path: root/transcripts/startup-science-2012/richard-price-academia.edu.mdwn

Richard Price http://academia.edu/

I just want to start off by saying what a pleasure it is to be here. I am glad
that more of a spotlight is being shined on science. I think a lot of VC and
engineering capital goes into photosharing apps, and I think there are huge
opportunities, I'm not one of those guys who is down on photosharing apps, but
I think there are enormous opportunities around science startups. I think now
is the time to act. Why is right now an interesting transformation time?

When I think about what is going on in science right now and how to predict the
future, the main lense that I view the future through is that I imagine that
science has historically been for the desktop, saved as a PDF and upload to the
web. Scientific content is going to be made for the web and it will inherit
native web content. As that transformation plays out, that will be exciting.

The first characteristic is instant distribution. The importance of this is
hard to over-state. There is abuot 12 months for submitting a paper and then
seeing it coming it out. There's at least 12 months to see a response from the
community. You don't see that time lag on web content on tweets or facebook
updates or anything in general. I hope we can see this in science.

A second use trend is that historically in science there has been a single mode
of publication problem-  all content is shared as a paper. I think over time we
will see different kinds of , increasing number of scientists sharing blog
posts, status updates, I think the key thing to unleash is credit metrics.
Ithink we're talking about that too.

So a lot of people this morning have bee talking about instant distribution.
The natural question is filtering and discovery and discovering the wheat from
the chaff. They think that science will fall apart. I think when you look at
the web as a whole, as a content distribution system, it doesn't have a formal
peer review process. You have discovery engiens like search engines asnd social
entwroks. The search engines show the links and they show these as
recommendations, then they apply algorithms to the structure of the web to
develop rankings and to extracct the sentiment of the web for each web page.

With social networks it is much more explicit. What do my peers recommend? I
think you see this in science. When you talk with scientists these days, and
this has been for 5 or 6 years, scientists to use search engines. Nobody walks
down to the library to read the papers- they use the browser. Search ranking is
doing the best it can to do this. It should be on the first page.

The second thing is social networks. You go through the current system, you're
looking at a paper and it's being recommended by two peer reviewers, well, what
about all your peers or choose the peers for the judgements that you trust? I
would name 20 philosophers that I would adore to know what they read on a daily
basis. That would be my reading list basically. That doesn't exist yet, but
we're working on it.

The main metrics historically has been, to get ahead in their career are things
like where to publish, the impact factor, and the last ten years it's been
citation counts, that sort of thing. I think the generally, that generally, as
scientific content becomes web content, you will see metrics based on
attraction to the content on the web. Some of these things will be follow
counts, some of these will be stackoverflow scores, there will be certain
niches where you will see recursive algorithms for like PageRank, right now we
have just raw citation counts, we don't even take into account who cited the
paper.. you will see that change. I think the opening up of metrics will be
powerful, you will get credit for datasets where as right now you can't get
credit for data sets or blog posts. I think that will chagne.

Academia.edu is on a mission to accelerate the world's research. We have a few
million users. The rough structure of the site and how it works. This is Steven
Pinker's page. You can upload papers. You can have questions and answers. You
can have blog posts. You can have document view counts. Here's a full text
paper that he uploaded. This is the news feed. It's a follow model like
twitter. You can follow Steven Pinker and get his personal news. You can see
who you are following. Every paper that gets uploaded gets into my news feed.
It gets into non-distribution, or something, or bioinformatis or something and
it gets immediately sent out to 5000 people's news feeds and that's
distribution.

A really cool part, and it goes into the metric revolution, people see a
dashboard of how many profile views they got, how many doc views they got, and
some facts about their visotrs, which countries they come from, it doesn't need
to go very deep you can see the dates and what search keywords which ones come
from Google, and people, on a site, really like the analytics, it's like crack.
So here are a few quotes from users and their thoughts on the analytics and on
their professional lives.

Tim Ritchie, a Lecturer in the Department of Psychology at the University of
Limerick

You said this was for uploading and sharing research. In order to contribute my
research, do I retain copyright, do I link to it where it was already
published? How does it work? You don't give us copyright.. they think of
academia.edu as sort of like arxiv, they upload their papers, tehy upload
working drafts, preprints, and increasingly many journals are uploading an
author version, incorporation of peer reviewed comments, ... increasing
journals are adding a published version with formatting, so when that happens
they do that to..

If I was looking to browse, if I was just reading and following other
publishers, I would have to have access to read thsoe journals? No, they upload
the full text to our site.

I have a quick question about your incentivization of metrics. One of my super
fun jobs is that I have to analyze my faculty's portfolios and get to do lots
of scorings and go to committees and tell them that this person is worthy of
attending.. which department? USC. My question is.. we have a whole bunch of
fudge factors to make our guys look good. If you want new and better stats for
academic community to be accepted, it's the unviersity level where they set
what is the criteria for advancement, what are your thoughts for legitimizing
your metrics among those old bodies?

It's a great question.. when it comes to new credit metrics.. it's a main
catalysis for scientific transformation.. what we're seeing is that when we
built our analytics dashboard we didn't know that people would take
screensshots and use those in promotion packages. We used that for internal ego
boosting, but we didn't know that people were including that in their grant
applications. As we see that take off, increasingly we will see that these
cocnepts will socialize.. how many page views do you have, and increasingly,
maybe 10% of applicants this year will include metrics in their grant
applications. It will start to reach a tipping point where 90% or 75% have
metrics, and these 25% have no metrics and just rely on journal titles, I don't
know, not enough data there. I think more data is good. The competition for
funding is so unbelievably intense, even for jobs and products, academics are
desperate to see any areas where they can get ahead. People are doing this more
and more to see.. understand these concepts. It happened with citations, I
think it's very much on the uptick, they probably won't fix... hiring
committees and grant committees 10 years ago.. and I think it's a bottom up
grass roots movement. I think that's how we're going to see.. these stats and.

We don't have an API yet and we are planning to build one. It's so obvious that
Facebook and Twitter advantage from an API and we need more engineers. If there
are scientists in the audience, shoot me an email at richard@academia.edu, so
tap me on the shoulder.