Engineering as a noun versus as a verb really threw me for a loop on this one.
The methodology is cool, the scale of the experiment is very cool (16k meetings), the conclusion is kinda workmanlike - as if the question was specifically 'is shared knowledge necessary to generate new ideas'.
"Overall, this study takes a critical step towards identifying the processes that explain when serendipitous encounters shape knowledge production outcomes among innovating individuals.
We show that brief, information-rich interactions between people with some overlapping knowledge interests can have a productive effect on knowledge transfer, creation and diffusion."
This tertiary point was distracting as well, seems out of place.
"Third, we make methodological contributions by highlighting the benefits of long-term studies that amalgamate multiple forms and uses of data. Prospective experiments can support multiple lines of investigation involving both near-term and long-term outcomes that may not be possible in retrospective, archival studies and suggests the use of multiple sources of data for unpacking the dynamics of knowledge production."
The take away I got was that people whose fields are too similar don’t knowledge share and generate good knowledge creation nearly as well as situations where the knowledge transfer is with someone with only “some” overlapping interests.
Ie there is a Goldilocks zone for knowledge creation that involves some, but not too many, overlapping interests between collaborators.
It kinda makes sense once you think about it, but was nice to see the sophistication they took to show it experimentally. (But I’m not an expert, this isn’t my field and I only spent 5 minutes skimming the first portions of it.)
Considering that yesterday in a meeting I was casually arguing for peer groups within an scattered enterprise, perhaps time would be better spent with something less rivalrous like semi-peer groups? Too much of a shared interest is actually a bad thing? Worth thinking about.
Interesting. I wonder how this would differ for programming as a practice and models of hiring. Programming differs from scientific research in that in scientific research publications of new knowledge is a frequent primary goal. In programming, generally speaking, both as a practice and in hiring the goal of compatibility selection greatly exceeds both notions of competence and discovery. That distinction results in different sets of biases by which shared knowledge is accepted or discarded.
The methodology is cool, the scale of the experiment is very cool (16k meetings), the conclusion is kinda workmanlike - as if the question was specifically 'is shared knowledge necessary to generate new ideas'.
This tertiary point was distracting as well, seems out of place.