The History of Computing

Ross Perot built two powerhouse companies and changed the way politicians communicate with their constituents. Perot was an Eagle Scout who went on to join the US Naval Academy in 1949, and served in the Navy until the late 1950s. He then joined the IBM sales organization and one year ended up meeting his quota in the second week of the year. He had all kinds of ideas for new things to do and sell, but no one was interested.

So he left and formed a new company called Electronic Data Systems, or EDS, in 1962. You see, these IBM mainframes weren’t being used for time sharing so most of the time they were just sitting idle. So he could sell the unused time from one company to another. Perot learned from the best. As with IBM he maintained a strict dress code. Suits, no facial hair, and a high and tight crew cut as you’d find him still sporting years after his Navy days.

And over time they figured out many of these companies didn’t have anyone capable of running these machines in the first place, so they could also step in and become a technology outsourcer, doing maintenance and servicing machines. Not only that, but they were perfectly situated to help process all the data from the new Medicare and Medicaid programs that were just starting up. States had a lot of new paperwork to process and that meant computers.

He hired Morton Meyerson out at Bell Helicopter in 1966, who would become the president and effectively created the outsourcing concept in computing. Meyerson would become the president of EDS before leaving to take a series of executive roles at other organizations, including the CTO at General Motors in the 1980s before retiring.

EDS went public in 1968. He’d taken $1,000 in seed money from his wife Margot to start the company, and his stake was now worth $350 million, which would rise sharply in the ensuing years as the company grew. By the 1970s they were practically printing cash. They were the biggest insurance data provider and added credit unions then financial markets and were perfectly positioned to help build the data networks that ATMs and point of sale systems would use. By the start of 1980 they were sitting on a quarter billion dollars in revenues and 8,000 employees. They continued to expand into new industries with more transactional needs, adding airlines and travel.

He sold in 1984 to General Motors for $2.5 billion and Perot got $700 million personally. Meyerson stayed on to run the company and by 1990 their revenues topped $5 billion and neared 50,000 employees.

Perot just couldn’t be done in business. He was good at it. So in 1988 he started another firm, Perot Systems. The company grew quickly. Perot knew how to sell, how to build sales teams, and how to listen to customers and build services products they wanted. Perot again looked for an effective leader and tapped Meyerson yet again, who became the CEO of Perot Systems from 1992 to 1998. Perot’s son Ross Jr took over the company.

In 2008, EDS and their 170,000 employees was sold to Hewlett-Packard for $13.9 billion and in 2009 Perot Systems was sold to Dell for $3.9 billion. Keep in mind that Morton Meyerson was a mentor to Michael Dell. When they were sold, Perot Systems had 23,000 employees and $2.8 billion in revenues. That’s roughly a 1.4x multiple of revenues, which isn’t as good as the roughly 2x multiple Perot got off EDS - but none too shabby given that by then multiples were down for outsourcers.

Based on his work and that of others, they’d built two companies worth nearly $20 billion - before 2010, employing nearly 200,000 people.

Along the way, Perot had some interesting impacts other than just building so many jobs for so many humans. He passed on an opportunity to invest in this little company called Microsoft. So when Steve Jobs left Apple and looked for investors he jumped on board, pumping $20 million into NeXT Computer, and getting a nice exit when the company went to Apple for nearly half a billion.

Perot was philanthropic. He helped a lot of people coming home from various armed services in his lifetime. He was good to those he loved. He gave $10 million to have his friend Morton Meyerson’s name put on the Dallas Symphony Orchestra’s Symphony Center. And he was interested in no BS politics. Yet politics had been increasingly polarized since Nixon.

So Perot also ran for president of the US in 1992, against George Bush and Bill Clinton. He didn’t win but he flooded the airwaves with common sense arguments about government inefficiency and a declining market for doing business. He showed computer graphics with all the charts and graphs you can imagine. And while he didn’t get even one vote in the electoral college did manage to get 19 percent of the vote. His message was one of populism. Take the country back, stop deficit spending just like he ran his companies, and that persists with various wings of especially the Republican Party to this day. Especially in Perot’s home state of Texas. He didn’t win, but he effectively helped define the Contract with America that that Newt Gingrich and the 90s era of oversized suit jacket Republicans used to as a strategy.

He argued for things to help the common people - not politicians. Ironically, those that took much of his content actually did just the opposite, slowed down the political machine by polarizing the public. And allowed deficit spending to increase on their watch. He ran again in 1996 but this time got far less votes and didn’t end up running for office again.

He had a similar impact on IBM. Around 30 years after leaving the company, his success in services was one of the many inspirations for IBM pivoting into services as well. By then the services industry was big enough for plenty of companies to thrive and while sales could be competitive they all did well as personal computing put devices on desks across the world and those devices needed support.

Perot died in 2019, one of the couple hundred richest people in the US. Navy Lieutenant. Founder. Philanthropist. Texan. Father. Husband. His impact on the technology industry was primarily around seeing waste. Wasted computing time. Wasted staffing where more efficient outsourcing paradigms were possible. He inspired massive shifts in the industry that persist to this day.

The Osborne Effect isn’t an episode about Spider-Man that covers turning green or orange and throwing bombs off little hoverboards. Instead it’s about the impact of The Osborne 1 computer on the history of computers. Although many might find discussing the Green Goblin or Hobgoblin much more interesting.

The Osborne 1 has an important place in the history of computing because when it was released in 1981, it was the first portable computer that found commercial success. Before the Osborne, there were portable teletype machines for sure, but computers were just starting to get small enough that a fully functional machine could be taken on an airplane.

It ran 2.2 of the CP/M operating system and came with a pretty substantial bundle of software. Keep in mind, there weren’t internal hard drives in machines like this yet but instead CP/M was a set of floppies. It came with MBASIC from Microsoft, dBASE II from Ashton-Tate, the WordStar word processor, SuperCalc for spreadsheets, the Grammatik grammar checker, the Adventure game, early ledger tools from PeachTree Software, and tons of other software. By bundling so many titles, they created a climate where other vendors did the same thing, like Kaypro. After all, nothing breeds competitors like the commercial success of a given vendor.

The Osborne was before flat panel screens so had a built-in CRT screen. This and the power supply and the heavy case meant it weighed almost 25 pounds and came in at just shy of $1,800. Imagine two disk drives with a 5 inch screen in the middle. The keyboard, complete with a full 10-key pad, was built into a cover that could be pulled off and used to interface with the computer. The whole thing could fit under a seat on an airplane. Airplane seats were quite a bit larger than they are today back then!

We think of this as a luggable rather than a portable because of that and because computers didn’t have batteries yet. Instead it pulled up to 37 watts of power. All that in a 20 inch wide case that stood 9 inches tall.

The two people most commonly associated with the Osborne are Adam Osborne and Lee Felsenstein. Osborne got his PhD from the University of Delaware in 1968 and went to work in chemicals before he moved to the Bay Area and started writing books about computers and started a company called Osborne and Associates to write computer books. He sold that to McGraw-Hill in 1979.

By then he’d been hanging around the Homebrew Computer Club for a few years and there were some pretty wild ideas floating around. He saw Jobs and Wozniak demo the Apple I and watched their rise. Founders and engineers from Cromemco, IMSAI, Tiny BASIC, and Atari were also involved there - mostly before any of those products were built. So with the money from McGraw-Hill and sales of some of his books like An Introduction To Microcomputers, he set about thinking through what he could build.

Lee Felsenstein was another guy from that group who’d gotten his degree in Computer Science at Berkeley before co-creating Community Memory, a project to build an early bulletin board system on top of a SDS 940 timesharing mainframe with links to terminals like a Teletype Model 33 sitting at Leopold’s Records in Berkeley. That had started up back in 1973 when Doug Englebart donated his machine from The Mother of All Demos and eventually moved to minicomputers as those became more available.

Having seen the world go from a mainframe the size of a few refrigerators to minicomputers and then to early microcomputers like the Altair, when a hardware hacker like Felsenstein paired up with someone with a little seed money like Osborne, magic was bound to happen. The design was similar to the NoteTaker that Alan Kay had built at Xerox in the 70s - but hacked together from parts they could find. Like 5 inch Fujitsu floppy drives.

They made 10 prototypes with metal cases and quickly moved to injection molded plastic cases, taking them to the 1981 West Coast Computer Faire and getting a ton of interest immediately. Some thought the screen was a bit too small but at the time the price justified the software alone. By the end of 1981 they’d had months where they did a million dollars in sales and they fired up the assembly line. People bought modems to hook to the RS-232 compatible serial port and printers to hook to the parallel port. Even external displays.

Sales were great. They were selling over 10,000 computers a month and Osborne was lining up more software vendors, offering stock in the Osborne Computer Corporation. By 1983 they were preparing to go public and developing a new line of computers, one of which was the Osborne Executive. That machine would come with more memory, a slightly larger screen, an expansion slot and of course more software using sweetheart licensing deals that accompanied stock in the company to keep the per-unit cost down. He also announced the Vixen - same chipset but lighter and cheaper.

Only issue is this created a problem, which we now call the Osborne Effect. People didn’t want the Osborne 1 any more. Seeing something new was on the way, people cancelled their orders in order to wait for the Executive. Sales disappeared almost overnight. At the time, computer dealers pushed a lot of hardware and the dealers didn’t want to have all that stock of an outdated model. Revenue disappeared and this came at a terrible time.

The market was changing. IBM showed up with a PC, Apple had the Lisa and were starting to talk about the Mac. KayPro had come along as a fierce competitor. Other companies had clued in on the software bundling idea. The Compaq portable wasn’t far away. The company ended up cancelling the IPO and instead filing for bankruptcy. They tried to raise money to build a luggable or portable IBM clone - and if they had done so maybe they’d be what Compaq is today - a part of HP.

The Osborne 1 was cannibalized by the Osborne Executive that never actually shipped. Other companies would learn the same lesson as the Osborne Effect throughout history.

And yet the Osborne opened our minds to this weird idea of having machines we could take with us on airplanes. Even if they were a bit heavy and had pretty small screens. And while the timing of announcements is only one aspect of the downfall of the company, the Osborne Effect is a good reminder to be deliberate about how we talk about future products. Especially for hardware but we also have to be careful not to sell features that don’t exist yet in software.

Chess is a game that came out of 7th century India, originally called chaturanga. It evolved over time, perfecting the rules - and spread to the Persians from there. It then followed the Moorish conquerers from Northern Africa to Spain and from there spread through Europe. It also spread from there up into Russia and across the Silk Road to China. It’s had many rule formations over the centuries but few variations since computers learned to play the game. Thus, computers learning chess is a pivotal time in the history of the game.

Part of chess is thinking through every possible move on the board and planning a strategy. Based on the move of each player, we can review the board, compare the moves to known strategies, and base our next move on either blocking the strategy of our opponent or carrying out a strategy of our own to get a king into checkmate.

An important moment in the history of computers is when computers got to the point that they could beat a chess grandmaster. That story goes back to an inspiration from the 1760s where Wolfgang von Kempelen built a machine called The Turk to impress Austrian Empress Maria Theresa. The Turk was a mechanical chess playing robot with a Turkish head in Ottoman robes that moved pieces.

The Turk was a maze of cogs and wheals and moved the pieces during play. It travelled through Europe, beating the great Napoleon Bonaparte and then the young United States, also besting Benjamin Franklin. It had many owners and they all kept the secret of the Turk. Countless thinkers wrote about theories about how it worked, including Edgar Allen Poe. But eventually it was consumed by fire and the last owner told the secret. There had been a person in the box moving the pieces the whole time.

All those moving parts were an illusion. And still in 1868 a knockoff of a knockoff called Ajeeb was built by a cabinet maker named Charles Hooper. Again, people like Theodore Roosevelt and Harry Houdini were bested, along with thousands of onlookers.

Charles Gumpel built another in 1876 - this time going from a person hiding in a box to using a remote control. These machines inspired people to think about what was possible. And one of those people was Leonardo Torres y Quevedo who built a board that also had electomagnets move pieces and light bulbs to let you know when the king was in check or mate. Like all good computer games it also had sound.

He started the project in 1910 and by 1914 it could play a king and rook endgame, or a game where there are two kings and a rook and the party with the rook tries to get the other king into checkmate. At the time even a simplified set of instructions was revolutionary and he showed his invention off at the Paris where notable other thinkers were at a conference, including Norbert Weiner who later described how minimax search could be used to play chess in his book Cybernetics.

Quevedo had built an analytical machine based on Babbage’s works in 1920 but adding electromagnets for memory and would continue building mechanical or analog calculating machines throughout his career. Mikhail Botvinnik was 9 at that point and the Russian revolution wound down in 1923 when the Soviet Union was founded following the fall of the Romanovs. He would become the first Russian Grandmaster in 1950, in the early days of the Cold War. That was the same year Claude Shannon wrote his seminal work, “Programming a Computer for Playing Chess.” The next year Alan Turing actually did publish executable code to play on a Ferranti Mark I but sadly never got to see it complete before his death. The prize to actually play a game would go to Paul Stein and Mark Wells in 1956 working on the MANIAC. Due to the capacity of computers at the time, the board was smaller but the computer beat an actual human.

But the Russians were really into chess in the years that followed the crowing of their first grandmaster. In fact it became a sign of the superior Communist politic. Botvinnik also happened to be interested in electronics, and went to school in Leningrad University's Mathematics Department. He wanted to teach computers to play a full game of chess. He focused on selective searches which never got too far as the Soviet machines of the era weren’t that powerful. Still the BESM managed to ship a working computer that could play a full game in 1957.

Meanwhile John McCarthy at MIT introduced the idea of an alpha-beta search algorithm to minimize the number of nodes to be traversed in a search and he and Alan Kotok shipped A Chess Playing Program for the IBM 7090 Computer, which would be updated by Richard Greenblatt when moving from the IBM mainframes to a DEC PDP-6 in 1965, as a side project for his work on Project MAC while at MIT. Here we see two things happening. One we are building better and better search algorithms to allow for computers to think more moves ahead in smarter ways. The other thing happening was that computers were getting better. Faster certainly, but more space to work with in memory, and with the move to a PDP, truly interactive rather than batch processed.

Mac Hack VI as Greenblatt’s program would eventually would be called, added transposition tables - to show lots of previous games and outcomes. He tuned the algorithms, what we would call machine learning today, and in 1967 became the first computer program to defeat a person at the tournament level and get a chess rating. For his work, Greenblatt would become an honorary member of the US Chess Federation.

By 1970 there were enough computers playing chess to have the North American Computer Chess Championships and colleges around the world started holding competitions. By 1971 Ken Thompson of Bell Labs, in a sign of the times, wrote a computer chess game for Unix. And within just 5 years we got the first chess game for the personal computer, called Microchess. From there computers got incrementally better at playing chess. Computer games that played chess shipped to regular humans, dedicated physical games, little cheep electronics knockoffs. By the 80s regular old computers could evaluate thousands of moves.

Ken Thompson kept at it, developing Belle from 1972 and it continued on to 1983. He and others added move generators, special circuits, dedicated memory for the transposition table, and refined the alpha-beta algorithm started by McCarthy, getting to the point where it could evaluate nearly 200,000 moves a second. He even got the computer to the rank of master but the gains became much more incremental. And then came IBM to the party.

Deep Blue began with researcher Feng-hsiung Hsu, as a project called ChipTest at Carnegie Mellon University. IBM Research asked Hsu and Thomas Anantharamanto complete a project they started to build a computer program that could take out a world champion. He started with Thompson’s Belle. But with IBM’s backing he had all the memory and CPU power he could ask for.

Arthur Hoane and Murray Campell joined and Jerry Brody from IBM led the team to sprint towards taking their device, Deep Thought, to a match where reigning World Champion Gary Kasparov beat the machine in 1989. They went back to work and built Deep Blue, which beat Kasparov in their third attempt in 1997. Deep Blue was comprised of 32 RS/6000s running 200 MHz chips, split across two racks, and running IBM AIX - with a whopping 11.38 gigaflops of speed. And chess can be pretty much unbeatable today on an M1 MacBook Air, which comes pretty darn close to running at a teraflop.

Chess gives us an unobstructed view at the emergence of computing in an almost linear fashion. From the human powered codification of electromechanical foundations of the industry to the emergence of computational thinking with Shannon and cybernetics to MIT on IBM servers when Artificial Intelligence was young to Project MAC with Greenblatt to Bell Labs with a front seat view of Unix to college competitions to racks of IBM servers. It even has little misdirections with pre-World War II research from Konrad Zuse, who wrote chess algorithms. And the mechanical Turk concept even lives on with Amazon’s Mechanical Turk services where we can hire people to do things that are still easier for humans than machines.

The Soviet Union detonated their first nuclear bomb in 1949, releasing 20 kilotons worth of an explosion and sparking the nuclear arms race. A weather reconnaissance mission confirmed that the Soviets did so and Klaus Fuchs was arrested for espionage, after passing blueprints for the Fat Man bomb that had been dropped on Japan. A common name in the podcast is Vannevar Bush. At this point he was the president of the Carnegie Institute and put together a panel to verify the findings.

The Soviets were catching up to American science. Not only did they have a bomb but they also had new aircraft that were capable of dropping a bomb. People built bomb shelters, schools ran drills to teach students how to survive a nuclear blast and within a few years we’d moved on to the hydrogen bomb. And so the world lived in fear of nuclear fall-out.

Radar had come along during World War II and we’d developed Ground Control of Intercept, an early radar network. But that wouldn’t be enough to protect against this new threat. If one of these Soviet bombers, like the Tupolev 16 “Badger” were to come into American airspace, the prevailing thought was that we needed to shoot it down before the payload could be delivered.

The Department of Defense started simulating what a nuclear war would look like. And they asked the Air Force to develop an air defense system. Given the great work done at MIT, much under the careful eye of Vannevar Bush, they reached out to George Valley, a professor in the Physics Department who had studied nuclear weapons. He also sat on the Air Force Scientific Advisory Board, and toured some of the existing sites and took a survey of the US assets.

He sent his findings and they eventually made their way to General Vandenberg, who assigned General Fairchild to assemble a committee which would become the Valley Committee, or more officially the Air Defense Systems Engineering Committee, or ADSEC.

ADSEC dug in deeper and decided that we needed a large number of radar stations with a computer that could aggregate and then analyze data to detect enemy aircraft in real time. John Harrington had worked out how to convert radar into code and could send that over telephone lines. They just needed a computer that could crunch the data as it was received. And yet none of the computer companies at the time were able to do this kind of real time operation. We were still in a batch processing mainframe world.

Jay Forrester at MIT was working on the idea of real-time computing. Just one problem, the Servomechanisms lab where he was working on Project Whirlwind for the Navy for flight simulation was over budget and while they’d developed plenty of ground-breaking technology, they needed more funding. So Forrester was added to ADSEC and added the ability to process the digital radar information. By the end of 1950, the team was able to complete successful tests of sending radar information to Whirlwind over the phone lines.

Now it was time to get funding, which was proposed at $2 million a year to fund a lab. Given that Valley and Forrester were both at MIT, they decided it should be at MIT. Here, they saw a way to help push the electronics industry forward and the Navy’s Chief Scientist Louis Ridenour knew that wherever that lab was built would become a the next scientific hotspot. The president at MIT at the time, James Killian, wasn’t exactly jumping on the idea of MIT becoming an arm of the department of defense so put together 28 scientists to review the plans from ADSEC, which became Project Charles and threw their support to forming the new lab.

They had measured twice and were ready to cut. There were already projects being run by the military during the arms buildup named after other places surrounding MIT so they picked Project Lincoln for the name of the project to Project Lincoln. They appointed F Wheeler Loomis as the director with a mission to design a defense system. As with all big projects, they broke it up into five small projects, or divisions; things like digital computers, aircraft control and warning, and communications. A sixth did the business administration for the five technical divisions and another delivered technical services as needed.

They grew to over 300 people by the end of 1951 and over 1,300 in 1952. They moved offsite and built a new campus - thus establishing Lincoln Lab. By the end of 1953 they had written a memo called A Proposal for Air Defense System Evolution: The Technical Phase. This called for a net of radars to be set up that would track the trajectory of all aircraft in the US airspace and beyond. And to build communications to deploy the weapons that could destroy those aircraft.

The Manhattan project had brought in the nuclear age but this project grew to be larger as now we had to protect ourselves from the potential devastation we wrought. We were firmly in the Cold War with America testing the hydrogen bomb in 52 and the Soviets doing so in 55. That was the same year the prototype of the AN/FSQ-7 to replace Whirlwind.

To protect the nation from these bombs they would need 100s of radars, 24 centers to receive data, and 3 combat centers. They planned for direction centers to have a pair of AN/FSQ-7 computers, which were the Whirlwind evolved. That meant half a million lines of code which was by far the most ambitious software ever written. Forrester had developed magnetic-core memory for Whirlwind. That doubled the speed of the computer. They hired IBM to build the AN/FSQ-7 computers and from there we started to see commercial applications as well when IBM added it to the 704 mainframe in 1955.

Stalin was running labor camps and purges. An estimated nine million people died in Gulags or from hunger. Chairman Mao visited Moscow in 1957, sparking the Great Leap Forward policy that saw 45 million people die. All in the name of building a utopian paradise. Americans were scared. And Stalin was distrustful of computers for any applications beyond scientific computing for the arms race. By contrast, people like Ken Olsen from Lincoln Lab left to found Digital Equipment Corporation and sell modular mini-computers on the mass market, with DEC eventually rising to be the number two computing company in the world.

The project also needed software and so that was farmed out to Rand who would have over 500 programmers work on it. And a special display to watch planes as they were flying, which began as a Stromberg-Carlson Charactron cathode ray tube. IBM got to work building the 24 FSQ-7s, with each coming in at a whopping 250 tons and nearly 50,000 vacuum tubes - and of course that magnetic core memory.

All this wasn’t just theoretical. Given the proximity, they deployed the first net of around a dozen radars around Cape Cod as a prototype. They ran dedicated phone lines from Cambridge and built the first direction center, equipping it with an interactive display console that showed an x for each object being tracked, adding labels and then Robert Everett came up with the idea of a light gun that could be used as a pointing device, along with a keyboard, to control the computers from a terminal.

They tested the Cape Cod installation in 1953 and added long range radars in Maine and New York by the end of 1954, working out bugs as they went. The Suffolk County Airfield in Long Island was added so Strategic Air Command could start running exercises for response teams. By the end of 1955 they put the system to the test and it passed all requirements from the Air Force. The radars detected the aircraft and were able to then control manned antiaircraft operations.

By 1957 they were adding logic and capacity to the system, having fine tuned over a number of test runs until they got to a 100 percent interception rate. They were ready to build out the direction centers. The research and development phase was done - now it was time to produce an operational system. Western Electric built a network of radar and communication systems across Northern Canada that became known as the DEW line, short for Distant Early Warning.

They added increasingly complicated radar, layers of protection, like Buckminster Fuller joining for a bit to develop a geodesic dome to protect the radars using fiberglass. They added radar to what looked like oil rigs around Texas, experimented with radar on planes and ships, and how to connect those back to the main system. By the end of 1957 the system was ready to move into production and integration with live weapons into the code and connections.

This is where MIT was calling it done for their part of the program. Only problem is when the Air Force looked around for companies willing to take on such a large project, no one could. So MITRE corporation was spun out of Lincoln Labs pulling in people from a variety of other government contractors and continues on to this day working on national security, GPS, election integrity, and health care.

They took the McChord airfare online as DC-12 in 1957, then Syracuse New York in 1958 and started phasing in automated response. Andrews, Dobbins, Geiger Field, Los Angeles Air Defense Sector, and others went online over the course of the next few years. The DEW line went operational in 1962, extending from Iceland to the Aleutians. By 1963, NORAD had a Combined Operations Center where the war room became reality.

Burroughs eventually won a contract to deploy new D825 computers to form a system called BUIC II and with the rapidly changing release of new solid state technology those got replaced with a Hughes AN/TSQ-51. With the rise of Airborn Warning and Control Systems (AWACS), the ground systems started to slowly get dismantled in 1980, being phased out completely in 1984, the year after WarGames was released.

In WarGames, Matthew Broderick plays David Lightman, a young hacker who happens upon a game. One Jon Von Neumann himself might have written as he applied Game Theory to the nuclear threat. Lightman almost starts World War III when he tries to play Global Thermonuclear War. He raises the level of DEFCON and so inspires a generation of hackers who founded conferences like DEFCON and to this day war dial, or war drive, or war whatever.

The US spent countless tax money on advancing technology in the buildup for World War II and the years after. The Manhattan Project, Project Whirlwind, SAGE, and countless others saw increasing expenditures. Kennedy continued the trend in 1961 when he started the process of putting humans on the moon. And the unpopularity of the Vietnam war, which US soldiers had been dying in since 1959, caused a rollback of spending.

The legacy of these massive projects was huge spending to advance the sciences required to produce each. The need for these computers in SAGE and other critical infrastructure to withstand a nuclear war led to ARPANET, which over time evolved into the Internet. The subsequent privatization of these projects, the rapid advancement in making chips, and the drop in costs while frequent doubling of speeds based on findings from each discipline finding their way into others then gave us personal computing and the modern era of PCs then mobile devices. But it all goes back to projects like ENIAC, Whirlwind, and SAGE. Here, we can see generations of computing evolve with each project.

I’m frequently asked what’s next in our field. It’s impossible to know exactly. But we can look to mega projects, many of which are transportation related - and we can look at grants from the NSF. And DARPA and many major universities. Many of these produce new standards so we can also watch for new RFCs from the IETF. But the coolest tech is probably classified, so ask again in a few years!

And we can look to what inspires - sometimes that’s a perceived need, like thwarting nuclear war. Sometimes mapping human genomes isn’t a need until we need to rapidly develop a vaccine. And sometimes, well… sometimes it’s just returning to some sense of normalcy. Because we’re all about ready for that. That might mean not being afraid of nuclear war as a society any longer. Or not being afraid to leave our homes. Or whatever the world throws at us next.

IBM is the company with nine lives. They began out of the era of mechanical and electro-mechanical punch card computing. They helped bring the mainframe era to the commercial market. They played their part during World War II. They helped make the transistorized computer mainstream with the S360. They helped bring the PC into the home. We’ve covered a number of lost decades - and moving into the 90s, IBM was in one.

One that was largely created by an influx of revenues with the personal computer business. That revenue gave IBM a shot in the arm. But one that was temporary.

By the early 90s the computer business was under assault by the clone makers. They had been out-maneuvered by Microsoft and the writing was on the wall that Big Blue was in trouble. The CEO who presided during the fall of the hardware empire was John Akers. At the time, IBM had their fingers in every cookie jar. They were involved with instigating the Internet. They made mainframes. They made PCs. They made CPUs. They made printers. They provided services. How could they be in financial trouble? Because their core business, making computers, was becoming a commodity and quickly becoming obsolete.

IBM loves to own an industry. But they didn’t own PCs any more. They never owned PCs in the home after the PC Jr flopped. And mainframes were quickly going out of style. John Akers had been a lifer at IBM and by then there was generations of mature culture and its byproduct bureaucracy to contend with. Akers simply couldn’t move the company fast enough.

The answer was to get rid of John Akers and bring in a visionary. The visionaries in the computing field didn’t want IBM. CEOs like John Sculley at Apple and Bill Gates at Microsoft turned them down. That’s when someone at a big customer came up. Louis Gerstner. He had been the CEO of American Express and Nabisco. He had connections to IBM, with his brother having run the PC division for a time. And he was the first person brought in from the outside to run the now-nearly 100 year old company. And the first of a wave of CEOs paid big money. Commonplace today.

Starting in 1993, he moved from an IBM incapable of making decisions because of competing visions to one where execution and simplification was key. He made few changes in the beginning. At the time, competitor CDC was being split up into smaller companies and lines of business were being spun down as they faced huge financial losses. John Akers had let each division run itself - Gerstner saw the need for services given all this off-the-shelf tech being deployed in the 90s. The industry was standardizing, making it ripe for re-usable code that could run on this standardized hardware but then sold with a lot of services to customize it for each customer. In other words, it was time for IBM to become an integrator. One that could deliver a full stack of solutions.

This meant keeping the company as one powerhouse rather than breaking it up. You see, buy IBM kit, have IBM supply a service, and then IBM could use that as a wedge to sell more and more automation services into the companies. Each aspect on its own wasn’t hugely profitable, but combined - much larger deal sizes. And given IBMs piece of the internet, it was time for e-commerce. Let that Gates kid have the operating system market and the clone makers have the personal computing market in their races to the bottom. He’d take the enterprise - where IBM was known and trusted and in many sectors loved.

And he’d take what he called e-business, which we’d call eCommerce today. He brought in Irving Wladowsky-Berger and they spent six years pivoting one of the biggest companies in the world into this new strategy. The strategy also meant streamlining various operations. Each division previously had the autonomy to pick their own agency. He centralized with Ogilvy & Mather. One brand. One message.

Unlike Akers he didn’t have much loyalty to the old ways. Yes, OS/2 was made at IBM but by the time Windows 3.11 shipped, IBM was outmaneuvered and in so one of his first moves was to stop development of OS/2 in 1994. They didn’t own the operating system market so they let it go.

Cutting divisions meant there were a lot of people who didn’t fit in with the new IBM any longer. IBM had always hired people for life. Not any more. Over the course of his tenure over 100,000 people were laid off. According to Gerstner they’d grown lazy because performance didn’t really matter. And the high performers complained about the complacency. So those first two years came as a shock. But he managed to stop hemorrhaging cash and start the company back on a growth track.

Let’s put this perspective. His 9 years saw the companies market cap nearly quintuple. This in a company that was founded in 1911 so by then 72 years old. Microsoft, Dell, and so many others grew as well. But a rising tide lifts all boats. Gerstner brought ibm back. But withdrew from categories that would take over the internet. He was paid hundreds of millions of dollars for his work.

There were innovative new products in his tenure. The Simon Personal Communicator in 1994. This was one of the earliest mobile devices. Batteries and cellular technology weren’t where they needed to be just yet but it certainly represented a harbinger of things to come.

IBM introduced the PC Jr all the way back in 1983 and killed it off within two years. But they’d been selling into retail the whole time. So he killed that off and by 2005 IBM pulled out of PCs entirely, selling the division off to Lenovo.

A point I don’t think I’ve ever seen made is that Akers inherited a company embroiled in an anti-trust case. The Justice Department filed the case in 1975 and it ran until 1982 eating up thousands of hours of testimony across nearly a thousand witnesses. Akers took over in 1985 and by then IBM was putting clauses in every contract that allowed companies like Microsoft, Sierra Online, and everyone else involved with PCs to sell their software, services, and hardware to other vendors. This opened the door for the clone makers to take the market away after IBM had effectively built the ecosystem and standardized the hardware and form factors that would be used for decades.

Unlike Akers, Gerstner inherited an IBM in turmoil - and yet with some of the brightest minds in the world. They had their fingers in everything from the emerging public internet to mobile devices to mainframes to personal computers. He gave management bonuses when they did well and wasn’t afraid to cut divisions, which in his book he says that only an outsider could do. This formalized into three “personal business commitments” that contributed to IBM strategies.

He represented a shift not only at IBM but across the industry. The computer business didn’t require PhD CEOs as the previous generations had. Companies could manage the market and change cultures. Companies could focus on doing less and sell assets (like lines of business) off to raise cash to focus. Companies didn’t have to break up, as CDC had done - but instead could re-orient around a full stack of solutions for a unified enterprise. An enterprise that has been good to IBM and others who understand what they need ever since.

The IBM turnaround out of yet another lost decade showed us options for large megalith organizations that maybe previously thought different divisions had to run with more independence. Some should - not all. Most importantly though, the turnaround showed us that a culture can change. It’s one of the hardest things to do. Part of that was getting rid of the dress code and anti-alcohol policy. Part of that was performance-based comp. Part of that was to show leaders that consensus was slow and decisions needed to be made. Leaders couldn’t be perfect but a fast decision was better than one that held up business.

As with the turnaround after Apple’s lost decade, the turnaround was largely attributable to one powerful personality.

Gerstner often shied away from the media. Yet he wrote a book about his experiences called Who Says Elephants Can’t Dance. Following his time at IBM he became the chairman of the private equity firm The Carlyle Group, where he helped grow them into a powerhouse in leveraged buyouts, bringing in Hertz, Kinder Morgan, Freescale Semiconductor, Nielson Corporation, and so many others. One of the only personal tidbits you get about him in his book is that he really hates to lose. We’re all lucky he turned the company around as since he got there IBM has filed more patents than any other company for 28 consecutive years. These help push the collective conscious forward from 2,300 AI patents to 3,000 cloud patents to 1,400 security patents to laser eye surgery to quantum computing and beyond. 150,000 patents in the storied history of the company.

That’s a lot of work to bring computing into companies and increase productivity at scale. Not at the hardware level, with the constant downward pricing pressures - but at the software + services layer. The enduring legacy of the changes Gerstner made at IBM.

Today's episode on spam is read by the illustrious Joel Rennich.

Spam is irrelevant or inappropriate and unsolicited messages usually sent to a large number of recipients through electronic means. And while we probably think of spam as something new today, it’s worth noting that the first documented piece of spam was sent in 1864 - through the telegraph. With the advent of new technologies like the fax machine and telephone, messages and unsolicited calls were quick to show up.

Ray Tomlinson is widely accepted as the inventor of email, developing the first mail application in 1971 for the ARPANET. It took longer than one might expect to get abused, likely because it was mostly researchers and people from the military industrial research community. Then in 1978, Gary Thuerk at Digital Equipment Corporation decided to send out a message about the new VAX computer being released by Digital. At the time, there were 2,600 email accounts on ARPANET and his message found its way to 400 of them. That’s a little over 15% of the Internet at the time. Can you imagine sending a message to 15% of the Internet today? That would be nearly 600 million people.

But it worked. Supposedly he closed $12 million in deals despite rampant complaints back to the Defense Department. But it was too late; the damage was done. He proved that unsolicited junk mail would be a way to sell products. Others caught on. Like Dave Rhodes who popularized MAKE MONEY FAST chains in the 1988. Maybe not a real name but pyramid schemes probably go back to the pyramids so we might as well have them on the Internets.

By 1993 unsolicited email was enough of an issue that we started calling it spam. That came from the Monty Python skit where Vikings in a cafe and spam was on everything on the menu. That spam was in reference to canned meat made of pork, sugar, water, salt, potato starch, and sodium nitrate that was originally developed by Jay Hormel in 1937 and due to how cheap and easy it was found itself part of a cultural shift in America. Spam came out of Austin, Minnesota. Jay’s dad George incorporated Hormel in 1901 to process hogs and beef and developed canned lunchmeat that evolved into what we think of as Spam today. It was spiced ham, thus spam.

During World War II, Spam would find its way to GIs fighting the war and Spam found its way to England and countries the war was being fought in. It was durable and could sit on a shelf for moths. From there it ended up in school lunches, and after fishing sanctions on Japanese-Americans in Hawaii restricted the foods they could haul in, spam found its way there and some countries grew to rely on it due to displaced residents following the war. And yet, it remains a point of scorn in some cases. As the Monty Python sketch mentions, spam was ubiquitous, unavoidable, and repetitive.

Same with spam through our email. We rely on email. We need it. Email was the first real, killer app for the Internet. We communicate through it constantly. Despite the gelatinous meat we sometimes get when we expect we’re about to land that big deal when we hear the chime that our email client got a new message. It’s just unavoidable. That’s why a repetitive poster on a list had his messages called spam and the use just grew from there.

Spam isn’t exclusive to email. Laurence Canter and Martha Siegel sent the first commercial Usenet spam in the “Green Card” just after the NSF allowed commercial activities on the Internet. It was a simple Perl script to sell people on the idea of paying a fee to have them enroll people into the green card lottery. They made over $100,000 and even went so far as to publish a book on guerrilla marketing on the Internet. Canter got disbarred for illegal advertising in 1997.

Over the years new ways have come about to try and combat spam. RBLs, or using DNS blacklists to mark hosts as unable to send blacklists and thus having port 25 blocked emerged in 1996 from the Mail Abuse Prevention System, or MAPS. Developed by Dave Rand and Paul Vixie, the list of IP addresses helped for a bit. That is, until spammers realized they could just send from a different IP. Vixie also mentioned the idea of of matching a sender claim to a mail server a message came from as a means of limiting spam, a concept that would later come up again and evolve into the Sender Policy Framework, or SPF for short. That’s around the same time Steve Linford founded Spamhaus to block anyone that knowingly spams or provides services to spammers. If you have a cable modem and try to setup an email server on it you’ve probably had to first get them to unblock your address from their Don’t Route list.

The next year Mark Jeftovic created a tool called filter.plx to help filter out spam and that project got picked up by Justin Mason who uploaded his new filter to SourceForge in 2001. A filter he called SpamAssassin. Because ninjas are cooler than pirates.

Paul Graham, the co-creator of Y Combinator (and author a LISP-like programming language) wrote a paper he called “A Plan for Spam” in 2002. He proposed using a Bayesian filter as antivirus software vendors used to combat spam. That would be embraced and is one of the more common methods still used to block spam. In the paper he would go into detail around how scoring of various words would work and probabilities that compared to the rest of his email that a spam would get flagged.

That Bayesian filter would be added to SpamAssassin and others the next year. Dana Valerie Reese came up with the idea for matching sender claims independently and she and Vixie both sparked a conversation and the creation of the Anti-Spam Research Group in the IETF.

The European Parliament released the Directive on Privacy and Electronic Communications in the EU criminalizing spam. Australia and Canada followed suit.

2003 also saw the first laws in the US regarding spam. The CAN-SPAM Act of 2003 was signed by President George Bush in 2003 and allowed the FTC to regulate unsolicited commercial emails. Here we got the double-opt-in to receive commercial messages and it didn’t take long before the new law was used to prosecute spammers with Nicholas Tombros getting the dubious honor of being the first spammer convicted. What was his spam selling? Porn. He got a $10,000 fine and six months of house arrest.

Fighting spam with laws turned international. Christopher Pierson was charged with malicious communication after he sent hoax emails. And even though spammers were getting fined and put in jail all the time, the amount of spam continued to increase.

We had pattern filters, Bayesian filters, and even the threat of legal action. But the IETF Anti-Spam Research Group specifications were merged by Meng Weng Wong and by 2006 W. Schlitt joined the paper to form a new Internet standard called the Sender Policy Framework which lives on in RFC 7208. There are a lot of moving parts but at the heart of it, Simple Mail Transfer Protocol, or SMTP, allows sending mail from any connection over port 25 (or others if it’s SSL-enabled) and allowing a message to pass requiring very little information - although the sender or sending claim is a requirement.

A common troubleshooting technique used to be simply telnetting into port 25 and sending a message from an address to a mailbox on a mail server. Theoretically one could take the MX record, or the DNS record that lists the mail server to deliver mail bound for a domain to and force all outgoing mail to match that. However, due to so much spam, some companies have dedicated outbound mail servers that are different than their MX record and block outgoing mail like people might send if they’re using personal mail at work. In order not to disrupt a lot of valid use cases for mail, SPF had administrators create TXT records in DNS that listed which servers could send mail on their behalf. Now a filter could check the header for the SMTP server of a given message and know that it didn’t match a server that was allowed to send mail. And so a large chunk of spam was blocked.

Yet people still get spam for a variety of reasons. One is that new servers go up all the time just to send junk mail. Another is that email accounts get compromised and used to send mail. Another is that mail servers get compromised. We have filters and even Bayesian and more advanced forms of machine learning. Heck, sometimes we even sign up for a list by giving our email out when buying something from a reputable site or retail vendor.

Spam accounts for over 90% of the total email traffic on the Internet. This is despite blacklists, SPF, and filters. And despite the laws and threats spam continues. And it pays well. We mentioned Canter & Sigel. Shane Atkinson was sending 100 million emails per day in 2003. That doesn’t happen for free. Nathan Blecharczyk, a co-founder of Airbnb paid his way through Harvard on the back of spam.

Some spam sells legitimate products in illegitimate ways, as we saw with early IoT standard X10. Some is used to spread hate and disinformation, going back to Sender Argic, known for denying the Armenian genocide through newsgroups in 1994. Long before infowars existed. Peter Francis-Macrae sent spam to solicit buying domains he didn’t own. He was convicted after resorting to blackmail and threats. Jody Michael Smith sold replica watches and served almost a year in prison after he got caught.

Some spam is sent to get hosts loaded with malware so they could be controlled as happened with Peter Levashov, the Russian czar of the Kelihos botnet. Oleg Nikolaenko was arrested by the FBI in 2010 for spamming to get hosts in his Mega-D botnet. The Russians are good at this; they even registered the Russian Business Network as a website in 2006 to promote running an ISP for phishing, spam, and the Storm botnet. Maybe Flyman is connected to the Russian oligarchs and so continues to be allowed to operate under the radar. They remain one of the more prolific spammers.

Much is sent by a small number of spammers. Khan C. Smith sent a quarter of the spam in the world until he got caught in 2001 and fined $25 million.

Again, spam isn’t limited to just email. It showed up on Usenet in the early days. And AOL sued Chris “Rizler” Smith for over $5M for his spam on their network. Adam Guerbuez was fined over $800 million dollars for spamming Facebook. And LinkedIn allows people to send me unsolicited messages if they pay extra, probably why Microsoft payed $26 billion for the social network.

Spam has been with us since the telegraph; it isn’t going anywhere. But we can’t allow it to run unchecked. The legitimate organizations that use unsolicited messages to drive business help obfuscate the illegitimate acts where people are looking to steal identities or worse. Gary Thuerk opened a Pandora’s box that would have been opened if hadn’t of done so. The rise of the commercial Internet and the co-opting of the emerging cyberspace as a place where privacy and so anonymity trump verification hit a global audience of people who are not equal. Inequality breeds crime. And so we continually have to rethink the answers to the question of sovereignty versus the common good. Think about that next time an IRS agent with a thick foreign accent calls asking for your social security number - and remember (if you’re old enough) that we used to show our social security cards to grocery store clerks when we wrote checks. Can you imagine?!?!

The simple story of Yahoo! Is that they were an Internet search company that came out of Stanford during the early days of the web. They weren’t the first nor the last. But they represent a defining moment in the rise of the web as we know it today, when there was enough content out there that there needed to be an easily searchable catalog of content.

And that’s what Stanford PhD students David Philo and Jerry Yang built. As with many of those early companies it began as a side project called “Jerry and David's Guide to the World Wide Web.” And grew into a company that at one time rivaled any in the world. At the time there were other search engines and they all started adding portal aspects to the site growing fast until the dot-com bubble burst. They slowly faded until being merged with another 90s giant, AOL, in 2017 to form Oath, which got renamed to Verizon Media in 2019 and then effectively sold to investment management firm Apollo Global Management in 2021.

Those early years were wild. Yang moved to San Jose in the 70s from Taiwan, and earned a bachelors then a masters at Stanford - where he met David Filo in 1989. Filo is a Wisconsin kid who moved to Stanford and got his masters in 1990. The two went to Japan in 1992 on an exchange program and came home to work on their PhDs. That’s when they started surfing the web. Within two years they started their Internet directory in 1994. As it grew they hosted the database on Yang’s student computer called akebono and the search engine on konishiki, which was Filo’s.

They renamed it to Yahoo, short for Yet Another Hierarchical Officious Oracle - after all they maybe considered themselves Yahoos at the time. And so Yahoo began life as akebono.stanford.edu/~yahoo. Word spread fast and they’d already had a million hits by the end of 1994. It was time to move out of Stanford.

Mark Andreesen offered to let them move into Netscape. They bought a domain in 1995 and incorporated the company, getting funding from Sequoia Capital raising $3,000,000. They tinkered with selling ads on the site to fund buying more servers but there was a lot of businessing. They decided that they would bring in Tim Koogle (which ironically rhymes with Google) to be CEO who brought in Jeff Mallett from Novell’s consumer division to be the COO. They were the suits and got revenues up to a million dollars.

The idea of the college kids striking gold fueled the rise of other companies and Yang and Filo became poster children. Applications from all over the world for others looking to make their mark started streaming in to Stanford - a trend that continues today. Yet another generation was about to flow into Silicon Valley. First the chip makers, then the PC hobbyists turned businesses, and now the web revolution.

But at the core of the business were Koogle and Mallett, bringing in advertisers and investors. And the next year needing more and more servers and employees to fuel further expansion, they went public, selling over two and a half million shares at $13 to raise nearly $34 million. That’s just one year after a gangbuster IPO from Netscape. The Internet was here. Revenues shot up to $20 million.

A concept we repeatedly look at is the technological determinism that industries go through. At this point it’s easy to look in the rear view mirror and see change coming at us. First we document information - like Jerry and David building a directory. Then we move it to a database so we can connect that data. Thus a search engine. Given that Yahoo! was a search engine they were already on the Internet. But the next step in the deterministic application of modern technology is to replace human effort with increasingly sophisticated automation. You know, like applying basic natural language processing, classification, and polarity scoring algorithms to enrich the human experience.

Yahoo! hired “surfers” to do these tasks. They curated the web. Yes, they added feeds for news, sports, finance, and created content. Their primary business model was to sell banner ads. And they pioneered the field. Banner ads mean people need to be on the site to see them. So adding weather, maps, shopping, classifieds, personal ads, and even celebrity chats were natural adjacencies given that mental model. Search itself was almost a competitor, sending people to other parts of the web that they weren’t making money off eyeballs. And they were pushing traffic to over 65 million pages worth of data a day.

They weren’t the only ones. This was the portal era of search and companies like Lycos, Excite, and InfoSeek were following the same model. They created local directories and people and companies could customize the look and feel. Their first designer, David Shen, takes us through the user experience journey in his book Takeover! The Inside Story the Yahoo Ad Revolution. They didn’t invent pay-per-clic advertising but did help to make it common practice and proved that money could be made on this whole new weird Internet thing everyone was talking about.

The first ad they sold was for MCI and from there they were practically printing money. Every company wanted in on the action - and sales just kept going up. Bill Clinton gave them a spot in the Internet Village during his 1997 inauguration and they were for a time seemingly synonymous with the Internet.

The Internet was growing fast. Cataloging the Internet and creating content for the Internet became a larger and larger manual task. As did selling ads, which was a manual transaction requiring a larger and larger sales force. As with other rising internet properties, people dressed how they wanted, they’d stay up late building code or content and crash at the desk. They ran funny cheeky ads with that yodel - becoming a brand that people knew and many equated to the Internet. We can thank San Francisco’s Black Rocket ad agency for that.

They grew fast. The founders made several strategic acquisitions and gobbled up nearly every category of the Internet that has each grown to billions of dollars. They bought Four 11 for $95 million in their first probably best acquisition, and used them to create Yahoo! Mail in 1997 and a calendar in 1998. They had over 12 million Yahoo! Email users by he end of the year, inching their way to the same number of AOL users out there.

There were other tools like Yahoo Briefcase, to upload files to the web. Now common with cloud storage providers like Dropbox, Box, Google Drive, and even Office 365. And contacts and Messenger - a service that would run until 2018. Think of all the messaging apps that have come with their own spin on the service since.

1998 also saw the acquisition of Viaweb, founded by the team that would later create Y Combinator. It was just shy of a $50M acquisition that brought the Yahoo! Store - which was similar to the Shopify of today. They got a $250 million investment from Softbank, bought Yoyodyne, and launched AT&T’s WorldNet service to move towards AOL’s dialup services. By the end of the year they were closing in on 100 million page views a day. That’s a lot of banners shown to visitors. But Microsoft was out there, with their MSN portal at the height of the browser wars.

Yahoo! bought Broadcast.com in 1999 saddling the world with Mark Cuban. They dropped $5.7 billion for 300 employees and little more than an ISDN line. Here, they paid over a 100x multiple of annual revenues and failed to transition sellers into their culture. Sales cures all. In his book We Were Yahoo! Jeremy Ring describes the lays much of the blame of the failure to capitalize on the acquisition as not understanding the different selling motion. I don’t remember him outright saying it was hubris, but he certainly indicates that it should have worked out and that broadcast.com was could have been what YouTube would become. Another market lost in a failed attempt at Yahoo TV. And yet many of these were trends started by AOL.

They also bought GeoCities in 99 for $3.7 billion. Others have tried to allow for fast and easy site development - the no code wysiwyg web. GeoCities lasted until 2009 - a year after Google launched Google Sites. And we have Wix, Squarespace, WordPress, and so many others offering similar services today.

As they grew some of the other 130+ search engines at the time folded. The new products continued. The Yahoo Notebook came before Evernote. Imagine your notes accessible to any device you could log into. The more banners shown, the more clicks. Advertisers could experiment in ways they’d never been able to before. They also inked distribution deals, pushing traffic to other site that did things they didn’t.

The growth of the Internet had been fast, with nearly 100 million people armed with Internet access - and yet it was thought to triple in just the next three years. And even still many felt a bubble was forming. Some, like Google, had conserved cash - others like Yahoo! Had spent big on acquisitions they couldn’t monetize into truly adjacent cash flow generating opportunities. And meanwhile they were alienating web properties by leaning into every space that kept eyeballs on the site.

By 2000 their stock traded at $118.75 and they were the most valuable internet company at $125 billion. Then as customers folded when the dot-com bubble burst, the stock fell to $8.11 the next year. One concept we talk about in this podcast is a lost decade. Arguably they’d entered into theirs around the time the dot-com bubble burst. They decided to lean into being a media company even further. Again, showing banners to eyeballs was the central product they sold.

They brought in Terry Semel in 2001 using over $100 million in stock options to entice him. And the culture problems came fast. Semel flew in a fancy jet, launched television shows on Yahoo! and alienated programmers, effectively creating an us vs them and de-valuing the work done on the portal and search. Work that could have made them competitive with Google Adwords that while only a year old was already starting to eat away at profits. But media.

They bought a company called LaunchCast in 2001, charging a monthly fee to listen to music. Yahoo Music came before Spotify, Pandora, Apple Music, and even though it was the same year the iPod was released, they let us listen to up to 1,000 songs for free or pony up a few bucks a month to get rid of ads and allow for skips. A model that has been copied by many over the years.

By then they knew that paid search was becoming a money-maker over at Google. Overture had actually been first to that market and so Yahoo! Bought them for $1.6 billion in 2003. But again, they didn’t integrate the team and in a classic “not built here” moment started Project Panama where they’d spend three years building their own search advertising platform. By the time that shipped the search war was over and executives and great programmers were flowing into other companies all over the world.

And by then they were all over the world. 2005 saw them invest $1 billion in a little company called Alibaba. An investment that would accelerate Alibaba to become the crown jewel in Yahoo’s empire and as they dwindled away, a key aspect of what led to their final demise.

They bought Flickr in 2005 for $25M. User generated content was a thing. And Flickr was almost what Instagram is today. Instead we’d have to wait until 2010 for Instagram because Flickr ended up yet another of the failed acquisitions. And here’s something wild to thin about - Stewart Butterfield and Cal Henderson started another company after they sold Flickr. Slack sold to Salesforce for over $27 billion. Not only is that a great team who could have turned Flickr into something truly special, but if they’d been retained and allowed to flourish at Yahoo! they could have continued building cooler stuff. Yikes. Additionally, Flickr was planning a pivot into social networking, right before a time when Facebook would take over that market.

If fact, they tried to buy Facebook for just over a billion dollars in 2006. But Zuckerberg walked away when the price went down after the stock fell. They almost bought YouTube and considered buying Apple, which is wild to think about today. Missed opportunities. And Semmel was the first of many CEOs who lacked vision and the capacity to listen to the technologists - in a technology company.

These years saw Comcast bring us weather.com, the rise of espn online taking eyeballs away from Yahoo! Sports, Gmail and other mail services reducing reliance on Yahoo! Mail. Facebook, LinkedIn, and other web properties rose to take ad placements away. Even though Yahoo Finance is still a great portal even sites like Bloomberg took eyeballs away from them. And then there was the rise of user generated content - a blog for pretty much everything. Jerry Yang came back to run the show in 2007 then Carol Bartz from 2009 to 2011 then Scott Thompson in 2012. None managed to turn things around after so much lost inertia - and make no mistake, inertia is the one thing that can’t be bought in this world.

Wisconsin’s Marissa Mayer joined Yahoo! In 2012. She was Google’s 20th employee who’d risen through the ranks from writing code to leading teams to product manager to running web products and managing not only the layout of that famous homepage but also helped deliver Google AdWords and then maps. She had the pedigree and managerial experience - and had been involved in M&A. There was an immediate buzz that Yahoo! was back after years of steady decline due to incoherent strategies and mismanaged acquisitions.

She pivoted the business more into mobile technology. She brought remote employees back into the office. She implemented a bell curve employee ranking system like Microsoft did during their lost decade. They bought Tumblr in 2013 for $1.1 billion. But key executives continued to leave - Tumbler’s value dropped, and the stock continued to drop. Profits were up, revenues were down.

Investing in the rapidly growing China market became all the rage. The Alibaba investment was now worth more than Yahoo! itself. Half the shares had been sold back to Alibaba in 2012 to fund Yahoo! pursuing the Mayer initiatives. And then there was Yahoo Japan, which continued to do well. After years of attempts, activist investors finally got Yahoo! to spin off their holdings. They moved most of the shares to a holding company which would end up getting sold back to Alibaba for tens of billions of dollars. More missed opportunities for Yahoo!

And so in the end, they would get merged with AOL - the two combined companies worth nearly half a trillion dollars at one point to become Oath in 2017. Mayer stepped down and the two sold for less than $5 billion dollars. A roller coaster that went up really fast and down really slow. An empire that crumbled and fragmented.

Arguably, the end began in 1998 when another couple of grad students at Stanford approached Yahoo to buy Google for $1M. Not only did Filo tell them to try it alone but he also introduced them to Michael Moritz of Sequoia - the same guy who’d initially funded Yahoo!. That wasn’t where things really got screwed up though. It was early in a big change in how search would be monetized. But they got a second chance to buy Google in 2002. By then I’d switched to using Google and never looked back. But the CEO at the time, Terry Semel, was willing to put in $3B to buy Google - who decided to hold out for $5B. They are around a $1.8T company today.

Again, the core product was selling advertising. And Microsoft tried to buy Yahoo! In 2008 for over 44 billion dollars to become Bing. Down from the $125 billion height of the market cap during the dot com bubble. And yet they eventually sold for less than four and a half billion in 2016 and went down in value from there. Growth stocks trade at high multiples but when revenues go down the crash is hard and fast.

Yahoo! lost track of the core business - just as the model was changing. And yet never iterated it because it just made too much money. They were too big to pivot from banners when Google showed up with a smaller, more bite-sized advertising model that companies could grow into.

Along the way, they tried to do too much. They invested over and over in acquisitions that didn’t work because they ran off the innovative founders in an increasingly corporate company that was actually trying to pretend not to be. We have to own who we are and become. And we have to understand that we don’t know anything about the customers of acquired companies and actually listen - and I mean really listen - when we’re being told what those customers want. After all, that’s why we paid for the company in the first place.

We also have to avoid allowing the market to dictate a perceived growth mentality. Sure a growth stock needs to hit a certain number of revenue increase to stay considered a growth stock and thus enjoy the kind of multiples for market capitalization. But that can drive short term decisions that don’t see us investing in areas that don’t effectively manipulate stocks. Decisions like trying to keep eyeballs on pages with our own content rather than investing in the user generated content that drove the Web 2.0 revolution.

The Internet can be a powerful medium to find information, allow humans to do more with less, and have more meaningful experiences in this life. But just as Yahoo! was engineering ways to keep eyeballs on their pages, the modern Web 2.0 era has engineered ways to keep eyeballs on our devices. And yet what people really want is those meaningful experiences, which happen more when we aren’t staring at our screens than when we are. As I look around at all the alerts on my phone and watch, I can’t help but wonder if another wave of technology is coming that disrupts that model. Some apps are engineered to help us lead healthier lifestyles and take a short digital detoxification break.

Bush’s Memex in “As We May Think” was arguably an Apple taken from the tree of knowledge. If we aren’t careful, rather than the dream of computers helping humanity do more and free our minds to think more deeply we are simply left with less and less capacity to think and less and less meaning. The Memex came and Yahoo! helped connect us to any content we might want in the world. And yet, like so many others, they stalled in the phase they were at in that deterministic structure that technologies follow. Too slow to augment human labor with machine learning like Google did - but instead too quick to try and do everything for everyone with no real vision other than be everything to everyone. And so the cuts went on slowly for a long time, leaving employees constantly in fear of losing their jobs.

As you listen to this if I were to leave a single parting thought - it would be that companies should always be willing to cannibalize their own businesses. And yet we have to have a vision that our teams rally behind for how that revenue gets replaced. We can’t fracture a company and just sprawl to become everything for everyone but instead need to be targeted and more precise. And to continue to innovate each product beyond the basic machine learning and into deep learning and beyond. And when we see those who lack that focus, don’t get annoyed but instead get stoked - that’s called a disruptive opportunity.

And if there’s someone with 1,000 developers in a space, Nicholas Carlson in his book “Marissa Mayer and the Fight To Save Yahoo!” points out that one great developer is worth a thousand average ones. And even the best organizations can easily turn great developers into average ones for a variety of reason. Again, we can call these opportunities.

Yahoo! helped legitimize the Internet. For that we owe them a huge thanks. And we can fast follow their adjacent expansions to find a slew of great and innovative ideas that increased the productivity of humankind. We owe them a huge thanks for that as well. Now what opportunities do we see out there to propel us further yet again?

What is the nature of innovation? Is it overhearing a conversation as with Morse and the telegraph? Working with the deaf as with Bell? Divine inspiration? Necessity? Science fiction? Or given that the answer to all of these is yes, is it really more the intersectionality between them and multiple basic and applied sciences with deeper understandings in each domain? Or is it being given the freedom to research? Or being directed to research? Few have as storied a history of innovation as Bell Labs and few have had anything close to the impact.

Bell Labs gave us 9 Nobel Prizes and 5 Turing awards. Their alumni have even more, but those were the ones earned while at Bell. And along the way they gave us 26,000 patents. They researched, automated, and built systems that connected practically every human around the world - moving us all into an era of instant communication. It’s a rich history that goes back in time from the 2018 Ashkin Nobel for applied optical tweezers and 2018 Turing award for Deep Learning to an almost steampunk era of tophats and the dawn of the electrification of the world.

Those late 1800s saw a flurry of applied and basic research. One reason was that governments were starting to fund that research. Alessandro Volta had come along and given us the battery and it was starting to change the world. So Napolean’s nephew, Napoleon III, during the second French Empire gave us the Volta Prize in 1852.

One of those great researchers to receive the Volta Prize was Alexander Graham Bell. He invented the telephone in 1876 and was awarded the Volta Prize, getting 50,000 francs. He used the money to establish the Volta Laboratory, which would evolve or be a precursor to a research lab that would be called Bell Labs. He also formed the Bell Patent Association in 1876. They would research sound. Recording, transmission, and analysis - so science.

There was a flurry of business happening in preparation to put a phone in every home in the world. We got the Bell System, The Bell Telephone Company, American Bell Telephone Company patent disputes with Elisha Gray over the telephone (and so the acquisition of Western Electric), and finally American Telephone and Telegraph, or AT&T. Think of all this as Ma’ Bell. Not Pa’ Bell mind you - as Graham Bell gave all of his shares except 10 to his new wife when they were married in 1877. And her dad ended up helping build the company and later creating National Geographic, even going international with International Bell Telephone Company. Bell’s assistant Thomas Watson sold his shares off to become a millionaire in the 1800s, and embarking on a life as a Shakespearean actor.

But Bell wasn’t done contributing. He still wanted to research all the things. Hackers gotta’ hack. And the company needed him to - keep in mind, they were a cutting edge technology company (then as in now). That thirst for research would infuse AT&T - with Bell Labs paying homage to the founder’s contribution to the modern day. Over the years they’d be on West Street in New York and expand to have locations around the US. Think about this: it was becoming clear that automation would be able to replace human efforts where electricity is concerned. The next few decades gave us the vacuum tube, flip flop circuits, mass deployment of radio. The world was becoming ever so slightly interconnected. And Bell Labs was researching all of it. From physics to the applied sciences.

By the 1920s, they were doing sound synchronized with motion and shooting that over long distances and calculating the noise loss. They were researching encryption. Because people wanted their calls to be private. That began with things like one-time pad cyphers but would evolve into speech synthesizers and even SIGSALY, the first encrypted (or scrambled) speech transmission that led to the invention of the first computer modem. They had engineers like Harry Nyquist, whose name is on dozens of theories, frequencies, even noise. He arrived in 1917 and stayed until he retired in 1954. One of his most important contributions was to move beyond printing telegraph to paper tape and to helping transmit pictures over electricity - and Herbert Ives from there sent color photos, thus the fax was born (although it would be Xerox who commercialized the modern fax machine in the 1960s).

Nyquist and others like Ralph Hartley worked on making audio better, able to transmit over longer lines, reducing feedback, or noise. While there, Hartley gave us the oscillator, developed radio receivers, parametric amplifiers, and then got into servomechanisms before retiring from Bell Labs in 1950. The scientists who’d been in their prime between the two world wars were titans and left behind commercializable products, even if they didn’t necessarily always mean to.

By the 40s a new generation was there and building on the shoulders of these giants. Nyquist’s work was extended by Claude Shannon, who we devoted an entire episode to. He did a lot of mathematical analysis like writing “A Mathematical Theory of Communication” to birth Information Theory as a science.

They were researching radio because secretly I think they all knew those leased lines would some day become 5G. But also because the tech giants of the era included radio and many could see a day coming when radio, telephony, and aThey were researching how electrons diffracted, leading to George Paget Thomson receiving the Nobel Prize and beginning the race for solid state storage.

Much of the work being done was statistical in nature. And they had William Edwards Deming there, whose work on statistical analysis when he was in Japan following World War II inspired a global quality movement that continues to this day in the form of frameworks like Six Sigma and TQM. Imagine a time when Japanese manufacturing was of such low quality that he couldn’t stay on a phone call for a few minutes or use a product for a time. His work in Japan’s reconstruction paired with dedicated founders like Akio Morita, who co-founded Sony, led to one of the greatest productivity increases, without sacrificing quality, of any time in the world. Deming would change the way Ford worked, giving us the “quality culture.”

Their scientists had built mechanical calculators going back to the 30s (Shannon had built a differential analyzer while still at MIT) - first for calculating the numbers they needed to science better then for ballistic trajectories, then with the Model V in 1946, general computing. But these were slow; electromechanical at best.

Mary Torrey was another statistician of the era who along with Harold Hodge gave us the theory of acceptance sampling and thus quality control for electronics. And basic electronics research to do flip-flop circuits fast enough to establish a call across a number of different relays was where much of this was leading. We couldn’t use mechanical computers for that, and tubes were too slow. And so in 1947 John Bardeen, Walter Brattain, and William Shockley invented the transistor at Bell Labs, which be paired with Shannon’s work to give us the early era of computers as we began to weave Boolean logic in ways that allowed us to skip moving parts and move to a purely transistorized world of computing.

In fact, they all knew one day soon, everything that monster ENIAC and its bastard stepchild UNIVAC was doing would be done on a single wafer of silicon. But there was more basic research to get there. The types of wires we could use, the Marnaugh map from Maurice Karnaugh, zone melting so we could do level doping. And by 1959 Mohamed Atalla and Dawon Kahng gave us metal-oxide semiconductor field-effect transistors, or MOSFETs - which was a step on the way to large-scale integration, or LSI chips. Oh, and they’d started selling those computer modems as the Bell 101 after perfecting the tech for the SAGE air-defense system.

And the research to get there gave us the basic science for the solar cell, electronic music, and lasers - just in the 1950s. The 1960s saw further work work on microphones and communication satellites like Telstar, which saw Bell Labs outsource launching satellites to NASA. Those transistors were coming in handy, as were the solar panels. The 14 watts produced certainly couldn’t have moved a mechanical computer wheel. Blaise Pascal and would be proud of the research his countries funds inspired and Volta would have been perfectly happy to have his name still on the lab I’m sure. Again, shoulders and giants. Telstar relayed its first television signal in 1962. The era of satellites was born later that year when Cronkite televised coverage of Kennedy manipulating world markets on this new medium for the first time and IBM 1401 computers encrypted and decrypted messages, ushering in an era of encrypted satellite communications. Sputnik may heave heated the US into orbit but the Telstar program has been an enduring system through to the Telstar 19V launched in 2018 - now outsourced to a Falcon 9 rocket from Space X.

It might seem like Bell Labs had done enough for the world. But they still had a lot of the basic wireless research to bring us into the cellular age. In fact, they’d plotted out what the cellular age would look like all the way back in 1947!

The increasing use of computers to do the all the acoustics and physics meant they were working closely with research universities during the rise of computing. They were involved in a failed experiment to create an operating system in the late 60s. Multics influenced so much but wasn’t what we might consider a commercial success. It was the result of yet another of DARPA’s J.C.R. Licklider’s wild ideas in the form of Project MAC, which had Marvin Minsky and John McCarthy. Big names in the scientific community collided with cooperation and GE, Bell Labs and Multics would end up inspiring many a feature of a modern operating system.

The crew at Bell Labs knew they could do better and so set out to take the best of Multics and implement a lighter, easier operating system. So they got to work on Uniplexed Information and Computing Service, or Unics, which was a pun on Multics. Ken Thompson, Dennis Ritchie, Doug McIllroy, Joe Assana, Brian Kernigan, and many others wrote Unix originally in assembly and then rewrote it in C once Dennis Ritchie wrote that to replace B. Along the way, Alfred Aho, Peter Weinber, and Kernighan gave us AWSK and with all this code they needed a way to keep the source under control so Marc Rochkind gave us the SCCS, or Course Code Control System, first written for an IBM S/3370 and then ported to C - which would be how most environments maintained source code until CVS came along in 1986. And Robert Fourer, David Gay, and Brian Kernighan wrote A Mathematical Programming Language, or AMPL, while there.

Unix began as a bit of a shadow project but would eventually go to market as Research Unix when Don Gillies left Bell to go to the University of Illinois at Champaign-Urbana. From there it spread and after it fragmented in System V led to the rise of IBM’s AIX, HP-UX, SunOS/Solaris, BSD, and many other variants - including those that have evolved into the macOS through Darwin, and Android through Linux. But Unix wasn’t all they worked on - it was a tool to enable other projects. They gave us the charge-coupled device, which resulted in yet another Nobel Prize. That is an image sensor built on the MOS technologies. While fiber optics goes back to the 1800s, they gave us attenuation over fiber and thus could stretch cables to only need repeaters every few dozen miles - again reducing the cost to run the ever-growing phone company.

All of this electronics allowed them to finally start reducing their reliance on electromechanical and human-based relays to transistor-to-transistor logic and less mechanical meant less energy, less labor to repair, and faster service. Decades of innovation gave way to decades of profit - in part because of automation. The 5ESS was a switching system that went online in 1982 and some of what it did - its descendants still do today. Long distance billing, switching modules, digital line trunk units, line cards - the grid could run with less infrastructure because the computer managed distributed switching. The world was ready for packet switching.

5ESS was 100 million lines of code, mostly written in C. All that source was managed with SCCS. Bell continued with innovations. They produced that modem up into the 70s but allowed Hayes, Rockewell, and others to take it to a larger market - coming back in from time to time to help improve things like when Bell Labs, branded as Lucent after the breakup of AT&T, helped bring the 56k modem to market.

The presidents of Bell Labs were as integral to the success and innovation as the researchers. Frank Baldwin Jewett from 1925 to 1940, Oliver Buckley from 40 to 51, the great Mervin Kelly from 51 to 59, James Fisk from 59 to 73, William Oliver Baker from 73 to 79, and a few others since gave people like Bishnu Atal the space to develop speech processing algorithms and predictive coding and thus codecs. And they let Bjarne Stroustrup create C++, and Eric Schmidt who would go on to become a CEO of Google and the list goes on. Nearly every aspect of technology today is touched by the work they did.

All of this research. Jon Gerstner wrote a book called The Idea Factory: Bell Labs and the Great Age of American Innovation. He chronicles the journey of multiple generations of adventurers from Germany, Ohio, Iowa, Japan, and all over the world to the Bell campuses. The growth and contraction of the basic and applied research and the amazing minds that walked the halls. It’s a great book and a short episode like this couldn’t touch the aspects he covers. He doesn’t end the book as hopeful as I remain about the future of technology, though.

But since he wrote the book, plenty has happened. After the hangover from the breakup of Ma Bell they’re now back to being called Nokia Bell Labs - following a $16.6 billion acquisition by Nokia. I sometimes wonder if the world has the stomach for the same level of basic research. And then Alfred Aho and Jeffrey Ullman from Bell end up sharing the Turing Award for their work on compilers. And other researchers hit a terabit a second speeds. A storied history that will be a challenge for Marcus Weldon’s successor. He was there as a post-doc there in 1995 and rose to lead the labs and become the CTO of Nokia - he said the next regeneration of a Doctor Who doctor would come in after him. We hope they are as good of stewards as those who came before them.

The world is looking around after these decades of getting used to the technology they helped give us. We’re used to constant change. We’re accustomed to speed increases from 110 bits a second to now terabits. The nature of innovation isn’t likely to be something their scientists can uncover. My guess is Prometheus is guarding that secret - if only to keep others from suffering the same fate after giving us the fire that sparked our imaginations. For more on that, maybe check out Hesiod’s Theogony.

In the meantime, think about the places where various sciences and disciplines intersect and think about the wellspring of each and the vast supporting casts that gave us our modern life. It’s pretty phenomenal when ya’ think about it.

Once upon a time, people were computers. It’s probably hard to imagine teams of people spending their entire day toiling in large grids of paper, writing numbers and calculating numbers by hand or with mechanical calculators, and then writing more numbers and then repeating that. But that’s the way it was before the 1979. 

The term spreadsheet comes from back when a spread, like a magazine spread, of ledger cells for bookkeeping. There’s a great scene in the Netflix show Halston where a new guy is brought in to run the company and he’s flying through an electro-mechanical calculator. Halston just shuts the door. Ugh. Imagine doing what we do in a spreadsheet in minutes today by hand. Even really large companies jump over into a spreadsheet to do financial projections today - and with trendlines, tweaking this small variable or that, and even having different algorithms to project the future contents of a cell - the computerized spreadsheet is one of the most valuable business tools ever built. It’s that instant change we see when we change one set of numbers and can see the impact down the line. 

Even with the advent of mainframe computers accounting and finance teams had armies of people who calculated spreadsheets by hand, building complicated financial projections. If the formulas changed then it could take days or weeks to re-calculate and update every cell in a workbook. People didn’t experiment with formulas. Computers up to this point had been able to calculate changes and provided all the formulas were accurate could output results onto punch cards or printers. But the cost had been in the millions before Digital Equipment and Data Nova came along and had dropped into the tens or hundreds of thousands of dollars 

The first computerized spreadsheets weren’t instant. Richard Mattessich developed an electronic, batch spreadsheet in 1961. He’d go on to write a book called “Simulation of the Firm Through a Budget Computer Program.” His work was more theoretical in nature, but IBM developed the Business Computer Language, or BCL the next year. What IBM did got copied by their seven dwarves. former GE employees Leroy Ellison, Harry Cantrell, and Russell Edwards developed AutoPlan/AutoTab, another scripting language for spreadsheets, following along delimited files of numbers. And in 1970 we got LANPAR which opened up more than reading files in from sequential, delimited sources.

But then everything began to change. Harvard student Dan Bricklin graduated from MIT and went to work for Digital Equipment Corporation to work on an early word processor called WPS-8. We were now in the age of interactive computing on minicomputers. He then went to work for FasFax in 1976 for a year, getting exposure to calculating numbers. And then he went off to Harvard in 1977 to get his MBA. But while he was at Harvard he started working on one of the timesharing programs to help do spreadsheet analysis and wrote his own tool that could do five columns and 20 rows. Then he met Bob Frankston and they added Dan Fylstra, who thought it should be able to run on an Apple - and so they started Software Arts Corporation.

Frankston got the programming bug while sitting in on a class during junior high. He then got his undergrad and Masters at MIT, where he spent 9 years in school and working on a number of projects with CSAIL, including Multics. He’d been consulting and working at various companies for awhile in the Boston area, which at the time was probably the major hub.

Frankston and Bricklin would build a visible calculator using 16k of space and that could fit on a floppy. They used a time sharing system and because they were paying for time, they worked at nights when time was cheaper, to save money. They founded a company called Software Arts and named their Visual Calculator VisiCalc. Along comes the Apple II. And computers were affordable. They ported the software to the platform and it was an instant success. It grew fast.

Competitors sprung up. SuperCalc in 1980, bundled with the Osborne. The IBM PC came in 1981 and the spreadsheet appeared in Fortune for the first time. Then the cover of Inc Magazine in 1982. Publicity is great for sales and inspiring competitors. Lotus 1-2-3 came in 1982 and even Boeing Computer Services got in the game with Boeing Calc in 1985. They extended the ledger metaphor to add sheets to the spreadsheet, which we think of as tabs today.

Quattro Pro from Borland copied that feature and despite having their offices effectively destroyed during an earthquake just before release, came to market in 1989. Ironically they got the idea after someone falsely claimed they were making a spreadsheet a few years earlier.

And so other companies were building Visible Calculators and adding new features to improve on the spreadsheet concept. Microsoft was one who really didn’t make a dent in sales at first. They released an early spreadsheet tool called Multiple in 1982. But Lotus 1-2-3 was the first killer application for the PC. 

It was more user friendly and didn’t have all the bugs that had come up in VisiCalc as it was ported to run on platform after platform. Lotus was started by Mitch Kapor who brought Jonathan Sachs in to develop the spreadsheet software. Kapor’s marketing prowess would effectively obsolete VisiCalc in a number of environments. They made TV commercials so you know they were big time! And they were written natively in the x86 assembly so it was fast. They added the ability to add bar charts, pie charts, and line charts. They added color and printing. One could even spread their sheet across multiple monitors like in a magazine.

It was 1- spreadsheets, 2 - charts and graphs and 3 - basic database functions. Heck, one could even change the size of cells and use it as a text editor. Oh, and macros would become a standard in spreadsheets after Lotus.

And because VisiCalc had been around so long, Lotus of course was immediately capable of reading a VisiCalc file when released in 1983. As could Microsoft Excel, when it came along in 1985. And even Boeing Calc could read Lotus 1-2-3 files. After all, the concept went back to those mainframe delimited files and to this day we can import and export to tab or comma delimited files.

VisiCalc had sold about a million copies but that would cease production the same year Excel was released, although the final release had come in 1983. Lotus had eaten their shorts in the market, and Borland had watched. Microsoft was about to eat both of theirs. Why? Visi was about to build a windowing system called Visi-On. And Steve Jobs needed a different vendor to turn to. He looked to Lotus who built a tool called Jazz that was too basic. But Microsoft had gone public in 1985 and raised plenty of money, some of which they used to complete Excel for the Mac that year. Their final release in 1983 began to fade away

And so Excel began on the Mac and that first version was the first graphical spreadsheet. The other developers didn’t think that a GUI was gonna’ be much of a thing. Maybe graphical interfaces were a novelty! Version two was released for the PC in 1987 along with Windows 2.0. Sales were slow at first. But then came Windows 3. Add Microsoft Word to form Microsoft Office and by the time Windows 95 was released Microsoft became the de facto market leader in documents and spreadsheets. That’s the same year IBM bought Lotus and they continued to sell the product until 2013, with sales steadily declining.

And so without a lot of competition for Microsoft Excel, spreadsheets kinda’ sat for a hot minute. Computers became ubiquitous. Microsoft released new versions for Mac and Windows but they went into that infamous lost decade until… competition. And there were always competitors, but real competition with something new to add to the mix. Google bought a company called 2Web Technologies in 2006, who made a web-based spreadsheet called XL2WEB. That would become Google Sheets. Google bought DocVerse in 2010 and we could suddenly have multiple people editing a sheet concurrently - and the files were compatible with Excel.

By 2015 there were a couple million users of Google Workspace, growing to over 5 million in 2019 and another million in 2020. In the years since, Microsoft released Office 365, starting to move many of their offerings onto the web. That involved 60 million people in 2015 and has since grown to over 250 million. The statistics can be funny here, because it’s hard to nail down how many free vs paid Google and Microsoft users there are. Statista lists Google as having a nearly 60% market share but Microsoft is clearly making more from their products. And there are smaller competitors all over the place taking on lots of niche areas.

There are a few interesting tidbits here. One is that the tools that there’s a clean line of evolution in features. Each new tool worked better, added features, and they all worked with previous file formats to ease the transition into their product. Another is how much we’ve all matured in our understanding of data structures. I mean we have rows and columns. And sometimes multiple sheets - kinda’ like multiple tables in a database. Our financial modeling and even scientific modeling has grown in acumen by leaps and bounds. 

Many still used those electro-mechanical calculators in the 70s when you could buy calculator kits and build your own calculator. Those personal computers that flowed out in the next few years gave every business the chance to first track basic inventory and calculate simple information, like how much we might expect in revenue from inventory in stock to now thousands of pre-built formulas that are supported across most spreadsheet tooling. Despite expensive tools and apps to do specific business functions, the spreadsheet is still one of the most enduring and useful tools we have. Even for programmers, where we’re often just getting our data in a format we can dump into other tools!

So think about this. What tools out there have common file types where new tools can sit on top of them? Which of those haven’t been innovated on in a hot minute? And of course, what is that next bold evolution? Is it moving the spreadsheet from a book to a batch process? Or from a batch process to real-time? Or from real-time to relational with new tabs? Or to add a GUI? Or adding online collaboration? Or like some big data companies using machine learning to analyze the large data sets and look for patterns automatically? 

Not only does the spreadsheet help us do the maths - it also helps us map the technological determinism we see repeated through nearly every single tool for any vertical or horizontal market. Those stuck need disruptive competitors if only to push them off the laurels they’ve been resting on. 

Microsoft went from a fledgeling purveyor of a BASIC for the Altair to a force to be reckoned with. The biggest growth hack was when they teamed up with IBM to usher in the rise of the personal computer. They released apps and an operating system and by licensing DOS to anyone (not just IBM) and then becoming the dominant OS they allowed clone makers to rise and thus broke the hold IBM had on the computing industry since the days the big 8 mainframe companies were referred to as “Snow White and the Seven Dwarfs.”

They were young and bold and grew fast. They were aggressive, taking on industry leaders in different segments, effectively putting CP/M out of business, taking out Lotus, VisiCalc, Novell, Netscape, `and many, many other companies.  

Windows 95 and Microsoft Office helped the personal computer become ubiquitous in homes and offices. The team knew about the technical debt they were accruing in order to grow fast. So they began work on projects that would become Windows NT and that kernel would evolve into Windows 2000, phasing out the legacy operating systems. They released Windows Server, Microsoft Exchange, Flight Simulators, maps, and seemed for a time to be poised to take over the world. They even seemed to be about to conquer the weird new smart phone world.

And then something strange happened. They entered into what we can now call a lost decade. Actually there’s nothing strange about it. This happens to nearly every company.

Innovation dropped off. Releases of Windows got buggy. The market share of their mobile operating system fell away. Apple and Android basically took the market away entirely. They let Google take the search market and after they failed to buy Yahoo! they released an uninspired Bing. The MSN subscriptions that once competed with AOL fell away. Google Docs came and was a breath of fresh air. Windows Servers started moving into cloud solutions where Box or Dropbox were replacing filers and Sharepoint became a difficult story to tell. 

They copied features from other companies. But were followers - not leaders. And the stock barely moved for a decade, while Apple more than doubled the market cap of Microsoft for a time. What exactly happened here? Some have blamed Steve Ballmer, who replaced Bill Gates who had led the company since 1975 and if we want to include Traf-O-Data - since 1972.

They grew fast and by Y2K there were memes about how rich Bill Gates was. Then a lot changed over the next decade. Windows XP was released in 2001, the same year the first Xbox was released. They launched the Windows Mobile operating system in 2003, planning to continue the whole “rule the operating system” approach. Vista comes along in 2007. Bill Gates retires in 2008. Later that year, Google launches Chrome - which would eat market share away from Microsoft over time. Windows 7 launches in 2009. Microsoft releases Bing in 2009 and Azure in 2010. The Windows phone comes in 2010 as well, and they would buy Skype for $8.5 billion dollars the next year. The tablet Microsoft Surface coming in 2012, the same year the iPad was released.

And yet, there were market forces operating to work against what Microsoft was doing. Google had come roaring out of the dot com bubble bursting and proved how money could be made with search. Yahoo! was slow to respond. As Google’s aspirations became clear by 2008, Ballmer moved to buy them for $20 billion eventually growing the bid to nearly $45 billion - a move that was thwarted but helped to take the attention of the Yahoo! team away from the idea of making money.  That was the same year Android and Chrome was released. Meanwhile, Apple released the iPhone in 2007 and were shipping the 3G in 2008, taking the mobile market by storm. By 2010, slow sales of the Windows phone were already spelling the end for Ballmer. 

Microsoft had launched Windows CE in 1996, held the smaller Handheld PC market for a time. They took over and owned the operating system market for personal computers and productivity software. They were able to seize a weakened and lumbering IBM to do so.  And yet they turned into that lumbering juggernaut of a company. All those products and all the revenues being generated, Microsoft looked unstoppable by the end of the millennium. Then they got big. Like really big. And organizations can be big and stay lean - but they weren’t. 

Leaders fought leaders, programmers fled, and the fiefdoms caused them to be slow to jump into new opportunities. Bill Gates had been an intense leader - but the Department of Justice filed an anti-trust case against Microsoft and between that and just managing hyper-growth along the way they lost a focus on customers and instead focused inward. And so by all accounts, the lost decade began in 2001. Vista was supposed to ship in 2003 but pushed all the way back to 2007. Bing was a dud, losing billions out of the gate. By 2011 Google released Chrome OS - an operating system that was basically a web browser bootstrapped on Linux and effectively what Netscape founder Marc Andreesen foreshadowed in a Time piece in the early days of the browser wars.

Kurt Eichenwald of Vanity Fair wrote an article called MICROSOFT’S LOST DECADE in 2012, looking at what led to the lost decade. He pointed out the arrogance and the products that, even though they were initially developed at Microsoft, would be launched by others first. It was Bill Gates who turned down releasing the ebook, which would evolve into the tablet. The article explained that moving timelines around pushed developing new products back in the list of priorities. The Windows and Office divisions were making so much money for the company that they had all the power to make the decisions - even when the industry was moving in another direction. 

The original employees got rich when the company went public and much of the spunk left with them. The focus shifted to pushing up the stock price. Ballmer is infamously not a product guy and he became the president of the company in 1998 and moved to CEO in 2000. But Gates stayed on in product. As we see with companies when their stock price starts to fall, the finger pointing begins. Cost cutting begins. The more talented developers can work anywhere - and so companies like Amazon, Google, and Apple were able to fill their ranks with great developers. 

When organizations in a larger organization argue, new bureaucracies get formed. Those slow things down by replacing common sense with process. That is good to a point. Like really good to a point. Measure twice, cut once. Maybe even measure three times and cut once. But software doesn’t get built by committees, it gets built by humans. The closer engineers are to humans the more empathy will go into the code. We can almost feel it when we use tools that developers don’t fully understand. And further, developers write less code when they’re in more meetings. Some are good but when there are tiers of supervisors and managers and directors and VPs and Jr and Sr of each, their need to justify their existence leads to more meetings.

The Vanity Fair piece also points out that times changed. He called the earlier employees “young hotshots from the 1980s” who by then were later career professionals and as personal computers became pervasive the way people use them changed. And a generation of people who grew up with computers now interacted with them differently. People were increasingly always online. Managers who don’t understand their users need to release control of products to those who do. 

They made the Zune 5 years after the iPod was released and had lit a fire at Apple. Less than two months later, Apple released the iPhone and the Zune was dead in the water, never eclipsing over 5 percent of the market and finally being discontinued in 2012. Ballmer had predicted that all of these Apple products would fail and in a quote from a source in the Vanity Fair article, a former manager at Microsoft said “he is hopelessly out of touch with reality or not listening to the tech staff around him”.

One aspect the article doesn’t go into is the sheer number of products Microsoft was producing. They were competing with practically every big name in technology, from Apple to Oracle to Google to Facebook to Amazon to Salesforce. They’d gobbled up so many companies to compete in so many spaces that it was hard to say what Microsoft really was - and yet the Windows and Office divisions made the lions’ share of the money. They thought they needed to own every part of the ecosystem when Apple went a different route and opened a store to entice developers to go direct to market, making more margin with no acquisition cost to build a great ecosystem. 

The Vanity Fair piece ends with a cue from the Steve Jobs biography and to sum it up, Jobs said that Microsoft ended up being run by sales people because they moved the revenue needle - just as he watched it happen with Sculley at Apple. Jobs went on to say Microsoft would continue the course as long as Ballmer was at the helm. Back when they couldn’t ship Vista they were a 60,000 person company. By 2011 when the Steve Jobs biography was published, they were at 90,000 and had just rebounded from layoffs.

By the end of 2012, the iPhone had overtaken Microsoft in sales. Steve Ballmer left as the CEO of Microsoft in 2014 and Satya Nadella replaces him. Under his leadership, half the company would be moved into research later that year. Nadella wrote a book about his experience turning things around called Hit Refresh. Just as the book Microsoft Rebooted told the story of how Ballmer was going to turn things around in 2004 - except Hit Refresh was actually a pretty good book.

And the things seemed to work. The stock price had risen a little in 2014 but since then it’s shot up six times what it was. And all of the pivots to a more cloud-oriented company and many other moves seem to have been started under Ballmer’s regime, just as the bloated company they became started under the Gates regime. Each arguably did what was needed at the time. Let’s not forget the dot com bubble burst at the beginning of the Ballmer era and he had the 2008 financial crises. There be dragons that are micro-economic forces outside anyones control. 

But Nadella ran R&D and cloud offerings. He emphasized research - which means innovation. He changed the mission statement to “empower every person and every organization on the planet to achieve more.” He laid out a few strategies, to reinvent productivity and collaboration, power those with Microsoft’s cloud platform, and expand on Windows and gaming. And all of those things have been gangbusters ever since. They bought Mojang in 2014 and so are now the makers of Minecraft. They bought LinkedIn. They finally got Skype better integrated with the company so Teams could compete more effectively with Slack.

Here’s the thing. I knew a lot of people who worked, and many who still work at Microsoft during that Lost Decade. And I think every one of them is really just top-notch. Looking at things as they’re unfolding you just see a weekly “patch Tuesday” increment. Everyone wanted to innovate - wanted to be their best self. And across every company we look at in this podcast, nearly every one has had to go through a phase of a lost few years or lost decade. The ones who don’t pull through can never turn the tide on culture and innovation. The two are linked.

A bloated company with more layers of management inspires a sense of controlling managers who stifle innovation. At face value, the micro-aggressions seem plausible, especially to those younger in their career. We hear phrases like “we need to justify or analyze the market for each expense/initiative” and that’s true or you become a Xerox PARC or Digital Research where so many innovations never get to market effectively. We hear phrases like “we’re too big to do things like that any more” and yup, people running amuck can be dangerous - turns out move fast and break things doesn’t always work out. 

We hear “that requires approval” or “I’m their bosses bosses boss” or “you need to be a team player and run this by other leaders” or “we need more process” or “we need a center of excellence for that because too many teams are doing their own thing” or “we need to have routine meetings about this” or “how does that connect to the corporate strategy” or “we’re a public company now so no” or “we don’t have the resources to think about moon shots” or “we need a new committee for that” or “who said you could do that” and all of these taken as isolated comments would be fine here or there. But the aggregate of so many micro-aggressions comes from a place of control, often stemming from fear of change or being left behind and they come at the cost of innovation. 

Charles Simonyi didn’t leave Xerox PARC and go to Microsoft to write Microsoft Word to become a cog in a wheel that’s focused on revenue and not changing the world. Microsoft simply got out-innovated due to being crushed under the weight of too many layers of management and so overly exerting control over those capable of building cool stuff. I’ve watched those who stayed be allowed speak publicly again, engage with communities, take feedback, be humble, admit mistakes, and humanize the company. It’s a privilege to get to work with them and I’ve seen results like a change to a graphAPI endpoint one night when I needed a new piece of data. 

They aren’t running amuck. They are precise, targeted, and allowed to do what needs to be done. And it’s amazing how a chief molds the way a senior leadership team acts and they mold the way directors direct and they mold the way managers manage and down the line. An aspect of culture is a mission - another is values - and another is behaviors, which make up the culture. And these days I gotta’ say I’m glad to have witnessed a turnaround like they’ve had and every time I talk to a leader or an individual contributor at Microsoft I’m glad to feel their culture coming through.

So here’s where I’d like to leave this. We can all help shape a great culture. Leaders aren’t the only ones who have an impact. We can all innovate. An innovative company isn’t one that builds a great innovative product (although that helps) but instead one who becomes an unstoppable force due to lots of small innovations at every level of the organization. Where are we allowing politics or a need for control and over-centralization stifle others? Let’s change that.