Musk: There are no more researchers, only engineers. LeCun: That's completely wrong.

For a long time, the roles of scientists (researchers) and engineers have been clearly defined.

This distinction exists not only in the academic community but is also deeply rooted in popular culture. For example, in the American TV series The Big Bang Theory, the physicist Sheldon Cooper often regards himself as a "real scientist" and sneers at Howard Wolowitz, an engineer. The occupational differences between the two even become an important source of material for the comedy.

If we use an informal but easy - to - understand statement to summarize, scientists are committed to discovering the laws of nature and understanding "why the world is the way it is"; while engineers are more concerned about "what we can do with this knowledge." They hope to transform existing scientific principles into real - world technologies, tools, and systems. One pursues truth, and the other pursues feasibility.

However, Elon Musk, the world's richest man, has recently challenged this deeply ingrained concept.

In the comments of reposting a recruitment tweet from an xAI employee, Musk claimed that this wrong naming that distinguishes between "researchers" and "engineers" is actually an implicit description of a two - tier engineering system.

He announced that xAI will no longer distinguish between them from today: "There are only engineers here."

He also said: "The word 'Researcher' is an antiquated term in the academic world."

Below this tweet, Musk continued, somewhat sarcastically, "SpaceX has done more meaningful and cutting - edge'research' in rocket and satellite development than the sum of all university academic laboratories on Earth. But we don't use the pretentious and irresponsible term'researcher'. (We use) engineers."

It's imaginable how much controversy such remarks will arouse. Apart from fence - sitters, onlookers, and meme - makers, there are many supporters and opponents of the "non - existence of researchers" theory.

Among the supporters, engineer David Brown believes that the division between researchers/engineers is just to exclude those without a doctoral degree.

Anaïs Howland, the founder of ParadigmShiftAI, said that in large laboratories, this division between researchers/engineers is just used to divide status.

Jordan Ford, a robotics entrepreneur and a Ph.D. in robotics from CMU, also said that he prefers the title of engineer.

The views of the opponents are also very clear. Some people try to patiently explain the concepts:

Some people question it by citing the great scientist Newton as a counter - example.

Some people also sarcastically retort, "The word 'Engineer' is an antiquated term since the appearance of the engine."

Some people joke that Musk made such remarks to prevent Mark Zuckerberg from poaching xAI's researchers.

Among the opponents, Yann LeCun, a Turing Award winner, is definitely a heavyweight. On X, he directly retorted with a "False."

In a linked LinkedIn post, LeCun elaborated on the differences between research and engineering in four aspects: mode of operation, methodology, openness, and evaluation criteria.

The following is the detailed content:

There are differences between research and engineering in these aspects: (1) Mode of operation (2) Methodology (3) Openness (4) Evaluation criteria.

Research uses the scientific methodology to discover new principles, prove that they can be effective in practice, analyze their advantages and limitations, and interact with the broader research community for criticism, verification, replication, comparison, and improvement. The criteria for research are conceptual simplicity, theoretical beauty/interpretability, and obvious performance advantages over existing technologies on some recognized indicators. This is true for both academic and industrial research.

Engineering, on the other hand, integrates methods (usually developed in the research mode) to build working systems. The concept of engineering is to use the first set of methods that perform well enough on the current task. It usually involves a lot of tinkering, adjustment, fine - tuning, and occasional chaos, with the current goal of improving the performance of practical tasks. The criterion for determining whether the method is absolutely the best is less important than whether it is sufficient to complete the task at hand.

Researchers are mainly evaluated based on their intellectual influence. Evaluating research is a difficult task because the impact of the results may not be apparent until many years (sometimes decades) later. Therefore, evaluations usually rely on the collective opinion of the research community, which must be achieved through proxies such as publications, citations, invited speeches, and awards. This is one of the reasons why research results must be published.

Engineers are mainly evaluated based on product influence, and sometimes proxy indicators (such as pull requests, lines of code, etc.) are also used.

By operating in the engineering mode, researchers can be motivated to do incremental work. If you do not distinguish between these two activities and do not evaluate researchers and engineers with different criteria, you may stifle breakthrough innovation. True breakthroughs require teams with a long - term vision and as few restrictions from product development and management as possible.

In the past, the industrial research laboratories that left an indelible mark on technological progress (Bell Labs Area 11, IBM Research, Xerox PARC, etc.) all had research departments clearly separated from the engineering departments.

But then again, what exactly is the precise definition of an "engineer"?

Perhaps, as Richard Hamming, a Turing Award winner and mathematician, once said, "When doing scientific research, if you know what you're doing, you shouldn't be doing it. When doing engineering development, if you don't know what you're doing, you shouldn't be doing it."

Of course, clarifying the definitions is important, but more importantly, we should focus on what we are actually doing.

What's your opinion on this "researchers vs. engineers" debate?

References

https://x.com/ylecun/status/1950844792805507339

https://x.com/elonmusk/status/1950254103474446728

This article is from the WeChat official account "MachineHeart" (ID: almosthuman2014), author: Panda, and is published by 36Kr with authorization.