Metasphere

Hierarchies became such a stronghold when it comes to how we structure information because they make complex representations easier to digest by replacing relations that should be curved or crooked with straight lines extending from a shared node, discarding every other possible relationship. Hierarchies are constructs that reflect only small portions of networks to make them digestible. But they reduce complexity instead of embracing it and break with a simple circular link. The reason why they still stick around is not that they are a good visual representation, but rather because they are good enough, easy to understand, have been around since the first folk taxonomies millennia ago and fit a “taxation mindset”. Much didn’t change since the first scribes started sorting scrolls into shelves and kicking off the categorization madness. That’s not because there weren’t smart alternatives, but because we are simply lazy and stick with the things that work just well enough.

Today’s metaphors, emerged from the viewpoint of the office worker and what surrounded them. We need to find a suitable successor to those desktops with files and folders, screens and — oh look, a few meters away — windows and all the other tangible representations which have been around for so long. One where we step out of the office and into the world to embrace the complex, fractal, multi-relational, mind-expanding, world-spanning nature of the information space and our quest to explore, while still remaining uncluttered and accessible at first glance.

metasphere uses an approach that pays respect to the multi-relational nature of information.

Instead of organizing your knowledge in a hierarchical way, with metasphere you create literal worlds of wisdom, where every bit of information can be related to others. But that comes with one caveat: complexity. Even though the network meta­phor is far better suited for organizing knowledge, we lack the capability to easily comprehend and navigate abstract network structures at first glance.

Luckily there is another visual representation that encompasses the same charac­teristics. And we have been trained to read and understand it for centuries: geospatial information as expressed in topologies and maps — worlds.

Need a space to hold all the things in our minds? Let us build worlds. Maps and the visual representation of our spatial surroundings are cultural techniques that likely even precede said drive to put things into hierarchies. We have become adept at navigating geospatial representations — i.e. representations of clustered objects, forming networks — without effort. It’s what we do. We live in that space and dream up worlds all the time.

Worlds, where information that goes along well with each other, tends to settle in the same areas. Where neighborhoods of like-minded pieces of information develop. Those, in turn, would be incorporated into cities, countries and finally continents. Pieces of information that have few relations with others, would become isolated on small islands while those with the most connections form the plateaus of your interests, appearing as huge mountains.

It’s so obvious, we talk about it all the time, especially when referring to the information space: domains, information highways, cyberspace, spheres of knowledge, mind maps, mountains of data, seas of information, routes, hubs — all spatial metaphors.

While there have been forays into the spatialization of information, all of the fascinating projects working with spatial metaphors for information organization I’ve come across, from maps to physics engines to projecting information into space, all their approaches lack in a crucial way — they do not embrace complexity.

In order to achieve fast information retrieval, all of the landforms would have to have distinct shapes, ones that we will learn to recognize and discern without difficulty.

Our brains have evolved perfect pattern recognition tools. We see them everywhere. Dreams are our brains, trying to make out patterns in random noise. But to be able to do this, we need noise in the first place. We need to have that one neuron firing a cascade of electrical activity up the network, eventually having that one neighboring neuron tune in, triggering auto-association, leading to the recognizing of a face in a cloud — or a landform for that matter. Why deprive our thinking with the simplest representations, stripping us of this great ability when we can have rich worlds of wisdom? We need real topography with weird forms, not colorful dots connected by lines, looking all the same.

Because we are adept at differentiating geospatial patterns information retrieval becomes fast and intuitive.

The fractal nature of information would also be suitably represented. When you look at how matter is organized in our universe, the same pattern always emerges: from microwave background radiation and the filaments of dark matter to galaxy clusters, universes and star constellations right down to the topographies of our lands, cities and even further down into our brains and cells — it’s a huge fractal dive down into an immensely complex networked structure, very much like our thoughts. So by adjusting the metaphor to a scale and look that we are familiar with, we would make the vast space of our interests navigable. You would zoom in from the solid mass of a continent, along a winding mountain chain right down to a grain of sand that could be a planet in itself.

After populating our worlds, we could then share parts of them, explore different worlds or do route planning and offer scenic rides our train of thought.

When I first started thinking about the usefulness of using topographical formations as a metaphor aimed at the information retrieval problem and after some initial research, I initially feared that it would be a niche application — it could only be with all this research already done, all those great ideas and still no working solutions. But my motivation reached an inflection point and was reinvigorated because of the developments in two technological fields: the accelerated progress in machine learning and the widespread acceptance and growth/hype of what I would call “decentralized trust environments”, i.e. blockchains.

When learning more about the possibilities of those technologies, I realized that you could actually pull this project off on a wider scale because of two reasons:

The embedding of the information could be automated

The cumbersome process of individually tagging and correlating information can be automated by machine learning techniques. When we look at how we humans process information, we see that we are incredibly good at memorizing the underlying themes of a piece that we read, but not at all when it comes to memorizing them word by word.

When we digest information, that’s exactly what we do in the literal sense. We go through a stream of data, picking out the patterns that interest us and flag them if they resonate with what we know so far. Everything else is filtered out and doesn’t leave a dent in our perception or memories. This is where we turn mere data into information. The next step is turning information into knowledge, by incorporating it into what we already know. We take a specific piece of information and put it on a map — near like minded information, be it complimentary or contradictory to what there already is. This in turn makes up what we perceive as knowledge — interconnected and annotated patterns of information. Take this a step further and you might end up with something like wisdom.

The one critical component in understanding things is not merely just reading it, but processing it. By dissecting, correlating and revisiting information, we acquire knowledge.

That’s why we want to encourage to dissect information while reading it. As a starting point metasphere gives you an algorithmically created summary on top of each text. It tries to determine the neighborhood the text would fit in and tags it accordingly. Whenever you highlight a passage, those highlights inform the algorithm of what is important to you and the summary is reshaped. Sections with no highlights get folded away, allowing you to quickly get up to speed whenever you revisit a text. And when another user reads the same text, they in turn profit from the highlights that you made. Remember the time when you got that schoolbook with the old annotations from the previous owner in it? metasphere wants to be the Half-Blood Prince for everybody.

You could incentivize the actual “understanding” of texts

Traditional media outlets are closed ecosystems. The most common mode of compensation for them is through advertisements. But readers want to immerse themselves in information without distractions. So they install adblockers. Those in turn get detected by the sites offering the content and access gets cut-off. Or they try to build paywalls around their content. Everybody knows that this conservative approach will break in the long run. Information inherently wants to be freely distributed, shared and discussed.

Because we believe in the freedom of information, we designed metasphere to offer an alternative to the monolithic information silos that are the norm today. It is a decentralized network in which content is not stored on a single server, but rather on a distributed network of individual nodes. By offering incentives to the authors of content and also the readers who work with it, we want to let content creators be independent of the grips of the advertising industry.

We also don’t want to compete with other platforms facilitating the free flow of information, but to support them. We love what projects like Medium do for the information landscape. That’s why we aim to give authors even more reasons to publish their content freely and use a system that makes it possible for authors to claim authorship and hence receive compensation for their pieces.

The gaining acceptance of blockchains as means of conducting transactions between unrelated entities without a trusted middleman gave way to the possibility to incentivize the actions of persons on a scale and in a granularity previously unavailable, simply because traditional delivery of money is conducted with units that couldn’t be suitably subdivided and — more importantly — because we didn’t have the technical and societal means to tie financial transactions to a specific set of events in a secure and transparent fashion. Because a critical mass of people began to value cryptocurrencies, no matter how twisted and self-referential their motives may be at times, this is now ultimately possible.

Facilitating micropayments, not in a single-directional (i.e. consumer/user to the content-creator), but in a bi-directional way appears to me to be the key to the success of a system like this. Instead of simply compensating the original creator of a piece of information, no matter how granular the scale, people would pay for blindly just “dumping” information in the system, but at the same time be rewarded for reading, dissecting, connecting, annotating, fact-checking and sharing pieces of information, thereby providing “proof of work” that they gain understanding while also training the underlying machine learning algorithm, even if it’s just a single highlight that they make in a given text. So by incentivizing the users’ behavior and giving extrinsic motivation — in a monetary but more importantly also in a cultural sense — one would artificially shift peoples inherent drive for entropy and inertia that naturally comes with overwhelming floods of information towards a motivation to accumulate and disseminate knowledge — for themselves and the greater good.

This way metasphere creates an ecosystem that helps organize knowledge — and makes the labor of understanding legible.