Mark Weiser: The World Is Not a Desktop

Our computers should be like our childhood: an invisible foundation that is quickly forgotten but always with us, and effortlessly used throughout our lives. - Mark Weiser ^[1]

Mark Weiser wrote the above quote back in 1993 in his amazing and short paper, The World is Not a Desktop. Throughout his career ^[2], Weiser explored many areas around what a computer is, what it should be, where it should be, and how it should behave. (Likely this newsletter will come back to his work again and again.) In this concise editorial, Weiser takes our generic understanding of computers and, in my opinion, blows it out of the water.

What is the metaphor for the computer of the future? The intelligent agent? The television (multimedia)? The 3-D graphics world (virtual reality)? The StarTrek ubiquitous voice computer? The GUI desktop, honed and refined? The machine that magically grants our wishes? I think the right answer is "none of the above", because I think all of these concepts share a basic flaw: they make the computer visible. ^[1:1]

Over the past few decades, computers have gone from being relatively rare, such as mainframes lurking in the basements of bank headquarters, to being ubiquitous, including, of course, smartphones. Many households will have dozens of computers. We have succeeded in making computers fast, cheap, and plentiful—not for everyone but for many—yet we have not made them invisible; in fact, quite the opposite. These numerous computers are right in our faces and close to us all the time. They are the antithesis of invisible.

Weiser argues that "a good tool is an invisible tool," ^[1:2] and he illustrates this with the example of glasses, or a cane used by someone with limited vision. There are things that we don't often think of as "technology." Everyday items such as glasses had to be invented, as did computers. However, they're not overwhelming; we look through the glasses to see the world; they're not opaque; they're on our face, but not in our face. The reality is that the way we use computers and the "metaphors" we give them end up making them more visible, not less. What he is talking about is not necessarily their physical structure, but the visibility and invisibility they take on in use and how they become part of the fabric of our daily lives, as opposed to being an intrusive presence.

Weiser takes on a number of technologies in the paper, exploring their value and why we try to use them the way we do.

• Multimedia

One perspective we seem to take is that computers should have visible interfaces, and those visible interfaces should be attractive; they should be capable of displaying information using various forms of media. We may have new form factors for computers other than the desktop and laptop, in that we have tablets and smartphones, but we still use them largely in the same way we have always used computers: we have a mouse (now a touchscreen), a keyboard (now virtual and predictive), and a screen, perhaps smaller, but nonetheless a screen. When we use these computers, whatever their shape or size, we still launch applications that open "windows" (which aren't windows at all, because we can't see through them like we can through glasses). So we spend considerable energy trying to make interfaces look better, but "attractiveness is the opposite of invisibility." ^[1:3]

• Magic

I'm as guilty as anyone in terms of hoping that computers can become magical. However, Weiser has a different take and says sardonically, "I wish my computer would show me what I am interested in." ^[3] Thirty years after this paper was published, there are a number of ways to interpret this. One is algorithmic content curation, which most of us will be familiar with if we use tools and sites like Youtube and Twitter (now, oddly enough, X.com), through which a (usually) secretive computer program tries to show us what we should see, based on our content history (not necessarily what we want to see) and in a way that can monetize the product. In this regard, Weiser's sarcasm is certainly justified. The second is that with tools like generative artificial intelligence (GenAI), we might be able to do the same thing but better: get computers to "magically" show us the things we need to see. However, in attempting to read between the lines of The World Is Not a Desktop it seems that Weiser would object to GenAI an AI attempting to make computing "magic."

• Virtual Reality (and Augmented Reality)

...virtual reality is only a map, not a territory. ^[3:1]

Weiser is particularly hard on VR, and I think rightly so. The problem we have today is that computers are too pushy, and in many ways, VR is the most invasive interface we can have with computers, bringing our body out of the real world and into so-called "cyberspace." While it's useful in art, like novels and films, to represent what is effectively the inside of a computer, because otherwise it is difficult to visially explain, it's only an extension of the "interface," now taking up as many senses as possible. If computers are to become invisible, then we certainly don't need to take the interfaces we have now and double down on them.

• Intelligent Agents

Why should a computer be anything like a human being? Are airplanes like birds, typewriters like pens, alphabets like mouths, cars like horses? ^[1:4]

Weiser makes an interesting point here, especially in the context of recent advances in AI and GenAI, where, for example, we can now converse with computers in natural language. However, as we have discovered, tools like GenAI can feel alien, and they are clearly not human intelligence as we know it, and yet...we badly want them to be. Ultimately, what we have today is token prediction, but our goals are to create "intelligence" and, presumably, "intelligent agents."

My interpretation of Weiser's main point in terms of intelligent agents is that we want human intelligence embedded in our computing systems so that they become our interface with them. However, Weiser asks whether this is a desirable goal at all: "Are human interactions so free of trouble, misunderstanding, and ambiguity that they represent a desirable computer interface goal?" While I do not believe that GenAI is intelligent, it already shows some of the issues that humans have in our own interactions, in that GenAI occasionally lies or makes things up and does so in a confident way that makes it feel like a kind of trickster—perhaps the way an intelligent agent, should we be able to build one, would behave.

Conclusion

This is a brief introduction to some of Weiser's writing. What I appreciate about his work is that he asks the kinds of questions that we seem to have forgotten to ask. We assume that computers and their interfaces need to look a certain way and that they should be more meddlesome. What's more, when we consider our ambitious goals for computers and AI, the designs we come up with end up looking like human intelligence, which may not be how we want them to act and perform, so we have to be careful what we wish for.

The idea that computers could become a distant memory, like our childhoods, and be recalled when necessary is striking—so much so that it is hard to imagine a time when it would be remotely possible. However, computers are still in their infancy, with the science only a few decades old.

There are some common themes to which this newsletter will return:

• Electronic computers are a relatively recent invention, and we are still discovering their full potential and embarking on a grand journey ahead.
• That computers should be invisible, not constantly in our faces.
• Using technology solely for the sake of technology is not always the most effective approach. As we continue to improve our technological capabilities, we will discover that reevaluating the socio-technical structure is one of the most effective ways to advance our work.
• That general-purpose computers can run any program, and we tend to try to get computers into the right shape, the right form, rather than using them in the right way; perhaps the form doesn't matter at all.

---

1. Weiser, M. (1993). The World is not a Desktop. https://www.dgsiegel.net/files/refs/Weiser%20-%20The%20World%20is%20not%20a%20Desktop.pdf ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

2. Unfortunately Weiser passed away in 1999. ↩︎

3. Weiser, M. (1991). The Computer for the 21 st Century. Scientific American, 265(3), 94–105. http://www.jstor.org/stable/24938718 ↩︎ ↩︎