for learning.

It gathers people who otherwise might never have met.

It gives them problems for which the textbooks have not yet been written.

It forces different disciplines into contact.

It teaches techniques.

But it also teaches taste.

Standards.

Ambition.

Confidence.

Who is good.

Who can be trusted.

What is possible.

What everyone else is wrong about.

Then, for one reason or another, the container fails to contain them.

And what happens next determines its historical importance.

Some institutions disappear and leave nothing.

Others explode like stars.

The star is destroyed.

Its material becomes the raw material for new worlds.

Fairchild Semiconductor was not merely a company that eventually lost its position.

It was a stellar nursery.

Robert Abel & Associates was not merely a studio whose business proved unsustainable.

It was a temporary concentration of people from which a larger visual medium could form.

General Magic was not merely a failed device company.

It was a room containing Tony Fadell, Andy Rubin, Pierre Omidyar, Megan Smith and others before the world knew quite what to do with people who thought that way.

The Pythagorean communities were not simply an ancient school interrupted by political violence.

Their ideas became embodied in people who could cross geography and generations.

The center could not hold.

But perhaps holding was never the most consequential thing it did.

There is a temptation to turn this into a law:

Great institutions die so their ideas can spread.

History is not that tidy.

Many institutions die and are simply forgotten.

Many generative institutions remain alive.

Fairchild’s fragmentation cannot be explained by the same forces that drove attacks on Pythagorean communities, and neither resembles the economic pressures surrounding an early computer-graphics studio.

The causes differ.

The pattern lies in the consequence.

When knowledge has become sufficiently embodied in a community, destroying or fragmenting the community does not necessarily destroy the knowledge.

It changes its geometry.

What had been vertical becomes horizontal.

Instead of knowledge flowing from a center to its members, it begins flowing between descendants.

Instead of one institution exploring one future, many institutions explore many futures.

Authority declines.

Variation increases.

Recombination accelerates.

And occasionally the resulting network becomes far more powerful than the institution that produced it.

That is why the moment of collapse is so easy to misunderstand.

Stand inside Fairchild while Noyce and Moore are packing up and you see a company losing two extraordinary employees.

Stand decades later and you see Intel.

Watch Jay Last, Jean Hoerni and Eugene Kleiner leave and you see another wound.

Stand farther away and you see Amelco, venture capital, and another set of branches forming.

Stand outside Robert Abel & Associates in 1987 and you see a company disappearing.

Follow Bill Kovacs and you arrive at Wavefront.

Follow John Hughes and his colleagues and you arrive at Rhythm & Hues.

Follow Tim McGovern and you arrive at Sony Imageworks.

Follow Richard Hollander and you move through VIFX, Rhythm & Hues and Pixar.

Stand at a burned Pythagorean meeting place and you see a tradition being destroyed.

Follow Lysis and you arrive in Thebes.

Follow Philolaus and Archytas and the intellectual tradition changes shape.

The event has not changed.

Your distance from it has.

The same departure can look like failure in the present and propagation in retrospect.

We tend to imagine knowledge as something stored.

In libraries.

In universities.

In companies.

In laboratories.

In schools.

But knowledge is also something carried.

It travels in human beings.

Bill Kovacs carried things out of Abel that could never have been reconstructed merely by purchasing Abel’s computers.

Noyce and Moore carried things out of Fairchild that were not contained in Fairchild’s patents.

Lysis could walk into Thebes without carrying a building from Croton.

Brewster Kahle could leave Thinking Machines without taking a Connection Machine with him and still carry away a sense of what enormous information systems might make possible.

People carry more than facts.

They carry techniques.

Intuitions.

Standards.

Questions.

Relationships.

Stories.

Grudges.

Ambitions.

Unfinished problems.

They remember what worked.

They remember what failed.

They remember the colleague who solved an impossible problem at three in the morning.

They remember the idea management refused to fund.

They remember which assumption everyone once believed and which experiment finally broke it.

And when they enter another institution, they do not reproduce the old one exactly.

They combine what they inherited with a new environment.

The idea mutates.

That mutation is not necessarily corruption of the tradition.

It may be the mechanism by which the tradition remains alive.

And so the history of ideas may resemble neither a monument nor a straight line.

It looks more like a family tree.

The trunk matters.

But eventually the branches are where the future appears.

Pythagoras could not have foreseen every form that later Pythagorean thought would take.

William Shockley could not have foreseen that driving eight employees away would help create Fairchild.

Fairchild could not have mapped the companies that would emerge from Noyce, Moore, Last, Kleiner, Sanders and generations of people who followed them.

Robert Abel could not have followed every line running from Kovacs, Hughes, Hollander, McGovern, Roberts, Feeney, McKenna, Gibson, Legato and the others who passed through his studio.

General Magic could not control what Fadell, Rubin or Omidyar would do with the future they had glimpsed together.

That is precisely the point.

The deepest influence escapes control.

An institution creates people capable of leaving it.

Those people carry fragments of its worldview elsewhere.

The fragments recombine.

The descendants become more numerous than the ancestors.

Eventually the original institution disappears from everyday memory.

But everywhere, quietly, the world begins to resemble the future that institution once saw.

And that leaves us with a different way to think about disappearance.

Sometimes a center collapses because its experiment failed.

Sometimes it collapses because the world changed around it.

Sometimes its people simply discover that the thing they most want to build cannot be built inside it.

And sometimes, in the most interesting cases, the center becomes less important because what was once concentrated inside it has escaped into the world.

The institution dies.

Lysis walks to Thebes.

Noyce and Moore start Intel.

Kleiner starts funding the next generation.

Kovacs builds Wavefront.

Hughes and his colleagues open Rhythm & Hues.

Wedge and five coworkers turn MAGI’s collapse into Blue Sky.

Crane and three programmers turn their argument with Atari into Activision.

Kahle turns a life among Thinking Machines into a quest to preserve the world’s information.

Fadell and Rubin carry a failed mobile dream toward the iPhone and Android.

The building empties.

The network fills.

And only then do we discover what the institution was really making.

People who knew how to make the next thing.