The Perfect Writing Machine

Recording and sharing text is not a very difficult technical problem compared to many others, and was already solved at the dawn of the Internet and the personal computer. The original Macintosh computer, in 1984, may have been the perfect writing machine, and since then we’re only getting refinements. It met the most important requirements: a simple and efficient way to record text, format it, print it and share it electronically.

It provided a giant leap in utility over the typewriter. First prototyped in the early 1700’s and developed gradually over the next hundred and fifty years, early models (like the one pictured above) were rejected or used only reluctantly by writers such as Nietzsche, who hated it. Once it reached the right level of refinement, however, the typewriter was brought to mass market and has been an essential piece of equipment for writers since.

It was an amazing device. It’s possible to type much more quickly than it is to write longhand. Average to quick typists can turn out words at 90 or 120 per minute, while most people write longhand at around 30 words per minute. Indeed, the typewriter did many things better than the early personal computers: it was portable, it ran without electricity and could never run out of batteries, and it immediately recorded everything on a permanent medium (paper has its fragility, but a power outage or unsaved document has made some hours of work vanish more easily from the RAM of a writer’s PC).

The key limitation of the typewriter is its capacity to store, backup, copy and share literary productions. T.E. Lawrence infamously lost an estimated 250,000 words of the first draft of his “Seven Pillars of Wisdom” when he forgot his briefcase on a train. Paper is a durable medium but as with all analog media, presents obstacles to easy duplication and relies on “master copies” which are vulnerable to physical loss and destruction.

Easy storage, backup and transmission of massive amounts of text were possible even with the earliest personal computers. And today, it’s trivial for a writer to store their entire works on an SD card hardly bigger than a paper clip. Many writers now skip physical media backups entirely, and immediately sync their writings to giant server banks in the cloud, which are redundantly backed up and accessible from anywhere.

Even the energy requirements of digital portable writing machines are closing in on the advantages of the typewriter. New ultralight laptops can run for a full working day on a single battery charge. But the greater advancement has come with e-paper and its ability to passively display text for long periods of time using very little energy. The Kindle and comparable e-readers can carry massive amounts of text, and only need to be recharged once a month or so.

A new kind of device, currently in prototype (see the Hemingwrite) will undoubtedly become a favorite of writers seeking to escape distraction: ultra-low-energy, durable e-ink Word Processors. “Word Processors” (taken to mean single-purpose portable computers rather than the software programs on general-purpose computers that have come to replace them) had a brief heyday in the 1990’s. They are poised to return as e-ink improves and other technologies like sophisticated off-the-shelf components and 3D printing makes such devices easy to manufacture in the limited runs which the niche will demand.

The advantages of a laptop over such a limited-purpose device are obvious. But for a writer, certain limitations (the inability to watch video, for example) are a feature, not a bug. Escaping constant distractions has become an all-out war for those trying to reach deep states of concentration and tap their creativity. A light, hardy device with a passive e-ink display and comfortable keyboard, just powerful enough to send backup transmissions to the cloud or a neighboring device, would be ideal for many writers. They could take it on a remote retreat for weeks without needing a charge, even hike into the backcountry, and when they are writing, focus strictly on composition.

There are new frontiers, of course, in recording and transmitting text, and one that may finally now be coming of age is speech-to-text. The software is still somewhat unwieldy, though it has improved dramatically over recent years. We may yet reach the day that Orwell depicted in 1984, when Winston Smith sits down to write in his illegal diary and finds that “he was not used to writing by hand. Apart from very short notes, it was usual to dictate everything into the speak-write.”

The age of the natural user interface, in which human speech and gestures commonly replace the graphical user interface metaphors we’re currently familiar with, is still only in its infancy. We have the Kinect, we have Siri, we have touch interfaces on the iPad, but for doing real work, we use a mouse and keyboard.

But just as programmers still resort to the command line interface that hides below the surface, I suspect that most writers will want to be ‘closer to the metal’ or in their case ‘closer to the alphabet’ for a very long time. My own experiments with speech-to-text tools have illuminated to me just how different writing is from speech. The importance of commas and periods and quotation marks, which are clumsy to dictate, is a huge hurdle, and there is also the way that keeping ones mind open and mouth shut seems to lend itself to clear thinking.

And yet, if the software becomes subtle enough, using sophisticated algorithms and machine learning to handle punctuation, many writers may increasingly turn to speech-to-text. Dostoyevski found it expedient to dictate “The Gambler” to a secretary, so if that day should come, I doubt it will mean the tragic end of literary creativity. Though it would probably shift literary style, just as the computer as a writing medium has to be at least partially responsible for an age of literary gigantism (check the word lengths of current literary icons like Roberto Bolaño).

For journalists, truly mature speech-to-text software promises to be extremely useful. Anyone who has transcribed a long interview knows what a chore it can be. This labor-saving technology means that the writer can spend more time on editing, and less playing back a quickly spoken mumbled sentence that may get cut in the end anyway.

Further off, more radical and most intriguing, is the possibility of direct transmission of thought to text. This is highly speculative at this point, but not impossible, as a 2015 study published in the journal “Frontiers of Neuroscience” demonstrates. If the neural mechanism that powers speech could be interpreted as signals understood by a computer, you could merely think the words and see them appear.

As ever, only a well-developed thought will turn into text worth reading. Whatever efficiency in literary production gained by such technology will only be culturally valuable in so far as it enables the writer to spend more time refining their craft.

Like what you read? Give Justin Allen a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.