“Do things.” A Google executive has simple advice on how to pursue a career in the age of AI
Mateusz Madejski, journalist of Business Insider Polska: How many engineers does Google have in Poland?
Ryan J. Salva, Senior Director of Product at Google Cloud: Approximately 3,000 people work at Google in Poland. people in my team – about 70. As for my team, there are employees primarily focused on the development of tools such as Gemini or Antigravity. I'm probably not completely objective, but I think these are the best solutions on the market. Importantly, in Poland we create solutions not only for engineers or large companies. If someone creates personal solutions, for example a website related to their wedding, we also want to help them. The programs we create are designed to translate ordinary language – regardless of whether it is Polish, English or German – into appropriate software solutions.
The rest of the article is below the video:
So there is something to the fact that everyone will be a programmer in the future? Because anyone can create a prompt that “turns” into solutions?
I think we're already at the point where basically anyone who can speak or write can build code – using natural language. When it comes to Google solutions, you can use Antigravity, Gemini CLI or AI Studio. It's not like only people who have studied programming at the appropriate university can program… For example, I have a friend who is a lawyer. He's never written code in his life, but now he creates programs on weekends. I also have other friends who are involved in this – even a professor of philosophy is among them. He creates solutions that allow him, for example, to summarize his notes or lectures. So what I'm saying is that not everyone who creates programs today is a programmer – they are often people who have completely different, “non-technological” goals. They just want to solve their problems – and you don't need to hire professional programmers to do it.
So amateur programmers can now do everything themselves?
No, I absolutely don't want to say that. Let's go back to a professor of philosophy or a lawyer – but they create different things from programs that are created, for example, for health care systems – or for financial institutions. With projects like this, everything just has to go right – there is no margin for error. Not to mention legal or privacy restrictions. So for advanced projects, for example those for large companies or institutions, programmers and specialists are still needed who analyze the code on an ongoing basis, but also monitor the direction in which a specific project is going.
How do such specialists work?
Me and my colleagues, well, we use AI to do AI (laughter). For example, they write some code, then use AI tools to summarize it, analyze it, adapt it – and keep writing. So it's a symbiosis of writing code and using artificial intelligence tools. So AI tools are definitely democratizing access to software development.
See also: Polish companies began to limit recruitment. The reason is obvious
3 billion programmers?
But?
I have always loved music, and one of my favorite bands as a child was Pink Floyd. The band's guitarist, David Gilmour, once said that anyone can get a synthesizer, but it won't make everyone Pink Floyd (laughter). There will always be a distinction between those who code on a more amateur basis – and those who are professionals. And that's very good – creating blogs or wedding websites is different than creating complicated programs. Although “amateurs” and “professionals” will use the same programs, the latter will do it in a more advanced way.
Okay, but what about those professional programmers? Will there be much fewer of them in the future than now?
There are now around 30 million programmers around the world. How many will there be in the future? Maybe 300 million. Or maybe 3 billion? Of course, it all depends on the definition. My point is that specialists will – technically – produce code. But for example, to organize work, to deal with automation… Writing code will simply spread to other professions, which we can already see. What about these professional engineers? I think their daily work will look different than it does now.
How?
They will write code less using tools such as Java or C# (C Sharp programming language – editor's note), and will focus more on the architecture itself, on developing the entire system. Their work will therefore be taken to a higher level. Changes are already visible. Just about a year ago, I had hundreds of engineers on my teams. They were divided like this: there was a director who had about 30 engineers under him who worked on solutions, releasing solutions every 2-3 weeks. And today? We have slightly different structures. We no longer want to have 30 people working on the same solutions. We prefer to divide them into teams of 2-5 people. Overall, we employ a similar number of people, but such teams are much more effective. Instead of a new solution every 2-3 weeks, we see up to six new solutions.
How did it happen?
Before the era of dynamic AI development, each programmer was a specialist in his field – so we had specialists and an architect who looked after them and coordinated them. Now programmers can easily use AI knowledge, so in a sense each of them is an architect. It's not just about productivity – smaller groups simply work better, people are much more engaged and understand each other better than in large teams of 30 people. People are also more motivated and fulfilled because they see the effects of their work more clearly. In ten years, work in such teams will change even more. Specialists will “take care” of AI much less, they will spend less time with the machines themselves. They will think more about the final solutions and what the final product should look like – to be safe, perform well and be relevant to customers.
These will be the positions of the future
Will the requirements change?
Bright! Now, as a manager, I am interested in whether a candidate can write code and understands it. In ten years, I will be more interested in whether a candidate understands business priorities, the market environment and legal restrictions… Today's job titles are “data scientist”, “software engineer”. I think they will change. Rather, the jobs of the future will integrate today's requirements. The engineer will have to understand the market environment just like today's “product manager”.
How to prepare for this change?
This may not be anything new or surprising, but critical thinking skills will certainly be necessary – regardless of the industry. As the focus becomes more and more on business priorities, it will become more and more important to ask the questions: “Okay, but why are we doing this?” “What is the priority?” Engineers can no longer just know a programming language. They must indicate what is important, what is critically important, what the organization wants to achieve… Critical thinking helps in indicating all this. In a word – the key to future professions is not “how” to write code. The key is rather to understand “why does it arise?”
So what should an 18-year-old today do who wants to work on large IT projects in the future?
There is no one way or one recipe. I studied philosophy myself, not computer science. But I have one piece of advice, quite universal.
I'm listening.
Do things.
That is?
Create things, you can go to a programming bootcamp, you can create things in your room, you can create things at university… But the important thing is that you create. By creating things and testing them in practice, people learn at an incredible pace. When I talk to candidates, I always pay attention primarily to whether they have built something or tested their ideas in practice. It's best if they created something in cooperation with other people. Because it shows that such a candidate not only knows how to create, but also knows how to cooperate. After all, few great things come into being in isolation, developed by one person. Great things are created through collaboration – and we stick to this principle at Google.
However, the labor market is currently difficult. I understand that it's worth “creating” things, but people need a salary to survive.
It's true, the job market is difficult. I admit, I was once in a situation where I was looking for a job and it is not easy. In my opinion, the current situation is not necessarily related to AI, but rather a phase of the labor market cycle. It was caused by various political and economic turmoil – they led us to where we are. And it's a fact that nowadays you have to look for a job longer, and this is difficult for many people. But it's not like nothing is happening. We are currently looking for several dozen people in Google alone in Warsaw. Really, the deals are on the market. And a difficult cycle, like every cycle, will end someday. So I encourage you – let's create things, even if we are looking for a job. This will definitely help us. Not to mention the fact that, for example, many NGOs can benefit from our projects. It's worth thinking about.
“Further chapters of the revolution are certain”
Let me ask you a personal question. Why did you decide to enter the IT world after philosophy?
On the one hand, sure, I realized that there are not that many job offers for people with philosophy degrees (laughter). But that wasn't the most important thing. In college, I learned a lot about communication theory. I realized that forms of communication haven't really changed for thousands of years. Painting, writing, creating sculptures, telling stories… We have known all this for thousands of years. Until I realized that software development was something new – in fact, the first “new” outlet for human creativity in a very long time in human history. I felt that this was something new – and that I wanted to be part of this revolution. Well, that's how I learned to code. I must admit that I was very lucky to find employers who wanted to support me in this.
The program as a medium. I admit, it's an interesting thought.
As I mentioned, currently I mainly create programs for people who create programs. So I kind of like to think that I'm doing something that the creators of brushes used to do for the first painters (laughter).
We see an interesting trend – there are more and more robots, often based on LLMs. Will there be more and more of them? Will we work with them?
In a sense, this is already happening. My thermostat knows when I've arrived home – and adjusts the temperature. My friends have better cars than me – and their cars also set their favorite temperature for them as soon as they get behind the wheel. But we also have more and more AI agents who summarize our work or help us work. They often take notes on our behalf or simply help us perform our duties. In a sense, they already become our collaborators. In fact, current LLMs are mainly text and image based. But in the future they will probably be integrated with gyroscopes, accelerators or pressure sensors… So LLMs will become more and more “robotic”, they will be able to do more and more. Ultimately, this will only be a new input to current solutions. Further chapters of the AI revolution are therefore basically certain. Although it is difficult for me to predict when they will become a reality.
Ryan J. Salva currently works on Google Cloud solutions, but has over twenty years of experience working with development tools and software development.
See also: “Poland was first.” We're talking to Uber's head of security
