Urs Hölzle: «Computer science is actually not a narrow career choice, as it leaves plenty of options open to you.»

Mr. Hölzle, you began your studies in 1983. Why did you opt for ETH back then?

I wanted to study computer science and there was practically only one place to do so at the time: ETH in Zurich. Seen like that, my choice was very simple.

What did you learn from your time at ETH?

The department was no more than two years old when I arrived. So the curriculum was still very broad in those days, certainly a bit broader than what’s offered today. Back then, we did more math than computer science.

What I really liked was the environment. I could attend lectures and seminars that made me think. They weren’t just about computer science, but included scientific topics of general interest. And of course what I also really liked was that ETH is an international university. I had a lot of contact with people from other cultures and countries.

What made you want to study computer science?

The subject really interested me. Also, I was aware that I would be working for something like 40 years after completing my education. Computer science gave me huge flexibility in this regard. Although I would be working in IT for many years, I could choose the industry or specialist field of work. Computer scientists are needed everywhere – in medicine, biology, and the energy industry. In this sense, computer science is actually not a narrow career choice, as it leaves plenty of options open to you.

Did you know from the outset that you wanted to do your doctorate abroad?

No, not at all. I didn’t even know if I wanted to do a doctorate. It just worked out like that. I took things as they came, year by year. I really started thinking about a doctorate only in the last year of my ETH course. In the end, I wanted to go to the US because the subject that I was interested in was not covered in Switzerland or anywhere in Europe. Back then, I spent a lot of time on object-oriented programming languages. Almost all the researchers working in this field were based in the US. I chose Stanford University on account of my later doctoral supervisor. After I read his publications, I knew that I wanted to apply to him.

So doing a doctorate at ETH was never an option for you?

No. I didn’t want to do a doctorate just for the sake of the title, but to further my skills in a specific field that fascinated me. From that perspective, I chose the people with whom I wanted to collaborate rather than the place.

Did your ETH degree prepare you sufficiently for Stanford University?

Absolutely; even back then ETH was one of the best institutions in the world. I passed the entrance exam to Stanford University with very little effort.

But after that it was probably like entering a completely different world, wasn’t it?

Yes, the difference was really enormous, especially in terms of operations. Even in those days, Stanford was less of a school and much more of a meeting place for like-minded people. The university maintained close contact with industry already early on, and it was not unusual for professors to be actively involved in start-ups.

The equipment was much better at ETH; even the stuff that was available to students. I can still remember that we were allowed to use Apple Macintosh computers and Sun workstations at ETH. At Stanford, the terminals were 24×80 pixels, without a graphical user interface. Doctoral candidates were required to join a research group. That’s where the money was, and the equipment was better. The decision of which research group you wanted to join was something you had to figure out by yourself. It took me nearly two years to find the subject area I wanted to do my doctorate in.

Is there anything that was not taught sufficiently, or at all, in your ETH degree?

The greatest difference back then was that the American universities prepared the students much more in the art of communicating their ideas. No matter whether you were giving a talk or writing a paper, the expectations in terms of communication and presentation of the content you created were enormously high and were then also practiced accordingly. That was certainly not considered so important at ETH in my day. I personally view communication as a key success factor – not only in the academic world but also in working life. I have the feeling that communication is learned at school in the US, even before going to university, and in Switzerland you’re more likely to learn it on the job. This has no doubt changed greatly at ETH over the last 25 years. Nevertheless, the presentations given by Americans at international meetings are still some of the best.

Can you explain more specifically why you think communication is so important?

Because IT is teamwork. These days, there are only a very few areas of IT where an individual can make a big impact, such as in cryptology or compression techniques. What matters most there is the algorithm. This is where an individual can make ground-breaking progress. But normally it’s teams of at least two or three engineers working together closely who achieve something jointly. That’s why it’s crucial for you to be capable of communicating your ideas to other people and of talking about the work you’re doing. If you know what would be the right thing to do, but can’t convey that to the others in the team or the company, the right thing will never happen. So the way I see it, people who can’t express themselves in a clear way can never be successful.

In other words, if someone wants to have a career at Google, then in addition to having the right technical skills, they must also be able to communicate well?

Definitely. Like other companies, we at Google distinguish between a technical track and a management track in a career path. This means you have two ways of developing professionally at Google: via leadership or by using in-depth technical knowledge. In both areas, the ability to communicate is crucial for success. At first glance, this seems more understandable applied to managers rather than tech leads. But if you’re the person responsible for technology decisions, you’re the one who needs to make sure that everybody in the organization understands what you’ve decided and why. It is fundamentally important that everyone is aware of the company’s technical orientation and understands the reasons for it.

Is the situation similar when it comes to technical knowledge? Do managers at Google need to understand technology?

Yes, I think it’s one of Google’s strengths that we also expect managers to have a good understanding of technology. Even if managers don’t take the technical decisions, it is important for managers and tech leads to cooperate closely. To do so, a manager must understand both the technology and the facts and figures that underpin the decisions. In the issues we deal with, there is often no simple yes or no, and no simple A or B. If the tech lead finally goes with a judgment call, the manager also needs to understand the big picture, and this includes the technological aspects alongside the economic and organizational considerations. This explains why people normally work their way up to tech lead with us first, even if they want to take on management roles later. You might say they develop as engineers and then switch to management.

The technologization of the world is advancing in great strides. Is being an engineer beneficial in business?

I certainly think so, because engineers can generally better assess and unlock the possibilities that the internet brings to light. I believe that the internet is the real revolution. An engineer who uses the internet for his benefit can work anywhere, meaning in any industry and physically at any location.

Could you explain that in more detail?

It is the internet that has brought about many more changes in the world, rather than the world becoming more technical. The world was already technical 20 years ago. Electricity, the telephone and the car were all innovative technologies, developed by outstanding engineers. Unlike today, it would not have been possible for those engineers to have reached a global audience – and hence a global market – so quickly with a good idea. Software and the internet have simply made this revolution possible.

It is of course still incredibly difficult to have a good idea and then to implement it technically. Thanks to software and the internet, these days engineers can share their idea globally at relatively little expense. You can develop things in small teams and scale up quickly if you’re successful. You don’t need much starting capital to invest in production or in a distribution network.

In the past, if you came up with something, you couldn’t simply put it on the shelves of a store and sell it. Anyone can market and sell anything on the internet. Thanks to the internet, global access to ideas has become much easier. And because you can also win recognition with ideas, they are far more important.

What would you say to a school-leaver who is currently planning a career?

I would tell them I don’t rate career planning. Looking back, I was way off the mark with all the predictions of what I would be doing for the next two years when I was studying and first started working. And I have to say, I actually think that’s pretty good. You shouldn’t worry too much about the future. I think it’s more important to concentrate on whatever it is you’re doing right now. If you do what you’re doing well, the next opportunity will open up of its own accord.

So you’re saying that no plan is the better plan?

Of course I realized that it wouldn’t hurt to go to Stanford to do my doctorate there. But if you’d have asked me that question back then, I would have been sure that I would return to Switzerland after four years to do something else. But that isn’t the way it panned out. My dissertation at Stanford, my time at the University of California, my first start-up, the time at Sun Microsystems and the encounter with Larry and Sergey. All of that wasn’t planned; that’s just how it turned out.

Did you have the self-confidence partly because you knew that you’d had the best possible start at ETH?

If you learn a trade that is in demand and the market, and you’ve been educated at such a prestigious institution, you don’t really have anything to worry about. You can’t really afford to get anxious if things don’t work out at a certain job. If you’re working at a place you don’t like, there’s another position out there waiting for you. What we have in Silicon Valley is unique of course. If you make a mistake, it’s no big deal. The important thing is not to make the same mistake twice. Here in Silicon Valley, you’re successful if you keep making new mistakes instead of repeating the same mistakes.

Of course that type of culture makes it considerably easier to stay relaxed …

That’s true. At the same time, though, I think that as an engineer, and especially as a computer scientist, you have the best prerequisites for professional success. The knowledge and skills of good engineers will always be in demand. You’ve just got to be open and have the courage to accept change in life as a challenge and to seize opportunities that open up for you along the way. The world will continue to change rapidly in the future. You just have to know how to deal with it.