Cupertino, December 2001. A manager named Joe walks into a quiet office and asks the question that smells of the guillotine: “Show me what you spend your days on, because right now I can’t justify your salary.” Across from him, an engineer, John Kullmann, says nothing. He switches on an ordinary Intel PC, one of those grey boxes you would have found in any accounting department at the time. On the screen appears “Welcome to Macintosh.” Joe is speechless (Engadget).
That grey box had just saved Apple’s laptop business, four years before anyone outside the company heard about it. And it started as the tolerated hobby of one man nobody had asked for anything.
A five-year double life
In June 2005, on the WWDC stage, Steve Jobs announced what no one expected: Apple was leaving the PowerPC chips from IBM and Motorola to move to Intel. Then he dropped the line that stuck. Mac OS X had been living “a secret double life” for five years. Every public release of the system, compiled for PowerPC, had also been quietly compiled for Intel. To prove it, that day’s demo was already running on a 3.6 GHz Pentium 4 (Macworld).
Five years of parallel development, kept secret, ready to fire the day the official strategy collapsed. That is what I call a real plan B. Not a slide in a steering committee, not an intention noted in some minutes. Code that boots, tested at every release.
Ask yourself, honestly: does your organization have a plan B that boots, or only a plan B on PowerPoint?
The invariant was not the chip, it was a human
The secret project, later named Marklar (after the parody alien race from South Park, a name that appeared only in October 2002), was not born from a leadership decision (MacRumors). It was born from the curiosity of one engineer.
Around 2000, John Kullmann has a rare permission to work from home, partly to care for his young child. At home, he decides to see whether Mac OS X can run on standard Intel hardware. For about eighteen months, it stays a one-man project, tolerated rather than encouraged, known to six people at most inside the company (The Register). Then comes that December 2001 day when his manager asks him to justify his time, and the grey PC shows the Mac startup screen.
I have said it for years: the invariant of innovation was never technology, it is the human being, with their fears, their biases and their curiosity (Book), Chapter 11, the organizations and structures pillar. NeXTSTEP, the system Apple bought in 1996 to build Mac OS X, was designed to be processor-independent and already ran on Motorola, Intel, SPARC and PA-RISC. The technical capability existed. But a technical capability does nothing until a human being decides to grab it one evening, at home, out of sheer desire to know.
Ambidexterity: hold the present and prepare the future
Here is the point that makes most executives uncomfortable. During those five years, Apple was publicly selling PowerPC, promoting it, committing its customers and developers to that architecture. And at the same time, a small team kept the emergency exit toward Intel in working order.
I call this capability ambidexterity: holding the profitable short term and the long-term vision under the same roof, without one smothering the other. Most organizations cannot do it. They defend their present success with such energy that they forbid, sometimes silently, any work on the scenario that will replace it. Apple did the opposite. It kept milking the PowerPC cow while feeding, out of sight, the Intel calf.
In January 2002, two more engineers join Kullmann, and the team is moved off campus to limit leaks (The Register). The project stays tiny, discreet, underfunded relative to the stakes. But it lives, it updates, it boots at every build. That is the definition of an ambidextrous organization: it does not ask permission to cover its back.
The G5 that ran too hot for a laptop
Why did this plan B become vital? Because of heat, quite simply.
Apple’s flagship processor, the PowerPC G5 made by IBM, delivered on its power promises at the cost of a thermal output that was hard to contain. The fastest desktop models, at 2.5 and 2.7 GHz, had to move to liquid cooling to work inside a tower (512 Pixels). Steve Jobs had publicly promised a 3 GHz G5 within a year. IBM never got there (512 Pixels).
A processor that needs a massive aluminum heat sink, then a liquid circuit, does not fit inside a laptop. So Apple found itself with a laptop line doomed to stall, precisely when the laptop was becoming the heart of the market. Without a credible alternative, Apple’s laptop business was stalling. The secret plan B became plan A.
Remember my conviction: a company that doesn’t innovate is already dead, it just doesn’t know it yet. Apple could have died slowly against that thermal wall. It had prepared the exit before it even knew it would need it.
Kill the assumption before it kills you
What strikes me most in this story is the experimental discipline.
The comfortable assumption, in 2000, said: “PowerPC is our future, IBM and Motorola will keep the pace.” A seductive assumption, carried by past success, exactly the kind of belief that blinds a leadership team. Assumptions are highly dangerous and should be killed with experimentation as early as possible. Apple, by design or by luck, kept a living experiment measuring, continuously, whether its official bet still held.
Every build compiled for Intel was a full-scale test answering one verifiable question: “If PowerPC lets us down tomorrow, are we ready?” As long as the answer stayed “yes,” Apple could defend PowerPC without naivety. The day the answer became “we have to,” the switch took a few months instead of several years.
Compare the communication of a leader who says: “Our past success speaks for us, we cannot be wrong.” Then this one: “We ran an alternative continuously, here are the results, we can switch with confidence.” Which one moves you (Book), Chapter 8, the vision, mission and strategy pillar?
What your organization should take from this
When Apple announced the transition in 2005, the hard part was already done. The system had been running on Intel for years. All that remained was to give developers the tools to create Universal Binaries, applications that hold both PowerPC and Intel code in a single file. The result: the first Intel Macs arrived ahead of schedule (Apple), instead of a chaotic multi-year rebuild.
This success does not come down to the genius of one man, even if Kullmann earns his place in history. It comes down to an organization that let a clandestine experiment survive, then knew how to industrialize it at the right moment. Past success, the weight of committed relationships and assumptions become our most formidable enemies when the organization is not ready to experiment (Book), Chapter 11, the organizations and structures pillar.
The good news is that you do not need Apple’s cash to reproduce the principle. You need something rarer: the psychological safety that lets a Kullmann tinker with his plan B without getting fired, and the clarity of a management team that, when the day comes, says “yes, let’s go” instead of “we have always done it this way” (MacDailyNews).
The essentials
Remember three things:
- A real plan B is code that boots, not a slide. Apple compiled Mac OS X for Intel in secret for five years, ready to switch.
- The invariant is the human. The project was born from the tolerated curiosity of one engineer, on standard hardware, at home.
- Ambidexterity saves companies. Defending the PowerPC present and feeding the Intel future at the same time turned a thermal crisis into a controlled transition.
You do not have to predict your technological future. You have to experiment with it continuously, so you are ready the day your own PowerPC starts to overheat. Innovate or agonize, the choice is up to you.
References
- (Book) https://philippeboulanger.com/book/
- (MacRumors) https://www.macrumors.com/2012/06/10/a-bit-of-history-behind-the-mac-os-x-on-intel-project-marklar/
- (Engadget) https://www.engadget.com/2012-06-10-how-marklar-os-x-on-intel-owes-its-start-to-a-one-year-old-boy.html
- (The Register) https://www.theregister.com/software/2012/06/11/insider-cuts-into-apple-peels-off-intel-mac-os-x-port-secrets/1131444
- (Macworld) https://www.macworld.com/article/175877/liveupdate-8.html
- (512 Pixels) https://512pixels.net/2024/12/2005-the-end-of-the-macs-powerpc-era/
