10 ms·
Building has a different kind of rigor than academic writing, because when you build, you have to make something that works. that's stupidly offensive (i'm not
by ta0967 11y ago
Building has a different kind of rigor than academic writing, because when you build, you have to make something that works.
that's stupidly offensive (i'm not an academic).
- keepitsmall 11y agoThe idea that it was more important to demo and build than it was to publish was–and is–an important part of the Media Lab culture. It’s through the building of things that we learn, explore, and discover. Building has a different kind of rigor than academic writing, because when you build, you have to make something that works. I didn't read that as offensive. The lady doth protest too much, methinks.
- rayuela 11y agoStop with this context nonsense. I've already got my pitchfork at the ready!
- azinman2 11y agoNot really. Many write papers that are purely theoretical. The media lab is all about having the skills to realize your dreams, and in rendering it into practice you learn a lot. SV has a similar mindset of MVP/iterate because the product has to be good to survive. In academia you need just papers and citations, regardless if anything you write can be made into anything at all.
- rifung 11y agoI'm not sure that papers being theoretical means that they're less rigorous. You could argue the opposite as well; they're more rigorous if you prove strict upper bounds whereas practical papers just require you to implement something and measure it. I don't think either one is better or worse than the other; they all have their place and are just different.
- nl 11y agoIt's not offensive at all, and I think most academics would say that it is exactly right. The rigor in writing is around making sure your claims match your evidence. That's dramatically different - but no less or more hard - than building something that works. For example, it my area (ML) it is well understood in academic circles that some models that work well in research don't work well in production systems. This is usually because there is a lot of hard engineering that is outside the scope of a publishing a paper on a new ML model. (I'm not an academic but I run the engineering arm of an academic research program)