The Autonomy Software Engineers Need for Career Growth

After Kimi recently released its K3 model, it sparked a lot of discussion in the community because it outperformed Claude Fable in some benchmarks. Kimi's founder and CEO, Zhilin Yang, earned his PhD from Carnegie Mellon University (CMU). His former advisor, Russ Salakhutdinov, posted on social media that many people had asked whether Kimi's founder was one of his PhD students.

His answer was, "The answer is yes, and he is indeed brilliant." In the same post, he listed a number of PhD students he had advised in the past, including founding members of Thinking Machines, a co-founder of xAI, and several people who are now professors at MIT, Princeton, and other universities.

Google Distinguished Scientist Peyman Milanfar then reposted it with a comment: "This list of talent is certainly impressive, but more than 80% of the researchers Russ is proud of were advised while he was also handling academic work and holding major executive roles at Apple or Meta. In many cases, the students largely figured out their research on their own."

He specifically noted that he was not trying to take a shot at Russ. Top PhD programs naturally attract excellent talent, and those students' abilities support the entire lab. So when evaluating outcomes, we should be careful about how much credit an advisor should receive, and we should not underestimate the students' own talent and discipline.

In response, Professor Russ Salakhutdinov said, "Whether a student succeeds of course depends on more than the advisor. It is also closely tied to the environment they grow in. CMU has a very special atmosphere. It values collaboration deeply and brings together an astonishing amount of talent."

That part really resonated with me. Looking back on my own education, I genuinely feel that environment matters most. There is a reason the story of Mencius' mother moving three times became a classic. In junior high, I was a fairly conventional student. I studied whatever the teacher taught. But because the coursework in junior high was not that difficult overall, that approach was still enough for me to get into Taichung First Senior High School.

The real shock came after I entered that school, from the classmate sitting next to me. Whether it was math, physics, or chemistry, while the teacher was explaining the lesson, he was not listening at all. He had already started working through the problems in the textbook on his own, often moving faster than the class. That was when I realized that in learning, this kind of autonomy is possible, and maybe even necessary.

Later, after entering National Taiwan University, it no longer felt surprising to see many classmates taking graduate-level courses in their first or second year. For the very best students, learning the material was not a problem even without a professor teaching it to them directly.

High school was more than ten years ago now. The internet already existed then, but online learning resources were nowhere near as rich as they are today, and there was certainly no ChatGPT available at any moment. Compared with my younger self, students today are in a much better position to use the internet to teach themselves almost any foundational subject.

At the more advanced graduate-school stage, the ideal advisor is someone who can provide industry resources, let students freely explore research directions they care about, and offer guidance when needed. Even outside the AI era, if someone still needs an advisor to watch them closely the whole way, teaching them like a child, then graduate school is probably not the right fit.

Coming back to my own work as a software engineer, I think career growth in software engineering should follow the same attitude and method. Once you enter the workplace, no one will push you from behind the way teachers might have in elementary or secondary school. If you want to accelerate your career growth, actively learning new technologies and intentionally developing soft skills are both essential.

If you are thinking about whether to pursue software engineering as a career, it may be worth asking yourself first: do you actively try new technologies? Even when no one is watching, do you use evenings and weekends to keep improving yourself? If the answer is no, software engineering may not be the right path for you.

Support ExplainThis

If you found this content valuable, please consider supporting our work with a one-time donation of whatever amount feels right to you through this Buy Me a Coffee page.

Creating in-depth technical content takes significant time. Your support helps us continue producing high-quality educational content accessible to everyone.