Humarithm: why data without
people is not enough




And how algorithms lost the Content Curation battle



                                 Guillaume Decugis
                        Co-Founder & CEO – Scoop.it
                                            @gdecugis
Hi, my name is Guillaume.
I’m an Engineer and I like data too.




           And also now an Entrepreneur (Musiwave, Scoop.it)
We Engineers love algorithms.
In 1993, I built a neural network
algorithm…




                        …to predict stock
                      market performance
This is where I’d be if it had worked
People produce a
       lot of data
And yes, algorithm can filter/compute
data.
But smart systems use people to correct
data.
And people like to do it.
Most people don’t understand
algorithms.
But they can understand people who
                   understand data.
Data doesn’t distribute itself.
People do.
The Content Curation problem


    I’m talking
      to you!




                      I’m looking
                        for you!
Web Content in 1995. Curated.
Web Content in 1997. Algorithmed.
The Search for the perfect Content
                        Algorithm
“Curation starts when Search stops
             working” (Clay Shirky)
The Filter Bubble (Eli Pariser)
7 things that content personalization
algorithm do poorly

  •   Anticipation
  •   Risk Taking
  •   Big Picture
  •   Pairing
  •   Social Importance
  •   Mind-blowingness
  •   Trust
                             Eli Pariser, 2011
Man or Machine?
Humarithm
“The Algorithm + the Crowd are Not Enough”
Rand Fishkin, SEOmoz.




        We need humans data editors/curators.
Start from the API. Let Human Curate.
         The Scoop.it Example.
When it comes to technology, we all
                         want this.
So please don’t build this.
Thank you!


                                Guillaume Decugis
                                       @gdecugis

Author: Franz Steiner

Humanrithm: why data without people is not enough