Why I Never Work Weekends
On living a balanced start-up life
It’s Saturday and I just signed-off an email to a customer by adding, “Now I’m going to write an article about how I never work on weekends.” I guess it’s more true to say that I hardly ever work on weekends.
After graduating from college in 1995, when I started working for ST Microelectronics in Bristol in the UK, some days I would work on weekdays until seven, eight, or even nine at night. That was after starting at eight or nine in the morning. So some days I was at work for up to twelve hours. Pretty much everyone else went home at five; I was the only one there, in the near darkness; the room in which six of us usually sat was lit only by my desk lamp. I was working late partly because I was scared, scared that I would not achieve my goals and scared that something terrible might happen, a reflection of the random abuse that I received in my childhood. But I was also curious and passionate about what I was working on.
Throughout my life, the amount of hours I have worked has been correlated with the passion I have felt for what I was working on. It’s hard to stay focused, to not get distracted, and to pour everything into a project unless there is some sense of autonomy, mastery, and purpose. This is expressed extremely well in the book by Daniel Pink called Drive: The Surprising Truth About What Motivates Us. So the amount of time and effort I have put into work has varied throughout my life. At times, especially near the beginning of a project—before I have built momentum—I tend to find it challenging to be engaged. As the sense of mastery increases, and as I am trusted with more autonomy, I seem to build momentum and engagement. Eventually there’s nothing more I want to do than drive the project forward.
At ST in Bristol, I remember asking a colleague why he went home every day at five. He told me, “I work to live, not live to work.” I remember resonating with the sentiment, but at the same time thinking that it was sad that he spent forty hours per week doing something that he was not necessarily passionate about. My group at ST was partnered with a Silicon Valley based 3D graphics chip company called NVIDIA. Speaking in an American accent, one of my colleagues repeated a request that came from one of the managers in the US: “can you work weekends?” The sentiment of my colleagues was simply, “heck no!” One possible factor in the resistance might have been that, unlike those requesting the weekend work, we didn’t have pre-IPO stock options in a company that was bound for the S&P 500.
But even back then, I never worked on weekends. Over the weekend, I was still usually focused on being productive in some way, often on developing my skills and abilities, but each week I specifically took a break from my employment. I spent my weekends on other things, including projects focused on learning. In one of those projects I taught myself to program in C while using it to create a 3D graphics authoring application.
In 1996, based on the quality of my work, one of those managers at NVIDIA, probably the same one who asked, “can you work weekends?” requested that ST send me over to Sunnyvale, California to work on NV3, what later became known as Riva 128, NVIDIA’s first successful 2D/3D graphics accelerator chip. I loved working as part of that early NVIDIA team, and I felt more at home in the start-up culture where we did whatever was necessary to succeed.
For about four months, in my own office (with my own door) in a building owned by Amdahl Corporation on Tiros Way in Sunnyvale, I worked intensely on verifying the functionality of Riva 128. My manager was Gopal Solanki who went on to become a Senior Vice President at NVIDIA and then to become President and CEO of Magnum Semiconductor. One time after I had told him about what I had been working on, he said to me, “I’ve realized that I like it when you do things that I hadn’t told you to do.” It seemed like he was learning about the power of autonomy.
I also worked closely with Chris Malachowsky, co-founder and VP of engineering, on making sure that parts of the graphics engine he designed functioned correctly. I was working there during the holidays, and I remember seeing Jensen Huang, co-founder and CEO, delivering bonus checks to the employees in the neighboring offices. I didn’t have the pre-IPO stock options that these people had, and ST didn’t even give me a holiday bonus, but I was still passionate and deeply engaged with the work.
One Sunday afternoon at about four, I came to the office to pick up a personal item that I had accidentally left there over the weekend. I bumped into a friend who was an employee of NVIDIA. I asked him how he was doing and he said, “I feel really guilty because I’m leaving early today.” I asked him if he had worked on Saturday as well and he confirmed that he had. I told him that any amount of work on the weekend should be celebrated.
At the end of those four months of working with NVIDIA as an employee of ST, both Chris and Jensen stopped by my office to tell me that if I ever needed a job they would be happy to employ me. I remember Chris mentioning my “work ethic” to someone around that time.
Gopal was grateful for my help, and suggested that I take some time off to visit Yosemite National Park. I didn’t have the heart to tell him that I had already spent a weekend visiting Yosemite. I actually visited a lot of beautiful places in Northern California during those weekends. I only just recently realized that, even though my office was right next to his, even though I had not been physically there, and even though I didn’t have remote access to my Sun SPARCStation computer, he must have assumed that I had somehow been working on those weekends.
Some time later, sitting with him in his office, Jensen said to me, “it’s not every day that you get to make someone a millionaire.” I became NVIDIA employee number 197 and I got my pre-IPO stock options. If I had held onto that stock, it would be worth around $50 million now, and that’s after the recent halving of the stock price.
In 1998, located in a different office in the same Amdahl Corporation building on Tiros Way in Sunnyvale, Michael Nagy (another chip project manager), asked me to design the video scaling engine on NV10, a chip that was later marketed as GeForce 256, “the world’s first GPU” (Graphics Processing Unit). I worked alone on the architecture, micro-architecture, hardware-software (programming) interface, digital logic design, verification, gate-level synthesis, timing analysis, and lab bring-up.
Michael was managing a team of something like fifty people, and I almost never saw him. One day, he called me into a meeting room, sat me down and said, “we’re all pulling our weight on this project, Duncan. We’re relying on you to do you part.” I immediately understood that he thought I was not working hard enough. I had, in fact, been working really hard, and making steady progress. Feeling that he wasn’t appreciating my effort, tears started to well up in my eyes. I could tell that he was uncomfortable with this expression of emotion. I suggested that I could send him an email every week listing the progress that I had made. He agreed, and that’s what I did. Only while thinking about this recently, in 2019, did I realize that perhaps he had noticed that I was not in my office on the weekends.
After that, probably because he had visibility into my progress, he seemed to learn to trust me more. When we were getting ready for the first chips to come back from being manufactured, he said to me, “you’re going to have to work weekends in the lab to bring up the chip when it comes back.” I remember saying, “okay,” but thinking, “we’ll see.” It turns out that I had been so thorough that the part of the chip that I was responsible for was almost completely bug-free and bring-up went smoothly and quickly. There was one hard-to-avoid bug, which was relatively easy to fix. I never did end up working on the weekends.
Soon afterwards, I had a performance review with Michael. In that meeting he revealed to me that he had discovered the details about the development of the video scaling engine in a competitor’s chip. This was a sub-module of their chip that fulfilled the same requirements as what I had developed completely on my own for our chip. He told me that their team of sixteen engineers had take a year to develop the same thing as I had in six months. So somehow, without working on weekends, I had been able to produce the same output as thirty-two engineers in a competing company.
When I finished that project, I was assigned back to Gopal to work on the next chip. After telling me about my new assignment, Michael looked at me with fatherly care and said, “you make sure you send those status updates to Gopal.” I nodded while also knowing that Gopal didn’t actually need them from me. Gopal already trusted me.
I’m not totally sure how my apparently extreme productivity was achieved. I don’t believe that I’m particularly intelligent. I’ve been in many situations where it seemed to me that those around me were much smarter than me. It takes me a lot of time and effort to understand things. I make a lot of mistakes.
However, I am conscientious and focused. I identify goals and I move them to completion while not allowing myself to prematurely claim success. I am persistent. Even though I seem to have very little belief in myself, I seem to be able to set that aside and sit with the intense discomfort of feeling completely incompetent long enough to make progress. I seem to have really good attention to detail. I’m able to discover what I don’t know and can often summon the courage to stay present with that long enough to at least partly fill in the capability-gaps.
Every skill I have that enables me to be so productive is learnable. These are not intrinsic qualities that others cannot develop. I learned all of these skills through a long and painful feedback process of continual improvement, of noticing and focusing attention on strengthening weaknesses. Some success experts claim that you should only focus on what you’re good at. I don’t agree. Mastery always requires that, to at least some degree, we identify areas for improvement and work on them. For example, if a professional tennis player always fails to return a particular type of shot, focusing on returning an easier kind of shot is not a winning strategy.
I don’t work on weekends not only because I don’t need to work on weekends, but because I’m more productive by not working on weekends. Effectiveness is the result of a balanced process where all of the needs of the human are taken care of. Back in the late 90s I was successful in-spite of how unbalanced I was. I was actually unbalanced even though I wasn’t working on weekends. I was not taking care of my body and my relationships as well as I could have. I was not spending as much time as I wanted to in traveling and exploring other interests. For example, it took burn-out, a divorce, and then a mid-life retirement for me to finally get my socializing and friendship needs met.
Now, back at NVIDIA, I still don’t work on weekends. I write every day. I exercise regularly. I hike, surf, and paddle-board. I spend time with my wife and friends. I eat lunch with a different person almost every day. I have a small garden. I research artificial general intelligence (human-like machine intelligence) outside of work. Yet I’m more effective and more productive at work than I have ever been, and I’m more able to empower and encourage others, thereby amplifying my contribution. The rhythmic cycling of intense focus with relaxation and integration are critical for maximal human performance. It’s not possible to go faster, in any sustainable way, without first slowing down and engaging deeply with the journey.