The Internet is constantly evolving. As the speed, flexibility and complexity of connections increase exponentially, the Web is increasingly beginning to resemble a biological analog; the human brain. But what exactly is it that’s makes us, or the Web, smart?
In the brain, neurologists now believe that it is the density and flexibility of theconnectionsbetween neurons, not simply neurons themselves, which are at the root of intelligence. These chemically-mediated connections are called "synapses."
Even if the total number of brain cells, or neurons, begins to diminish in early adulthood, our ability to generate new connections between neurons and between different parts of the brain – what neurologist call “plasticity” - persists throughout life. What's more, the brain seems to be like any other muscle in that plasticity can be increased with exercise.
It is at the synapse, the "gap" between one neuron and another, where neural connections are consummated to create pathways that, when used, are reinforced with additional connections and, when unused, are "pruned" to make way for new, more useful pathways. It is these incredible chemical bridges that define the patterns of communication that, at any given time, define our cognitive capacity.
We believe that this evolving view of neural science provides an increasingly apt metaphor for what we call the "Synaptic Web" in that the connections between objects are more important than the objects themselves. The question is; how are these connections changing to create new experiences? In other words, there is an opportunity to stop looking at the nodes and start looking at the space between them.
The exploding variety, speed and flexibility of electronic connections - those between people, data sets, applications, the real world and the online world, gestures and meaning and content and communication – is at the root of what some have called an evolving “collective intelligence.” Thus, the Synaptic Web is about the evolution of the Internet from document delivery platform, to a platform for communication ("2.0") and now towards something much more profound: a dynamic web of adaptive "organic" and implicit connections whereby real-time information flows give structure and meaning to previously unconnected sets of data. The Internet is a sea of conversations streaming through connections, and these patterns have meaning.
Signs of the emerging Synaptic Web abound.
Where we once built destination websites that had loose connections in the forms of links, we now build widgets; lightweight pieces of functionality that connect countless sites and services using rich, deep and meaningful pieces of functionality. Like individual neurons, "sites" must now maximize their connections to outside data sources and applications in response to external stimuli or risk being "pruned" themselves.
The same is true for Social Networks. These once mighty silos of communication are becoming ever more deeply connected to the fabric of the internet. Ultimately only the connections between the people (the 'social graph') and between their social objects (images, profiles, links and groups) really matter. These connections are now becoming part of every website we visit and they will become ever more intrinsic to our user experience. Private social networks will of course persist, but it will be users themselves, not the platforms, who decide with whom, what, when and where connections occur.
Social profiles are becoming real-timestreams. If the old profile was a neuron, the stream is a neural pathway or pattern. It is the connective tissue between applications and people that feeds information from one node to another. Profiles come and go, people express themselves using countless tools and technologies - the stream, however, is the consistent and persistent channel that matters. It is the new presentation metaphor that increases the level of information we can consume while reducing our sense of overload. Just like synapses, they fire, and like synapses, it is the collective patterns of multiple firings - multiple signals or re-tweets - that creates a pattern. Patterns create meaning. Tune in, tune out, it doesn't matter. The information will find you ifitmatters. Implicit information derived from content and gestures is one of the great opportunities of the Synaptic Web. To observe a set of gestures and connect them together creates a dynamic profile of interests, intentions and friends that can be used for discovery and filtering.
Mobile phones once connected people via their killer app - voice. Their killer app now is computing and connecting in all flavors. From games to web browsing to location based services, phones are now hyper connected devices. Mobile devices represent powerful sensory nodes and are more pervasive, more personal and more aware (in terms of sight, sound, location) than the stationary PC platform ever could be. With geo-location data connected to messaging, mobiles are already enabling fortuitous connections between friends, friends of friends, like-minded strangers or countless commercial opportunities, and or all of which just happen to intersect in time and space. This is only possible when communication speeds reach near-real-time.
Databases are becoming data peers. Most leading thinkers now agree that storing and hording data is the surest way to devalue data and miss out on a much larger, more profitable opportunity. Connected, complete and current data stored in open formats that 3rd party tools can use and leverage has been the basis for every lasting Internet innovation. The emerging data network of interoperable peers builds on the TCP/IP and HTTP layers before it to become the underlying infrastructure for much of the future of the synaptic web. These are complete and current data flows that make their closed database cousins seem stale and insignificant. They will be powered by new standards like OpenID, OAuth, ActivityStrea.ms, PortableContacts, APML and more.
In the Synaptic Web, filtering is more important than search. At the very least it is the next generation tool to help navigate the content discovery problem. While search is about narrowing the infinite document web to a digestible set of pages, filtering is about narrowing the torrent of streams, nodes and networks into something that matches your current and evolving criteria. It's about defining and constantly refining your world view so that information can findyou.
Filtering is not just a synaptic web opportunity, it is a synaptic process in and of itself. Making connections between the content and conversations in your daily life and your current interests and future intents will result from machine learning, human teaching and user experience improvements. All of these will make the synaptic web easier to use, but also more powerful.
These changes can be summarized in terms of the following general characteristics:
More real-time connections E.g. Email vs. Streams
More types of connections E.g. Yellow pages directory (Search = Results) vs. Augmented Reality (Video + Geolocation + Directory + Personal Interests)
More density of connections E.g. Quicktime VR (One author, one image) vs. Microsoft Photosynth (100s of Authors with 100s of Pictures automatically stitched/connected together)
More organic connections E.g. Facebook connect hub-n-spoke model where the target API is known vs. Peer-2-peer DataPortability where the endpoints are open and interoperable
More implicit connections E.g. Typing in broad, static interests vs. Automatically detecting changing interests from lifestreams
Deeper connections E.g. Links to sites vs. Embedded widgets that combine functionality from multiple sites
So, what's a “synaptic” application? We think they reflect some combination of the following traits:
They connect two or more categories of things together (E.g. People and Data, Content and Communication, Data and Devices, Places and Companies)
They create or derive new/novel meaning or utility from implicit connections (E.g. Interest profiles, filtering, visualizations)
The connections they enable adjust in real or near-real time to changes in users’ behavior or other inputs
They bias towards implicit connections that are strengthened or weakened by actual behavior rather than explicitly stated connections that are arguably less accurate and relatively inflexible.
They use the web as the platform (E.g. Open Standards and Interoperable Endpoints)
They apply a variety of inputs to extend existing applications (E.g. GPS applied to maps, interests applied to dating)
They become stronger through network effects (E.g. crowd sourced images, social gestures etc)
Though they might have a companion site, they are defined by usership and information flows and are untethered from any given destination site
One of their primary inputs and/or outputs is the stream
The Synaptic Web is forming. It will be made of small pieces loosely and organically joined. More profoundly, though, the connections between those pieces will be just as important as the pieces themselves. The connections will be interoperable and create spontaneous meaningful interactions. The network will look less like a two dimensional spiders web and more like a three dimensional human brain. And like with all things, the more connections occur, the faster our pace of innovation and serendipity.
Comparisons to the Semantic Web
Not surprisingly, many have asked how the Synaptic web relates to the Semantic web. The question stems not just from the similarity in terms but also the observation that they address many of the same issues. This is no accident. Our ideas about what we’ve called the Synaptic web are directly informed by observations, our own and others', about the status of Semantic web ideas and initiatives that have been in play for well over a decade. Specifically:
1) semantic systems are still largely confined to the lab - however brilliant in the abstract, real-life technical and (more importantly) organizational and behavioral obstacles continue to block widespread adoption. It is now painfully obvious that machines often fail to make the right connections because only the most diligent humans have the patience and understanding to properly classify their work and others have incentives to intentionally misclassify content.
2) semantic analysis and data classification is simply one way to create richer connections on the network - it may be that semantic systems can achieve scale. We'd love to see it happen. Meantime, the advent of the real-time web, increasingly effective publishing, sharing and engagement tools make it easier to find connections between nodes in near-real time by leveraging the power of systems (machine learning, neural networks, etc.) to infer meaning and connections by observing human gestures at scale.
3) real progress toward a richer and more connected and “intelligent” web is occurring – we think this is happening as a result of the sheer mass of content and density of connections that now exist and through a process that more closely resembles biological/neurological change than an engineering effort.
It's important to note that, whereas the Semantic web has always been a prescriptive project - specifically concerned with architecting the web’s future - we’re more interested in building a descriptive framework to help us understand what’s actually happening, how and why and with the goal of keeping our own systems "firing".
The following table is our attempt to highlight what we think are some of the key distinctions between Semantic and Synaptic frameworks.
Prescriptive – A vision of the future web and framework for achieving it.
Descriptive – Attempt to frame what is actually happening
The purpose of this document is to present a straw man overview of emerging trends on the next generation web. We encourage participation and conversation about these proposals so that we, as participants in this ecosystem, can come to a communal understanding our current and emerging opportunities for the web.