Tuesday, January 10, 2017

graphical/universal interfaces, filters, filter bubbles, artificial intelligence, global mind



I want to say a few things about building new interfaces - universal interfaces, maybe you could call them. Let's see, how do we get from here to there? Well, I've been writing about, blogging about, this stuff for a few years. For a few years I've been talking about graphical internet - making it more fluid - making your internet interface a 3d graphical experience, like what we're used to with a lot of video games. So, that would be a project. I still feel strongly that it would be a very worthwhile project.

I don't think it's much beyond the difficulty and complexity of a lot of things that have already been created.  Minecraft and Roblox are examples of platforms that allow for creating 3d worlds and interacting with others in them. The next key innovations will have to do with connecting the different instances of the interface - different people's interfaces - together, so we'll be able to see each other, and making as much of the behavior of the interface as possible customizable by the user. 

What I'm describing here is much like visions of cyberspace that many science fiction authors and futurists have elaborated. I think we're very possibly to the point that we can really create it. And the reality, if and when it arrives, is sure to be amazing and profound in new, emergent, surprising ways, beyond what's been predicted in the fiction and the prognostications.

If you start by including, in the new interfaces, information from our current interfaces, data streams to & from from social networks, email, search, and other apps and so on, this can provide a starting point for populating the new medium. All of these largely text-based media can be included within a graphical superstructure. The old media become the first content of the new medium, and then we can go from there. So we'll immediately have all the same capabilities that we have now, plus the new capabilities inherent in a programmable graphical environment.

It really seems like something that could be built. I haven't dug into the programming and development of it myself yet. I would like to, and I might begin doing so soon, and I think many others have the ability to contribute very valuably too. I've never been much of a coder. I've been kind of a coder, years ago mostly. I feel like I know enough to say that, given our current resources - processing power, bandwidth, programming tools, browser 3-d graphics - something like what I'm discussing here could be put together in a few weeks, probably, with a team of a few people. Might take a few weeks. Getting more people involved might help too. There are ways for dozens or more people to be involved in a project at once, for sure.

However the development is organized, I want to suggest that it's important for it to happen. I think the pieces of this technological puzzle all basically exist, seemingly just waiting to be put together, and once we put them together, we will have created a profoundly new medium, a massive, and massively multiplayer, massively shared, new space, a new experience, so new that it might strike you as farfetched or hard to imagine. So let's try to flesh out the vision a bit more.

Roughly speaking, (anything we say in English words and sentences will be "rough" relative to actual code that implements these ideas; we're in an early stage of this project, the stage of trying to definitively plant this seed and get people more interested in this general topic) this interface will comprise a 3d space that we can navigate through, moving in all directions, zooming in and out, and adding, moving, and removing anything we wish anywhere we wish. Some entities in this space may be linked to specific resources, feeds, etc,. elsewhere in the world, and will be able to update and transform themselves in response to new information from the resources to which they're linked. Some entities will be programmed to act upon, interact with, other entities.

Virtual and augmented reality would be natural places to want to do this. But we have 3d graphics on our other screens too, on our desktops, laptops, and smartphones.

This can become the environment that we become used to entering into when we get online, instead of the current system, which is structured largely as separate pages. There will probably be tabs and windows and menus, etc., within the graphical environment -- we'll still use tools like these for switching between relatively separate, discrete components -- but a major part of the whole point of the new interface is integrating all our information into a single environment, allowing more visually apprehensible data to be included in a single landscape, allowing more fluid, continuous navigation through a visually rich data space. Among the multitudes of entities in this data space will be many varieties of tools that allow us to combine & organize all our data in new ways. With the availability of these new capabilities, it seems likely that we'll soon have much less use for making sudden, complete switches between entirely separate views, opting instead to organize our components, our views, our sub-windows, in many new, more flexible ways. When we do make relatively sudden, complete switches between views, we'll probably tend to find it advantageous to handle these switches within the graphical environment too, rather than using the tabs and windows of our browsers and operating systems. So it will be as if everything is happening within a single, all-encompassing page, rather than the myriad separate pages among which we're accustomed to shuffling.

When I wrote about this years ago, I speculated a lot about particular filtering algorithms that could be used for determining what appears in your interface (i.e. the "social graph transformation algorithms" in this blog's name). And that's still an interesting question. I'm not sure that it has the central, outsized importance that I seemed to ascribe to it in those years. It might still. It might still be that there are really interesting, hot, killer algorithms that could be used for that, still to be developed. But there are a lot of existing algorithms too, designed to handle similar tasks. So it's partially that this problem has been worked on and there are a lot of algorithms out there now. And partially, I think I've come to appreciate more fully the importance of the task of setting up setting up an environment that's fully programmable by the users, who will have available a whole ecosystem of methods by which to fine-tune the population of their data spaces. And so I'm not sure that the development of any specific filtering algorithms will be, as I used to speculate, a significant constraint or milestone in developing these interfaces.

Let's discuss the filtering some more. Once we have the 3d environment that has things imported from the media that we're used to, then we'll also have our methods of filtering what comes in. That function will be there - within the graphical environment we'll have settings for those filters. It's an interesting exercise to imagine ways of graphically representing the codes, the controls, that would be involved in determining the behavior of these filters for a person's universal interface. So people will be able to, within the graphical world, again, (I keep saying "graphical"; I think it's kind of boring, now that I just keep saying it, but it's going to be fun when it exists) -- within there, you're going to be controlling your filter that will control what appears for you. And that filter will be information and it will appear there in the graphical world just like everything else. So you'll be able to mix and match and share and show off your filters. And you'll have filters of filters. There'll be so many different filters, for so many different purposes, that we'll be filtering in order to narrow down which one or ones we want to select at a given time and place. So this scenario offers strategies for dramatically improving the quality of our filters, giving us dramatically richer, more accurate, more actionable, information about (depictions of) our world.

People have been talking a lot about the filter bubble phenomenon. And this graphical interface strategy is a definite way of addressing that. It addresses, and it's good for, a lot of other things besides that too. I don't know if I've really mentioned too much about the filter bubble phenomenon, by that name or otherwise. I don't know if I've written too much about it. I've written about a lot of other great benefits we'd see and enjoy from these new graphical/universal interfaces, and overcoming our filter bubbles is definitely another one of those benefits. By enabling our screens to be populated in richer, more fluid ways, ways that will be simultaneously more fine-tuned by us and more automated, we'll be bringing a bigger, more comprehensive picture of what's going on to everyone, in contrast to our current situation, where there's a balkanization, a lack of overlap, in the content that different people see. So graphicalization, at the same time that it brings us content more exquisitely tailored for every unique moment, will also bring us big, unifying pictures that will be shared by everyone.

It might seem that there are two competing considerations here: the desire to give users more control of their filters, and the desire to bring more automated intelligence to bear on the creation of the filters. But these two priorities will reinforce each other: when we create more intelligent filters, we're giving people better options to choose from. We want to allow everyone to participate in creating and sharing ever more sophisticated filters.

I hope what I'm saying is clear. I think it's along the lines of what a lot of people are talking about, curious about, interested in. I noticed a conversation online a few weeks ago. Someone asked, what can social media companies do to improve the filtering process, what steps, what recommendations, would you give to them? And somebody replied, give the users control over their filters. Make those aspects of the programming more transparent to, and tweakable by, the users. That's what somebody's idea was in response to that question, that request for ideas. And then a third person replied to that reply, and expressed concern that if we give people the ability to modify their own filters, that that will make even more of a filter bubble problem. You can see how you might be concerned about that. But no, I think it can only help, if it's very transparent, if it's moving in the direction of the kinds of things I've been talking about - if people can see what filters other people are using - and then you can share them, and see through those other filters. If I choose a filter that filters out a lot, and I'm only seeing a very narrow spectrum of what's going on, then yes, that's of course a concern, but I don't think people are going to do that when extremely sophisticated alternatives are readily available. Those narrow spectrums are what's going on now, with the algorithms, the filters, that are applied to our social media. They're pretty static, we don't have control over them, and we're getting too narrow a spectrum. We can widen the spectrum by opening up the process of programming the filters. Imagine seeing what filters other people are using. Sharing, mixing, matching, and going toward graphical. There's going to be too much going on, we're going to have too much of a desire to get too much information. I think this is the constraint we're experiencing now. Too much going on, too much that we really want, and can really benefit from seeing on our screens. And so there's a big incentive to create more sophisticated ways of generating, and presenting to our eyes, relevant and timely and vivid pictures of what's going on.

When we say that our media have helped encase us in filter bubbles, I don't think this implies that anyone has necessarily engineered this phenomenon intentionally. With very centrally organized media, only a limited number of people have the access to be able to contribute to these projects. It seems fair to say that bureaucratic/financial priorities have sometimes tended to stifle innovation to a notable extent. It also seems fair to say that we each have sometimes inadvertently cooperated in constructing these bubbles around ourselves, via the well-known dynamics of intellectual inertia that we speak of as confirmation bias, reality tunnels, etc. Our interfaces have largely followed a "pull" model whereby we explicitly choose to look at, or receive updates from, certain sources. We have a natural and even necessary tendency to stick to the comfortable and familiar to a significant extent. We have a limited amount of attention to allocate, so we often choose to expose ourselves to things that help us flesh out and reinforce the intelligible aspects of reality that seem to help anchor us in something solid and dependable. We must simplify in order to function at all, but we've often tended to oversimplify, for instance treating broad classes of people as caricatures and ignoring their input and their concerns. We're aware of the benefits of having a diverse variety of inputs, but our frequent, understandable preference for the familiar, combined with the limitations of our media environment, have contributed to this balkanization.

As the person mentioned above suggested, there may be ways for social media, within their current general structure, to give users significantly better filtering options. The focus of my efforts with this document is to help catalyze interest in the development of the next generation of media, which goes beyond the linear, textual general structure of our current interfaces. The new interfaces will be structured from the ground up as environments suited to providing maximally comprehensive and inclusive data spaces for us to interact with.

This talk of seeing other people's interfaces through one's own, and seeing how other people are configuring their filters, will likely strike some as a potential source of privacy and security risks. Any new technology that enables new types of sharing, new ways to interact, new forms of intersubjectivity, could generate such concerns. These interfaces will surely facilitate exposure of our lives and our minds to each other in many unprecedented ways, but we'll also have unprecedented control over the exact terms on which this occurs. So I think we can look forward to the new forms of connection that this will engender. Much new information will become available to each of us. And in fact we can expect to see many benefits specifically in the areas of privacy and security. Graphical interfaces can be expected to give us many new tools for protecting ourselves and for cooperating to detect and defeat security threats. Many of the significant privacy and security concerns of today seem to involve surreptitious gathering and dissemination of information and disinformation. Pushing forward into a new generation of media that allows all netizens to coordinate with each other more flexibly, comprehensively, transparently, resiliently, may be a significant part of the puzzle of overcoming these ongoing adversarial information dynamics. It's in the context of, under the influence of, these unfortunate current dynamics, that we have tendencies to feel suspicious of any new forms of connection. We'll want to exercise care and good judgement in many ways and participate in many types of growth and development - cultural and spiritual in addition to technological. If we can see more of what's going on at a time - make more and more information available to the eye, more quickly, more efficiently, more digestibly - it seems safe to say that this will ultimately be good for everybody. And that's the kind of thing that creating a 3d graphical environment, as the basis of our interface, the basis of our user experience, will really facilitate.

It may still be difficult to imagine how this shared graphical environment would work, and how it could lead to such far-reaching consequences as we've suggested. A crucial part of it is that there will be a kind of bootstrapping process. The far-reaching consequences will result from the cooperative participation of all of us, as users of the new interfaces. The job of creating the interface is the job of facilitating the bootstrapping - programming a programmable, massively shared, environment.

This graphical interface scenario may be very relevant to our anticipated artificial intelligence scenarios. Creating more intelligent ways of connecting these existing biological intelligences (ourselves), to each other and to software and information, could well help us congeal into a more effective, coherent global mind, which would begin to have many of the capabilities that we expect artificial intelligence will have.

It seems reasonable and interesting to imagine that many existing artificial intelligence projects, deep learning projects, mathematical models of intelligence, etc., could be readily applied to the task of developing, maintaining, and evolving the filters in our graphical interfaces - in other words the task of deciding how to populate our screens out of the vastness of what's available.

In 1939 Leó Szilárd and Albert Einstein wrote a letter to Franklin Roosevelt, regarding the serious possibility "that the element uranium may be turned into a new and important source of energy in the immediate future." In both that situation and our situation now, in relation to these hypothetically forthcoming interfaces, work was/is required to make the possibility into a reality, but some of the immense implications could/can be foreseen.

This could be a very ripe time, with much growth & development coming to fruition. Often we seem to accumulate information, experience, data, over relatively long periods, and then in relatively short bursts, our knowledge bears fruit in the form of solutions, artifacts, inventions, paradigm shifts.

Our minds themselves can be considered a major frontier for exploration, discovery, and invention. Our inner experiences encompass vast spaces of associations, synchronicity, creativity, etc. New media that give us better ways of expressing our inner experiences will clearly have enormous benefits.

I wonder if the time is so ripe for these new interfaces (and for other new developments of equal or greater profundity) that the relative or marginal utility of many of our current media and interfaces is significantly declining for many of us. In other words, either starting now or starting some time pretty soon, for at least some of us, our time may be better spent developing the new interfaces (and doing various other things) rather than continuing to engage much through some of our currently most-used media.

As we (hopefully) work toward graphical interfaces (and I'd be happy to talk with anyone who might like to collaborate on it), let's remember that of course there are other important tasks to be attended to also. Let's keep attending to the personal, spiritual, and social tasks and changes that improve life for everyone. Let's help those who are have been, and/or are being, hurt in the current way of things. And let's be open to major new ideas, significantly new ways of doing things. All of these seem to be significant aspects of a worthwhile enterprise, what has long been referred to as the Great Work.

My previous post on graphical internet interfaces (May 2016): From Agar.io to GGODD (global graphical online direct democracy)