Research vs. engineering

As annoying and repulsive as the world’s richest man’s behavior often is, he is a great source of social media entertainment. He recently complained that making a difference between researchers and engineers is a false dichotomy and hence xAI would do away with such a distinction. Obviously, it is xAI’s decision on how to label their job positions. More problematic is the extrapolation from researcher vs. engineer to research vs. engineering and that there is no point in distinguishing these two.

So let me explain the difference and why it is important. First of all, we are talking about activities (research and engineering) and not about job titles (research and engineer). In engineering, you are trying to solve problems, like getting from A to B. In research (or science) you are trying to understand the underlying model, like the x/y coordinate system, so that in the future, you can get from A to B faster.

The underlying model is commonly called a theory. It is regrettable that the word theory has received a bad rap in recent decades, as if it was opposed to practice. It is quite the opposite: Good theory is validated by practice and guides it (“there is nothing more practical than a good theory”). I teach all of this in my course Nailing Your Thesis, if you are interested in more.

Now back to the original brouhaha. The xAI CEO has a point in complaining about researchers vs. engineers. Some researchers like to claim that they are different from engineers and that they should not be confused. This is often done to shed or lower responsibility for the results of the funding flowing to them.

In reality, a good researcher-engineer switches between the two activities of researching and engineering. In complex situations, this extends to whole teams switching modes, going back and forth. Sometimes it also leads to job title and position separations, but it won’t work well for long if a Chinese wall would separate these two. Still, we are talking about different activities with different skill sets to be learned at university.

When building products, you rarely just solve a specific problem instance at hand. You usually solve the problem category so that your code and configuration reliably solve the same and similar problem instances over and over again. That is to say, in research mode, you build a theory about the problem and in engineering mode, you implement a comprehensive solution based on the theory.

So the richest man in the world has a point, but he (perhaps deliberately) made this point in the most incendiary way, which then overshadowed the grain of truth within. Oh well, social media of course.