Любопытная тёрка между Лекуном и Маском про инженеров и исследователей https://www.linkedin.com/posts/yann-lecun_there-is-a-difference-between-research-and-activity-7356606929554567169-_iT2 There is a difference between research and engineering in (1) modus operandi, (2) methodology, (3) openness, (4) evaluation criteria. Research uses the methodology of science to discover new principles, demonstrate that they can work in practice, analyze their advantages and limitations, and interact with the wider research community to criticize, validate, reproduce, compare, and improve. The criteria are conceptual simplicity, theoretical beauty/explainability, clear performance advantage over prior art on some accepted metrics. This is true for research in academia as well as in industry. Engineering integrates methods, often developed in a research mode, to build working systems. The philosophy is to go with the first set of methods that work well enough for the task. It generally involves a lot of tinkering, tweaking, fine-tuning, and an occasional kludge to get the performance up on a real task. Whether the method is the absolute best matters less than whether it is good enough for the tasks at hand. Researchers are evaluated largely on intellectual impact. Research evaluation is a difficult task because the product impact may occur years (sometimes decades) after the work. For that reason, evaluation must often rely on the collective opinion of the research community through proxies such as publications, citations, invited talks, awards, etc. That's one reason research must be published. Engineers are evaluated largely on product impact, sometimes through proxy metrics such as pull requests, lines of code, etc. By operating in engineering mode, researchers are incentivize to do incremental work. If you make no distinction between the two activities, if you don't evaluate researchers and engineers with different criteria, you run the risk of killing breakthrough innovation. True breakthroughs require teams with a long horizon and minimal constraints from product development and management. The industry research labs of yore that have left an indelible mark on scientific and technological progress (Bell Labs Area 11, IBM Research, Xerox PARC, etc) were all research divisions that were clearly separate from engineering divisions.