vim ai/impacts: Writing doesn't have a lock on mediating our thinking

Using writing to mediate thought, as is the norm today in schools, habits like journaling, or blogging, is not a timeless technological practice. It has a history. Which means it may or may not have a future.

One defense of human writing, amidst the onslaught of a whole lot of AI-assisted writing that is barreling towards us, is that human writing matters when it serves as a mechanism for thought. We can take this for granted easily, in the late-stage phase of blogging where everyone can be a creator of textual content. Indeed, it is a form of thought. Even if no one reads this, then it was helpful for me to write it, because it captures a moment of thought and helped me organize it in a way that connected it to all the other things I'm thinking about.

But using writing as a a primary means for thinking through things is not natural, a foregone conclusion, or inevitable. It is not even something likely to be valued generally into the future. In that sense, I wonder whether the current defense of education and, more specifically, writing in an educational context is doomed to fail, not because of where generated content is, but because of where these technologies will soon be.

If I can make characters, movies, story, and the like from hacked together generated component, all within an hour or so, will that become my mode of thought?

There is ample evidence that something like PowerPoint impacts thought. The bulletpointization of information leads us not only to expect information in that bite-sized chunk but also becomes a creative way to think through things. (often abused, it must be said, by assignments that allow students to create slide presentations as products for learning) Will AI content be any different? Is there really anything special about writing?

Historically speaking, no not really. It's worth recalling that for thousands of years before writing, public speaking was a mode of thinking through things in an organized way. The specific codification of rhetoric in the Greco-Roman tradition is relevant here, but practices of debate, community decision-making, storytelling, and public speaking across a wider swath of societies and peoples make the point more effectively. Speech was the de facto tool for conveying thought. Writing is but one option among many for organizing thoughts. It has distinct advantages, particularly in the serialization of time and in its searchability after the fact. But recording technologies have for a while been chipping away at those advantages. Other benefits have crept in (e.g. small size of data when stored digitally) with further technological changes. But ultimately there is nothing special about the fact that writing, like any technology, is a tool that helps us think and organize our thoughts.

The emerging generation of AI creation tools look now very much like writing did at its beginnings. A niche product perhaps, with some obvious limitations. It was something inauthentic, a fake voice compared to the default reality that everyone knew. There's no definite endpoint here, but thinking through the affordances of writing (durability, recall, the ability to put things in sequence and compare multiple sources) came quickly enough. Indeed, the key feature of early writing that matters is its ability to index and update and, moreover, communicate that list. There was an oral technology for list keeping too, but one that could be eclipsed by record keeping on durable material. As a tool of thought, that is writing's initial play.

The converging technologies of generative AI have different superpowers. They seem now to be more in the space of the imaginary, but I don't actually think that's the important part. They are highly sophisticated indexes, supercharged in ways that make a list look like a pile of sticks compared to a modern skyscraper. But ultimately that is all they are doing, taking in large amounts of patterns and spitting out patterns with varying degrees of fidelity or adjustment and which we humans find meaningful as derivative or novel or somewhere in between. They are tools for organizing thought and memory and showing it to us again. As such, it will not be very long before anyone using these tools finds them more effective for many domains of thought than other tools.

There are limiting factors -- computation, resources, cost. and I am never an advocate for the idea of big replacement. Few technologies ever disappear completely. Rather, their niche might get more restricted and people might find different kinds of meaning in working with those technologies. Current retro tech is a good field upon which to think of those questions. Why do young people pick up old cameras, turntables, typewriters, or simply hand-write in journals when there are digital analogs that are functionally "better"? It's not just nostalgia or romanticism. There's something about the way that such activities, shifted in meaning in light of current tech, fosters certain kinds of thinking and interaction.

We can't defend the value of writing by claiming it is something special. Quite the opposite, the fact that it is one of many options means that we can value it more in light of the fact that its function will not remain stable. It will not remain a default mode of thought. It may become a special mode of thought, one that has the added barrier of not just showing people, one that requires people to imagine, one that relies in large part upon the complicity of the reader in order to perform that imperfect act of mind-reading between writer and some reader.