When you leave school, you face the treat promised for perhaps decades: the real world. This is the place where what you’ve learned not only can be applied but is supposed to make sense. But think about it: How much like the real world you met was the world of school? I started in 1955 in the first grade, spent six years at Merrimac Elementary School, six years at Pentucket Regional High School (including seventh and eighth grades), five years at Northeastern University, one year at MIT, two years at the University of Illinois, and five years at Stanford, and when I finished it was 1981, I had been married and was on the way to divorce; and I had a two-year-old son. I started with Eisenhower, ended with Reagan, and passed through Kennedy, Johnson, Nixon, Ford, and Carter; I started with the Korean War just ended, witnessed Vietnam, the protests, the humiliation of Nixon, the rise and fall of liberalism, the fall and rise of conservatism; we had decided to go to the moon, done it, and all but abandoned the space program; I started a young boy and ended near middle age.
When I was but a few years into this trek, they told me that I was part of the future and whatever I needed to be a productive member of society was the most important thing of all. And when I finished about to realize that all-important productivity, my son’s future became most important and I was secondary. When was my time? I’d spent 25 years preparing, only to turn to prepare the next gener- ation. No one told me that what I was preparing to do was to prepare someone else.
School is a world where failure is hard but not fatal—money can be lost, but for me it was the furtive money of scholarships and cooperative earnings, fellow- ships, soft money. People’s careers could be damaged or even ruined, but rarely did they die, did they shuffle off in shame; everything could renew next quarter, restart in a semester or next fall—you could retake a course, change advisers, transfer to another school, reapply in the winter, hire a tutor, even cheat.
So after school and after my wife drove off to New York with my son and my red ’73 Beetle, I stayed for three years in academia because it seemed so easy and familiar. Later, after an exhilarating ride starting and running my own company, I would try my hand at failure.
After I graduated from Stanford in 1981, I spent a couple of years working for John McCarthy on the Advice Taker project and for Lawrence Livermore National Laboratories (LLNL) on implementing a dialect of Lisp for a supercomputer they were building. The Advice Taker was an artificial intelligence project based on the premise that it is not possible to program a computer to figure out how to do something unless you could program it to be told how to do it.
I worked with a philosopher named Lew Creary who was a jazz guitar aficio- nado and would talk for hours about jazz greats, strange fingerings, unusual chords and voicings, and his particular brand of philosophy of knowledge, based on the work of Frege. He thought up the ideas, and I implemented them in what I thought was a simple and effective arrangement that minimized the need for a heavy intellectual investment from me. My experience getting my degree had not been very wonderful, and McCarthy’s remark after he read my thesis before he hired me was, “it’s not very good.” Nevertheless, he needed a hacker, and I was certainly more of that than he had had in a long time if ever.
I socialized with Lew a bit, and though his wasn’t a world-beating personality, he had enough quirks and texture, enough raw stuff, to make for a good compan- ion. He seemed like he was wired into the late ’40s or ’50s—he looked pretty straight and even a little worn round the edges, though he wasn’t much older than me.
Lew was always making a clever, very small joke and laughing silently while watching your face for any sign that you got it. He eventually went off to Hewlett- Packard (HP) Labs and I lost track of him after a few years.
The group at LLNL actually was a group of ex-Stanford and ex-Berkeley stu- dents out to build the world’s most outrageous multiprocessor, the S1 Mark IIA. I was hired to implement a version of Lisp called NIL (New Implementation of Lisp) which was designed by Jonl White and Guy L. Steele Jr.
The Livermore group that I worked for was a spin-off from the O Group, which was the special projects group in physics, and was led by Lowell Wood, the bad-boy physicist. Lowell did things like charter jets when he missed the sched- uled flight to the Nevada Test Site or pick jetway locks when he missed the flight to DC where he would meet with White House officials to convince them of the need for Star Wars.
The O Group was filled with tremendous young physicists who were con- vinced that all that stood between peace and war with the Soviet Union was their
brains, which were locked in battle with their corresponding numbers in bomb labs in the USSR and China.
One of the guys at the Lab had two pictures—a geographic map of the Soviet Union and a geographic map of the moon, labeled “before” and “after”. Some of them carried arms in their cars for protection against the KGB and terrorists. One carried a .357 and once was up at Lowell’s giant log house showing it off—11,000 square feet on 100 acres on top of a hill above the Lab. A group was standing around in a quarter circle when he shouted, “look, a rabbit!” He pointed the gun, apparently at the rabbit, and tracked it as it ran behind the group, and it didn’t require a lot of thought to realize that the gun pointed at one knee followed by a thigh or a waist. When the rabbit emerged from behind the group and started to head down the hill, a shot ripped out—“damn it, I missed.”
Out of the LLNL experience I would start a company in 1984 that folded in 1994—Lucid, Incorporated. This experience was unlike anything in my life before and perhaps unlike anything that could ever happen again—I raised nearly $16 million of venture capital, I signed OEM deals with IBM, DEC (now Digital), Sun, HP, and almost all the other major computer manufacturers of the 1980s, I hired dozens of top-notch developers—the company grew to 75 at its peak—I was the president of it for four years, I was the de rigueur charismatic leader of the company, I hired and fired CEOs.
There was nothing like it: I worked 60- to 90-hour weeks—because I loved it. We hired fabulous people who were the most talented I’ve seen, who loved work- ing together toward a goal we felt worthy. We had three or four rock bands, we had parties at work and away, we had an extended family which, two years after the death of Lucid, still has picnics and potlucks, and still has an active mailing list.
A company, though, is like a working gold mine: A prospector, a man with a vision, finds a vein of gold in a desolate area; a company of men smart in the ways of building sluices, constructing extraction facilities, and minding books and deliveries uses the man’s vision to create wealth, a part of which is shared with the visionary. The man who puts up the grubstake takes more gold than the finder does. It is not enough to build a mine that employs 90 men well, it also must pro- vide wealth for a handful.
Because of that, all I have to show for the work are two extra-large tee shirts with white lettering on black that says:
I founded a Silicon Valley startup in the 1980’s, and all I got was this lousy tee shirt.
When the board of directors decided to close down the company, it employed 60 people and had a yearly revenue of $5 million to $6 million. If you consider that each employee takes about $100,000 per year to support, this revenue figure
is close to breakeven, meaning that the company was not that far off from being healthy enough to continue, perhaps indefinitely. The previous quarter was a record in the primary product line, and sales for that product were growing by nearly 30% per quarter. Deals were being made, and the company was nine months away from AT&T adopting one of its products as its official product. The primary product had just won the Jolt Cola Programmer Productivity Award from Computer Language, a computer magazine, and things, except the bottom line, were looking good.
One day in late June 1994, at 4:00 p.m., Bill Goerke, the interim CEO, called all the employees onto the second-story porch where most of the company meetings were held. It was a few days before payday, and he told them,“don’t bother coming to work tomorrow, we can’t pay you.”
Robert J. Hoder, the designated representative of the Credit Manager’s Associ- ation of California, the official liquidator, stood at the front of the Menlo Park council chambers and told the gathered creditors: “In the 20 years I’ve been doing this job, I have never seen a meeting where there was even a single employee. Every other company was savvy enough to know two weeks in advance of when it was about to run out of money and would let the employees go after their last paycheck.”
In 1985 when Brentwood Associates was wooing Lucid, Brad Jones pulled his BMW 700 series over to the side of the road and told Tony Slocum—Lucid’s CEO—and me that Brentwood was the kind of firm that stuck with its compa- nies. “Once we had to shut down a company, and Mr. Warren personally made sure that each employee got a full month’s pay. We will take care of you through success or failure.”
The company kept a couple of people to answer the phone in case customers called support and to ship product if that was required and to collect receivables. But to customers who had recently bought one of the company’s products and to the employees, the company simply disappeared for no reason at all.
I think this is a story of business incompetence built on top of technical excel- lence—I don’t assume any credit for that excellence, except that I happened to have hired extremely talented people. Perhaps there are lessons to be learned for the ’90s from this story.
I forgot to mention, when the company was about nine months away from folding, the then-current CEO read Crossing the Chasm by Geoffrey Moore (1991)
and learned one thing: When a company is trying to move beyond the early adopters to mainstream customers, sometimes the technical founder must go, so I went. After Lucid’s demise I read editorials about how the company could never grow out from my shadow (Binstock 1994) and how the company was “blinded by science” (Deger 1995).
Since then, as VP of Development at ParcPlace Systems, Inc., I learned that the general belief, at least among executives and venture capitalists (VCs) is that Lucid failed because I did. Perhaps I did fail, and it certainly felt like it. Blame is not something worth ascribing in most cases, and in this one blame can get passed around like a flu. I learned, though, a lot about business practices in the United States, and it seems to me that the “blame” for Lucid’s failure lies squarely at the feet of American business education, exemplified by Lucid’s next-to-last CEO: Bob Kadarauch.
The story started when some of the guys I worked with at LLNL had the idea that we could start a company by spinning out the technology we had developed at LLNL for the S1 project. The original project was to construct a CISC (Complex
Instruction Set Computer) multiprocessor, and the uniprocessor for it was suc- cessfully constructed by a small team in a short time (the S1 Mark IIA). However, the lesson of CISC was quickly learned: It is hard to write a compiler that takes advantage of the complex instructions and addressing modes, because that requires the compiler to recognize that these complex operations apply while looking at a number of program statements, and it is hard to make devices to go fast when you have to fetch and execute a lot of simple operations along with the complex ones. Hence, RISC (Reduced Instruction Set Computer) was revalidated and the next version of the S1—the S1 Mark IIB—was a RISC processor; this was the one we wanted to spin out.
On top of this was laser pantography. The idea was that the chemical process that deposited semiconductor subparts on a silicon wafer could be activated by a laser. In effect, rather than a photographic process we could use a drawing pro- cess. With laser pantography we could build and test a wafer-sized chip in a cham- ber and, to some extent, repair it. The idea was to combine the RISC design, this technology advantage, and our superior talent—cough, cough—into a super- computer company.
At the time I was consulting occasionally for a venture capital group in New York headed by Bill Scherlis’s Uncle Sid—remember, Bill was the guy who rented a room from me the first year at Stanford. Sid and his assistant Victoria would send me a business plan for a company in the Bay Area, and I would read it, visit the group, and assess the technical likelihood for its success. Of course, I knew nothing about business, so I could really assess only the technology.
For example, the first company I looked at had developed a simple ethernet local area network (LAN) for fast food joints. Each cash register would be a node on the network, which would download its transactions to a backroom com- puter every few minutes, thereby making it less likely for one of the poorly paid
hamburger salespeople to rip off the franchise. In fact, the design of the register itself made it almost impossible to steal from—the keyboard was bulletproof!
This sounded good to me, but I was concerned that their very slow communi- cations software and hardware coupled by the slow backroom computer would bog down during peak periods like lunchtime in large food emporia. I dutifully and diligently visited potential customers (franchise managers) and asked about peak traffic and did calculations based on the technical specs of the proposed products. I also quizzed the hackers about what happened when there were too many collisions on the ethernet and consulted my networking friends at Stanford.
I reported all this on a conference call. Sid and Vic listened carefully and asked probing questions. Then Victoria said, “Well, this is all very interesting, but it is clear these people have no business sense.” This didn’t seem clear to me because the business plan was thorough, they had researched their market and inter- viewed many potential customers. She continued, “Everyone knows that the way kids working in hamburger joints make money—given they’re paid the same or less than waiters but without tips—is by stealing from the cash register, and man- agement plans for the amount, which they expect to be around $20 per day per employee.”
Gleep, no, Victoria, not everyone knows that. How foolish these people were to design a theftproof cash register and to log transactions almost as soon as they were made. No venture capital for these naïve souls.
Armed with this newly found business acumen, I called up Victoria and talked about our plan. She came out to visit, and we explored the idea. I wrote a series of business plans for the Livermore group. It turned out that when faced with the potential of getting a few million bucks to get started and to commit to actually trying the venture outside the comforting chain-link fences at the Lab, away from the full weight and authority of not only the University of California but also the U.S. government as well, my colleagues declined.
But along the way, constructing the business plan, I had made the case that the rise of artificial intelligence (AI) in the marketplace, coupled with the need for commercial implementations of Common Lisp—the de facto Lisp standard I had initiated and politically led—on standard hardware was significant enough that it was a focus of the full-spectrum computer company we proposed. Victoria sug- gested that that part of the business plan was interesting to her and Sid.
I was surprised, but I looked around for additional folks for the venture and lined up 10 founders, two of which would be silent. One was Bill Scherlis and the other Scott Fahlman, who led the Common Lisp group—during the definition of the standard I did the backroom politicking with research and government groups.
Sid and Vic coached me on how to be a founder and what would happen. One of the disturbing things was that Victoria said that regardless of how things went,
one of the key founders would leave the company in a bad way within a year or two—forced out, shamed into quitting, or in a huff over a disagreement. I looked at the founders and only one wasn’t a friend, and I wondered who it would be who would leave in disgrace.
Tony Slocum would be the CEO. He had been the CEO at Intellicorp who took