Running rabbit

Relevant Information Delivery, Built for Speed

6 minute read
Peter Prosol avatar

We’ve spent a long time trying to solve what sounds like a simple problem: how to get people information when they need it. Sounds basic, but it can affect everything: what you think, what you do, what you say, what you create. Every minute you spend looking for the right data could have been a minute spent building something new, or having an experience, or strengthening a relationship, or gaining a new insight.

People have used everything from carving on rocks to library cards to systems for organizing paper notes and drawings to get a better handle on information. With the advent of computers, we got the Graphical User Interface, hypertext and search as several key technologies to find and retrieve information efficiently.

But as information continues to explode (over 90 percent of all data created by humans was created since 2011) the pressure keeps rising to build more effective tools. And now we have a new problem:


We already have an unimaginably complex variety of information. The Internet of Things opens up a whole new front in the information overload problem, and will require better ways to capture, index, retrieve and distribute information.

Caught in a Trap

But there's an obstacle. A lot of our information is currently trapped in complex, variously structured and formatted websites. If you wanted a data table representing the relationship of kittens to happiness right now, you would typically have to define a search term, process the resulting index of results, choose an entry and then browse one or more websites of wildly varying structure, navigational ease, etc., until you found your target. And you'd waste a lot of time and energy in the process.

The basic building block of the web is a site. Notice the static term—site. It’s a destination. It’s also a virtualization of an old concept: a document. The web was built to handle research papers and that intention has stayed in its DNA even though we have dramatically changed it in the meantime. HTML still very much looks like a sort of table of contents for a college paper with sections, paragraphs, bullets and so forth. Google and others have embarked on the mission to break down these sites a few notches (for example by picking out images), but in doing so they’re trying to correct the shortcomings of a relatively heavy, complex information structure with a top down solution.

Atomizing information and indexing the smaller, discrete components can transform the Internet. Content wants to be free. Enter the card.

A Little Bundle of Content

What are cards? They are small containers of arbitrary size to house micro-content. Micro-content refers to small, visual, meme-size chunks of knowledge that summarize other, richer information. Micro-content serves both as primary content itself and as an index and link to other content. A card can serve as an index, link, mini-site, piece of content and even app all in one little bundle.

Websites are lengthy, complex, coherent, collections of information. Cards are precise, context agnostic bits of information. Just as we couldn’t do quantum physics before we got down to the subatomic level, so we want to break the atoms that are websites into their smaller components and let all the energy escape.

Cards and micro-content are a table in a report versus the whole report, a video on a site versus the whole page, a paragraph versus the whole essay. They can be reshuffled, recombined, shared and cross-pollinated. If this essay were built using cards and more sites adopted card protocols, this paragraph could appear as a quotation on someone’s site, and if an edit were made here, it would instantly appear there too. 

Additionally, it's possible to track what readers find most interesting and sharable. Analytics could look at the interaction of users with content wherever it appears, whether on a site or an Apple Watch notification or a data point in an API. 

Just as every word, concept and piece of knowledge is a collection or summary of other information, so each card can be made up of other cards, which themselves comprise or link to other cards.


By atomizing all information into micro-content in card format for visual presentation and data management, we would have a flexible, scalable and precise system for indexing and navigating knowledge.

Learning Opportunities

But however efficient we make the entire system of storage, indexing and retrieval it will always have a critical flaw. 

We are too slow. 

Google is faster and slices through more information than a student looking for books in 1900 could have imagined. And yet we must still type into it (or increasingly, speak). We must review the results. We must click them and read what’s in each. And finally, after 10 seconds, a minute, a few hours, we find what we wanted. 

And we are already getting disrupted.

Built for Speed

Starting with more boilerplate, predictable types of information like upcoming flights, standard commute times and weather, Google Now is trying to put some of the key data you’ll probably want to retrieve on auto-pilot.


With advances in artificial intelligence and machine learning there is the promise that more and more information could be served up proactively by a computer that knows what you really need to see, and in fractions of a second versus however long you’d spend hunting for it. Progress on that front may make pretty much any design or organizational decisions we make today irrelevant, as computers inevitably make more rapid and effective improvements.

But for now, we must still manage our information with the help of computers. And cards offer an opportunity to do so effectively with better information taxonomy and more precise organization. Additionally, the insights that can be collected by watching billions of interactions of people with much more specific and discrete content would reveal network patterns an order of magnitude richer than those from site ranking algorithms or eyeball tracking studies. In an Internet built on cards versus sites we could more easily determine what people find relevant. Such a system could potentially provide fodder for AI to dig into and improve its own algorithms for information relevance and retrieval.

Cards and AI bring us closer to the dream of getting the best information at the best time with the least work. By cutting out unnecessary work and expanding our field of vision, this approach could give us the greatest potential to accomplish whatever inspired us to reach for knowledge in the first place.

Creative Commons Creative Commons Attribution 2.0 Generic License Title image by  Kumweni 

About the author

Peter Prosol

Peter Prosol is a software developer and serial entrepreneur. He is the Co-Founder of Cardstreams, which offers a way to create, manage, curate, and distribute bite-size content in activity streams, feeds, or timelines and display them as cards with a few lines of code.