A Detour at Intel

1983-1984
Greg Bryant

I played a tiny part during a key moment that arbitrarily and slightly shifted the direction of a vast movement, a torrent that transformed the world for the worse.

And I observed firsthand the destructive ignorance of 'our great captains of industry'.

I was hired to work at the headquarters of "Intel". How to describe this place? A high-tech, anti-union pillar of the United States' industrial hegemony? A blind, anti-human, profit-driven, immoral tyranny? I can't say enough bad things about the very existence of this category of institution, which is built on the belief that the accumulation of power is the primary goal, and that the best way to accumulate power is through intense, wasteful competition, which the rest of us just need to cope with somehow -- typically through hidden cooperation.

Anyway, I was recruited as an experienced UNIX guru with microprocessor expertise and many ideas.

I was tasked with taking the UNIX hacking load off of a young Pat Gelsinger -- an increasingly important problem-solver who needed time to get a microprocessor engineering degree. My hiring was approved by a sharp, and ultimately successful, insurgent group at Intel's global headquarters in Santa Clara. Like many insurgents, they had supporters among the oppressed at Intel, some secret comrades up the management ladder, but, ultimately, they had full establishment support when they were proving right. We were absolutely certain that microprocessor engineering needed to be treated as software engineering ... the 80386 or i386 in fact could not have been built without UNIX and its software tools.

(This is indirectly self-referential, because Gelsinger's book on the 80386 was the primary inspiration for the second free OS based on unix: linux. Here's an overhead presentation by Pat Gelsinger in early 1984 on UNIX, or the version we were licensing: Amdahl's UTS, which I maintained on a single VM within an IBM 3081. The presentation summarizes the material I and others collaborated on, with Gelsinger, in the preceding months).

This insurgency, and other tempests in a teapot, took place mostly on the top floor of a small, ugly, two-story building. The battles had world-changing consequences only because of the vast monies and energies focused on this particular engine, Intel, within the computer industry. The executives who founded Intel were not influential because they were brilliant. They were mostly clever, but that's beside the point. They were successful because they'd been seduced by 'the drive to win', and took advantage of opportunities to grow their portfolios of technology, power, and finance. They built armies to serve the establishment. They happily helped to grow industry and empire, no matter the cost to the world. They sucked up the brightest people and nature's precious resources for power and profit, justified by convenient collective delusions of human progress. They weren't the only ones deluded in this way, of course. The 20th and 21st centuries have been full of unleashed, uninhibited tech daddies, a fate I managed to turn away from, but whose temptations I felt quite clearly. If one looked at life in a totally selfish light, becoming a tech plutocrat seems like the only liberation available, in a world of alienation produced by extreme capitalism which, at that point, had already been extreme for a hundred years. Today, everyone is becoming aware that capitalist industrial growth has burned the planet, and impoverished nearly everyone.

Gordon Moore, famous for observing the exact pace of this destruction, was the CEO, and he would stalk the floor restlessly. He'd regularly pull a mildly displeased face at me, for purposefully not wearing my badge ... but he would not tell me to wear it. At that point, he wasn't completely sure if badges were a sufficiently important issue to lose a good engineer over. And, after all, everyone knew me -- why would I need a badge? He once sat down next to me in the cafeteria, and started chatting about someone who had left intel recently, and how easily people moved from company to company, and how that was his history too, and how that was simply the way of 'the valley' (this was my first job south of San Francisco) -- fueled by a kind of rebellion, later to be known as 'disruption'. I think he was trying to tacitly acknowledge the value of my rebellion against the badges, but gently encourage me to wear one, without saying so explicitly. After I left intel, this kind of tolerance would completely disappear.

Everyone at Intel was working unecessarily hard, and typically on busy-work, despite the popular "work smarter, not harder" slogan. I found myself rebelling against this obvious dissonance, and would take ever-longer lunch breaks offsite. The employees suffocated in an impossibly bureaucratic, autocratic, and sterile office-political atmosphere. It was managerial, hierarchical, irrational, primitive, and dull. Not only were unions forbidden: co-operation, even for the good of the company, needed to be covert.

There was a patronizing, soul-sucking internal marketing campaign to teach everyone the so-called 'Intel culture', which seemed to mean that creativity, innovation, opportunity, mutual aid, social responsibility, and quality-of-life were stifled at every turn.

Some of my immediate co-workers agreed with that assessment: the secret rebel group. But the engineering challenges were motivating, and consumed our attention. We enriched this new approach to producing microprocessors, while stitching together computing power to do the work. Again, the 80386 wouldn't have been buildable without unix -- itself a covert project within AT&T, a different tyrannical corporation -- as well as countless notions imported into electronic engineering from the world of software. The work we decided to take on in this airless environment (smoking was allowed, so this is a visceral memory for me) required tireless, diligent reasoning, systems-level rationality, and relentless bug-killing. Here's a typical email from one of the rebels: a 22-year-old Pat Gelsinger. I served as a technical consultant for a corporate-wide migration of engineering to unix, pushed by the rebels. I offered ideas, made observations, and debated, alongside another 'intel culture' affront: a nauseatingly unproductive and unsubtle divide-and-conquer mechanism promulgated by the executive management, known as 'constructive confrontation'. Which was only confrontation, and never constructive. Still, it wasn't hard for us to win arguments, because we were borrowing successful software principles and deploying them for the first time to complex chip design. All of that is laudable, but most of my time at Intel was spent working with Gelsinger and others to cobble together a custom heterogenous network of real and virtual unix machines to improve our computing capacity and performance, to move the 386 or 'p3' project forward:

Although there were certainly brilliant people at Intel, the insanity of the organization itself drove me crazy by osmosis. I resigned in the middle of an additional storm of stupidity, this time from the outside, when IBM decided that 32-bit computers (including the 386) were overpowered for desktop machines ... and so they wouldn't be using the chip at all! Many inside and outside the company also saw no merit in the microprocessor group's hard work on facilitating backwards-compatibility for software, which we forcefully advocated, arguments I'd learned in part from portability headaches that arose in the unix world (I was a C portability consultant prior to this job, so I longed for compatible hardware advancements). These simple ideas were barely on the radar of the establishment executives.

After over a year of these useless battles in the world's dullest environment, I gave my notice. Desperate, Pat Gelsinger and three senior rebels took me out to lunch.

In the tradition of Intel's 'renegades' -- who left one startup to start another, and left that one to start Intel -- they begged me to start a super-rational company, building 32-bit machines with the 386 after it was released, machines which could run all known software in virtual mode -- and then hire the 386 engineers away from Intel! They were appealing to various discussions we had during my time there, including the need to jettison corporate propaganda, and deprecate technologies that survived only because they were proprietary. (The free software of the future was also part of this discussion. Microsoft's importance was not yet secured, since its success was dependent on ties to IBM, and the nepotism that landed Bill Gates, and his monopoly criminality, into that position.)

But, deep down, they just thought I'd be a fun CEO, because I don't believe that anyone should indulge the fantasy that they can 'manage people'. Instead, people must work together to manage the project, and if possible, decide what the project is. Because that's what happens in successful projects anyway. Management is deluded if they think otherwise. It's better to clear away the chains of autocracy and get on with cooperative action.

But, I asked them: Why help Intel out of its stupor? Why take advantage of an obtuse IBM? Why promulgate computing at all? It's a dirty business, making chips and machines, despite its PR-cultivated post-industrial image. We were all fighting within the industry, but we weren't fighting to stop its problems! Is that doing the right thing? We fought out of frustration, and a sense of injustice, because we were at the bottom, and the people at the bottom know better than the people at the top. We were drawn to the technical problem of forcing 'business people', and even former engineers like our executives, to understand reality. But was that worthwhile? It depends on the situation, of course, and we need to force open these internal discussions. But I was so numb from Intel that I didn't feel anything in computing deserved that commitment and activism.

I also told them that, based on the examples I'd seen, I would become a worse person if I became a CEO.

So I said no.

So, they all stayed at Intel, and fought. Compaq decided to seize the opportunity -- or risk -- on the 386. The rebel project I'd helped for over a year ultimately became the most influential computer chip in history: the 80386, or i386, the microprocessor inspired by unix software practices.

And, after several exits and returns, Pat Gelsinger is CEO of Intel now.