One (imperfect) way of understanding the future of anything is to understand it at a meta level, by focusing on the salient trends. I recently thought about computing and how the introduction of new and new layers of abstraction is transforming it.

Consider what it was like to the generations before us. Early on, computer scientists and engineers were meticulously perforating and feeding tapes to machines that acted as glorified calculators — something the generation after them could take for granted. The questions that the machine could answer were limited in domain and most of the problem construction and solution happened outside of the computer. The computer was–one could argue–in its mathematically purest form, a Turing machine, a reading device moving over some tape. The device being faster than us and the tape storing more information than we could memorize, the computer merely supplanted our memory and made mathematicians into engineers (i.e. people who could add immediate value).

The introduction of rich user interface (and, I’d argue, the very much related introduction of programming languages, a powerful layer of abstraction on top of the simple computational box) ushered in a still-alive-and-well era of computers as aiding us in decision making. I still interview people who worked on wireframes in large companies, meticulously modeling behavior in Fortran. More and more of models were captured by the computer, who now enhanced our ability to visualize.

Connectivity (and, as a consequence of it, mobility) was the next big step. Today it’s something our generation takes for granted, so it may not seem as so revolutionary, but networked computers allowed information to flow more freely and be available anywhere. Just as humans achieve extraordinary things through interacting with other humans, computers

There are two dimensions of change here — one is the growth of each of the above concepts. When considering the future of technology, we will want to consider the increased intensity of each. Despite various efforts to debunk it, Moore’s law still holds, which means that computers will be able to give us answers more and more quickly, and hold more and more information. This allows us to solve more and more classes of problems. Similarly, user interfaces have been improving since the 70s and still are, the multi touch interface being the most recent (and shockingly impactful!) addition to the family. User interfaces get more intuitive, bridging the gap between how humans interact with the world and how they interact with machines. Haptic interfaces, and interfaces much more closely taking advantage of our senses (such as sensors attached to our eyes) will continue revolutionizing the field (as an aside, with such a power coming from the layers of abstraction, one must really redefine the term revolutionize to take into account the continuous nature of such a process in computing). Languages will be much less formal and much more expressive, akin to human languages (already a lot of the applications out there are simply large building blocks that are configured to work together). Finally, connectivity will turn into hyperconnectivity. All information will be connected in a kind of semantic web, making all these Hollywood OSs seem trivial.

The second dimension — new concepts that make their way into computing — is fascinating to me. I can only theorize what these concepts may be (and our capacity to see two or more layers above today is, I’d say, nonexistent). One such concepts is one of partnership — while today computers still perform tasks given to them by humans, tomorrow’s computing will be more of a cooperation between the human and the computer. We’re still a tad bit away from this concept, though: note that just as connectivity was made possible with advanced programming languages and better user interfaces and mobility couldn’t be doable without advanced computational power, partnership will require all the existing concepts to be sophisticated enough to support it.

Let me conclude with a bit of science fiction. Search is a difficult problem to solve with quality, but one that provides immense value when solved well: just look at the collective number of degrees its employees hold, and the market cap of Google, respectively. What makes search difficult is that so far we have to find good proxies for what makes one bit of information relevant to another.

In the future, we will be able to organize information in a much more intuitive way, semantically: that is, in a way that is much easier to describe connections between one bit of information and another. Our brains work the same way so searching for information will be easier as all the relevant supporting information will be easily extracted. We will be able to traverse trees of knowledge, but also ask questions that have to do with connections between concepts and not just the concepts themselves. For example, suppose that I’m looking for a good seafood restaurant in New York City. Today me best bet are specially designed websites that organize the restaurant information in a way that’s easy for me to search, or Google’s ability to find information that matches the words “seafood”, “restaurant” and “new york city”. Maybe we can do a little better with social networking and crowdsourcing (I’m sure many people search for good seafood restaurants in NYC). This works pretty well but there are still inaccuracies: a custom restaurant search website must have existed, Google may be stupid (for example, it may bring up diners that happen to offer some seafood entrees). But if we could find information related to the concept of restaurant and cross this with information related to the concept of seafood and the concept of New York City, we would get much more relevant information. Even better, further refinement would easily be possible (maybe I actually really just mean Manhattan, or restaurants within 20-block radius from my apartment, etc.).

Let’s not stop there. We’ll pretty quickly realize we’re limited by our interface to the computer. If the computer could read the information coming from our brain, new kinds of queries would be possible because they would no longer have to have been pre-programmed. In that way, the computer could learn the model of what we are looking for simply by matching the “mental signature” of our query with its model of the world (to get an answer to “find me a seafood restaurant in the City that’s 20 blocks from my apartment that has entrees I will like”). If we go a little further (yes, no doubt into freakishly scary territory), and allow for bidirectional communication, the computer could actually help us figure out what it is we’re looking for even if we’re not sure of it (“What do I feel like today?”) by interacting with our mental model: sending signals to it and seeing how we respond; I like to think of it–in gross simplification–as a kind of mental “20 questions” game.

One more crazy step — if we connect all such computer-brain interaction systems together (let’s skip all the dangers for now), we have an incredibly powerful collective system, a kind of Borg. Just imagine what kind of questions will be possible in such a system! Following with our example, it could be something like “find me something to do in the City tonight so that my friends and I will have a good time.” The system would simply connect my mind and my friends’ minds to get the answer, understanding our models of “fun” and maybe playing out some scenarios to see how they would resonate. Maybe have three candidates for activities, simulate an entire experience for each of the three, pick the one that maximizes the collective utility, and then wipe out our memories so we can actually experience the activity in reality.

If that feels familiar, you are right, this is the Matrix. Note that as soon as the computer can interact with our minds, it can fool us into thinking anything or remembering anything. We don’t need to build a sophisticated model of the world; we just need to tickle the right part of the brain (provide the right “sensory signature” to the brain) to make us think we’re in a seafood restaurant. Computing in this stage will, as a result, be indistinguishable from the real world, and computers will be gods, capable of generating a plethora of universes for us to live in.