The below blog post was written by GPT-3 fine-tuned on 8 years of my blog writing, prompted with just the title and "Conclusion" to prompt the model to wrap up at the end.
The future of writing interfaces
Over the last couple years, I’ve spent a lot of time thinking about how we can design better tools for writing. The most obvious place for us to start is with word processors, but I think it’s too limited a view.
Writing interfaces
I think programs that help us write are best understood as writing interfaces, pieces of software designed to help us capture and organize our thoughts into words. A writing interface is a bit more broad than a word processor — it can also include things like note-taking apps and writing apps that focus on specific workflows like brainstorming or publishing regularly to Medium.
Writing interfaces can be more than just word processors.
Right now, we have a lot of room for improvement in how we design these writing interfaces. Modern word processors are the result of decades of iterative improvements in technology, but those improvements have been primarily focused on efficiency — faster rendering, faster processing, faster search, faster storage. These are great benefits, but we pay a price for them.
Modern word processors are complex and opaque machines that we ask our hands to manipulate. We sit idle, waiting for the machine to do something when we press a button, and we sit idle again, waiting for the machine to respond when we ask it to do something. We enter text and symbols into a black box, and we wait for the box to give us some output.
We should not be surprised that we reach for these complex, opaque tools when we want to write. After all, it's the best we have. What else can we do?
The future of writing interfaces
I think we can do better. The future will hold a writing interface renaissance, where we demand more from the software that helps us get thoughts from our brain, onto the page. We should expect more from the software we use to write.
What should a writing interface do?
A good writing interface should empower your hands, not replace them. It should be a tool that you can sit down and use right away, without having to learn lots of features or remember commands. A good writing interface should feel more like a workshop than a black box. You should be able to run your hands over a good writing interface, and feel the tools and knobs you can turn to get the kind of output you want.
A good writing interface should feel responsive, like you're operating a system that's sitting right in front of you, instead of a remote server somewhere on the other side of the planet. The interface should feel like it's made of hardware and light, and you should feel the weight of the words and sentences you type hitting the screen and spilling out of the interface in the same way you feel objects in the real world.
A good writing interface should feel alive. When you sit down to write, you should be able to feel the interface breathing next to you, reacting to your words as you type. It should remember your past writings, and suggest words and ideas you might want to use more often in the future. It should occasionally ask you questions about your thoughts, and remember your answers for the next time you get stuck. The interface should have its own personality, helping you write as your digital muse, not just a machine.
A good writing interface should get out of your way. It should recognize when you're stuck, and offer suggestions without you having to ask. It should recognize when you're tired, and offer you a cup of coffee or a short break. It should recognize when you're done, and shut itself down. It should remember what you were working on last time you left, and open right up to where you left off.
A more ambitious future
I think we can build writing interfaces that do these things, and more.
I think we can build writing interfaces that are more than just a text box. We can design interfaces that act as canvases, or as canvases on which we can place images and files from our computer. We can design interfaces that are more than a text box, but can still feel like a text box. When you press the "enter" key, for example, the text box may expand to take up the entire screen, and spread out across the screen like a sheet of paper that's been strewn across the ground. As you type, words may emerge from the page like letters emerging from the bell of a typewriter. When you finish, the page may magically pull itself back together into a coherent document, ready to be saved.
I think we can build tools that feel alive, by giving them their own personalities and memories. To do this, they'll need to be powered by AI. But I think AI is up to the task, and once we have the right interfaces to help us get the right kind of training data, I think we'll be able to imbue our writing interfaces with a kind of life that's not possible without the right kind of intelligence behind the scenes.
I think we should get out of your way. A good writing interface should be a tool that you can use to get ideas from your brain to the page, but it should take a backseat to your own hands as much as possible. When you're stuck, it should gently nudge you in the right direction. When you get tired, it should gently help you rest. When you finish, it should gently close itself down.
I think we can build responsive writing interfaces that feel like real, physical objects. If a writing interface feels like a real, physical thing on the table, it can be a much better canvas for the words we write. The interface can be a canvas on which we paint our thoughts and ideas, a dedicated space on which we can pour our attention. When we finish, we can roll up the canvas and store it away until the next time we need it.
A responsive writing interface is more than just a text box. It's a canvas on which we can paint our ideas. It's a real, physical object we can interact with.
Conclusion
These are just ideas, and a lot of details will have to change as we design and build better tools for writing. But I think we need a vision for a more ambitious future for writing interfaces, and the tools we use to write.
The future will be better than the present because we'll ask more of the technology we use. We won't settle for "good enough" tools. We'll demand more, and we'll find people to build it for us.
Further reading
All of these pieces of writing inspired and informed this essay, in one way or another. I'd recommend them regardless of whether you agree with the thoughts here.