Quartz’s new app has come half way in reinventing news

You probably have heard about Quartz’s new innovative iPhone app. Over the past days, the news about it was all over the Internet. It’s too early to conclude whether this app will turn out to be a hit among users. The texting-styled user experience is definitely innovative and fits very well to this year’s hype topic, conversational interfaces. However, in the end, the possibilities to interact with the news items are pretty limited, which can quickly lead to boredom.

But that does not need to remain like this. In fact, Quartz has come half way in changing the distribution and presentation of digital news for good. What’s missing? The text field in which readers can write their questions, comments and requests and through which they can access all the information and knowledge they desire about a specific piece of news. Such a field would require what’s usually labeled “artificial intelligence”, but the absence of that does not surprise. Creating a personal news bot that is capable of interacting with users around news and that understands their remarks and inquiries was, until recently, pretty close to rocket science. And it’s still hard.

Thanks the recent advancements in regards to deep learning and artificial intelligence as well as to initiatives that plan to open source the underlying algorithms, maybe very soon the creation of conversational smart bots which natively “understand” the information they serve will become much easier. But we are not there yet, which means that from Quartz’s perspective it made sense to start with something simple.

Sign up for the weekly meshedsociety.com newsletter so you do not miss any posts. Example.

But if Quartz-owner Atlantic Media decides to continue the development of the app, adding the possibility for users to ask questions or to comment as a way to access additional information would turn this app, in my eyes, from a novel and sexy experiment into the future of news.


Take the example above. Quartz tells me that “Coca-Cola is selling more soda by shrinking its soda cans”. After reading that I am only offered a pre-defined question or I can move on to the next news. Again, the approach is fully understandable for a first experimental app launched by a mid-sized publishing house. But as a user, what I really would like to do is posing the questions that I actually care about. For example:

“What was the old size and what’s the new one?”

“How much more are they selling?”

“Do you have a photo comparing the two cans?”

“How often has the company done something similar in the past?”

“How is Pepsi responding?”

“What is the world record in drinking Coca Cola?

“Coca Cola is bad for your health!”

Basically, I want to do what I would do if I learn about this news from a friend or colleague: Asking more questions or voice my point of view. And then, the news bot would follow up by providing me with the requested information or with data on my potentially casual remarks, like “Here is a link to the latest research about the health effects of soda consumption”.

I have zero doubts that in the near to mid future, a news application or service will work that way. Once we have reached that point, the whole nature of how news are being presented will change, and chances are good that the static article as a container for news will slowly become neglected. Thus, the conversational interface for news could end one of the most annoying and inefficient practices of digital journalism: The fact that each and every article about news only contains about 20 % actual new information. The remaining 80 % are a summary of the status quo or of previous, related events (be aware: deliberate generalization! It’s not always 20/80). This practice is annoying for most readers and for the publisher, it is a waste of creative resources. Some do it on purpose though so that they can churn out a whole new article for a tiny piece of information in order to harvest ad impressions.

Conversational interfaces allow for a full personalization of the news experience. People who already know the whole backstory will not need to skim through a long static article just to find the 2-3 sentences containing the new information. On the other hand, those who are not experts on a certain topic can specifically request to receive a summary of what happened previously. Over time, the algorithm will be aware of their level of knowledge about frequently returning topics, allowing for even better customization.

First the Internet led to an atomization of the newspaper into many news pieces which can be distributed separately. Now, the next step is the atomization of the news itself into different bits which are being assembled on-demand and in a personalized, conversational manner.

After the failed news app Circle tried to work based on this philosophy but lacked the conversational aspects, Quartz gets closer to such a state. To fully accomplish this might still take a few more years, or it could happen very soon. Google and Facebook, the current frontrunners in artificial intelligence, will for sure not just watch.

If you like what you read, you can support meshedsociety.com on Patreon!


  1. An interesting follow-up question would be: If the future of news is a conversational presentation through bots, how will those bots be fed the information and who will do it; i.e. what does this mean for journalists?

    • I guess the role of journalists who cover day-to-day news would simply change slightly. Instead of writing a continuous text into a WYSIWYG editor, they feed the information separated into multiple paragraphs (of various length and depth) into a special CMS. Then, depending on the machine reading capabilities of the algorithm, they might have to tag the content blocks or in other ways reference them to each other. Also, they’d probably not publish everything at the same time, instead they’d add new information or update previous information step by step.
      Does that make sense? What are your thoughts?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.