Hacker News Reader

Comment by robotresearcher | Hacker News Reader

Most of those papers’ first authors were in industry shortly afterwards. Most grad students go directly to industry and take their knowledge with them.

View on HN

Is that really how research in CS works? Certainly we hear about grad students doing lots of the work in laboratory sciences like biology, but in more analytic subjects like mathematics or theoretical physics its not true at all. Grad students making significant contributions is extremely rare; basically all of the real research is done by tenured professors.

Can those of you who did postgrad CS tell us how much of the research is actually done by grad students vs professors?

CS PhD here from one of the top programs (currently in industry). Personal friends with faculty at several top schools.

Like 95% of research is done by grad students.

In my experience it's often (but definitely not always) the case that for new post-grad students the idea or general research direction often comes from the professor (especially if the research is funded by a grant proposal). Most of the heavy lifting is then done by the grad student with guidance and feedback from the professor. As the student matures they start to operate more independently.

In my experience (new professor), 90-95% of the work is done by grad students.

A good prof in my mind acts a mentor, providing frequent feedback on the research and contributes to writing of the paper. The feedback is often invaluable, but still just a small part in the overall amount of work compared to what the student puts in.

> Grad students making significant contributions is extremely rare; basically all of the real research is done by tenured professors.

...but most of these first authors (the people formally credited with doing the real work) are graduate students, so how does that fit in with your theory?

I specifically made that claim of mathematics and theoretical physics, and requested first hand accounts of how it worked in CS.

So how does it work in your field? Making your own significant contribution is literally the basic standard required for a PhD in most fields.

Well i think it just depends on how 'significant' you need the phd to be. As an example of what I mean, consider the first issue of this years Annals of mathematics [1] and note that there is not a single grad student on any of the papers, first author or otherwise.

[1] http://annals.math.princeton.edu/2018/188-1

While the general idea might be professor-inspired, 90% of legwork is carried out by grad students.

Depends a bit on the professor. Young ones are more involved than older ones for a variety of reasons. In either case, the bulk of the work is done by grad students. Professors are more like managers that provide direction, feedback, etc.

I don't see how "I can't learn more because they work for ExampleCo on proprietary technology" is an improvement on "I can't learn more because it's behind a paywall, along with everything it refers to and everything that refers to it".

Google was founded, quite literally, out of studying the graph of literature. Here's one way they can move that original vision a little closer to reality.

Are there any other influential companies that are like Google? I would love to know how well academic research transforms into industry.

IIRC, VMWare was started by a Stanford professor and directly based on his research in virtualization. I can't find the full story of the transition from academia to industry, but SimOS looks related for sure:

https://en.wikipedia.org/wiki/Mendel_Rosenblum

(If anyone has details, please chime in.)

- Pixar had a pretty long road to industry, but it has deep academic roots. Co-founder Ed Catmull published significant original research in computer graphics.

- Yahoo was also started by 2 Stanford grad students, although it wasn't based on any research that I'm aware of.

- Werner Vogels was one of my professors at Cornell, doing work on distributed computing, e.g. on gossip protocols, and he is now CTO of Amazon.

- Guido van Rossum worked in Andrew Tanenbaum's lab (a professor whose OS book you might have), and Python was at least partly inspired by that experience. Python now runs much of the industry :)

It would indeed be nice to see a list of "papers that companies are based on".

The PageRank paper is well known (although Google search no longer resembles it). I suspect you can do the same thing for VMWare and Pixar, at the very least. Those two companies relied on significant technical advances.

EDIT: Also Cisco and Stanford:

https://www.siliconvalleyhistorical.org/cisco-systems/

This goes all the way back to Hewlett Packard, although I'm not sure if they're closer to Yahoo's case (same people) or Google's case (same subject area):

https://en.wikipedia.org/wiki/Hewlett-Packard#History

Industrial research has a long and influential history -- IBM, AT&T's Bell Labs, MS Research -- all doing pure research with generous budgets.

But if their work is published by the IEEE CS or the ACM, you may not be able to read it. This situation has improved a lot in recent years as many authors will post copies on websites they control. But that doesn't account for the decades of work otherwise unavailable for direct reference.

Reply on news.ycombinator.com