TODOs before publishing
  • "Transpose" the matrix. Maybe start with a quick 1-5 rundown, but then do a deep dive on each rather than constant context switching.
  • Little demos of the specific features
  • Diff from previous versions. Scroll wheel?
  • Comment + credit
  • Combine "Revision" and "Continuous iteration"... former is a subset of the latter
  • Robert Caro example of editor continuity. "Gottlieb ... has continued to edit all of Caro’s books, even after officially leaving the company."


# I. Writing and the scientific method
"Science is a way of thinking more than it is a body of knowledge," said Carl Sagan.

Writing is similar. It is a process, not a position. It is as much about challenging and polishing an idea as it is about the idea itself, and it's not a solitary activity.

Unfortunately, this process is misunderstood and misrepresented by the tools available to us right now. Current publishing methods undermine several important aspects of writing:

1) Revision: Almost everything I publish is radically different from its original draft. Whenever I write something, I'm suspicious when my view doesn't change at least a little before I share it with the world. Discussion, feedback, and reflection mark the most valuable stage of the writing cycle. They might uncover a neglected argument, produce a stronger rhetorical structure, or cause me to question my assumptions all together. This journey is interesting in its own right!

2) Credit: None of my writing is truly "mine alone". Dozens of people have participated in this blog in one way or another. Some have contributed so much to my thinking that I've invited them to be listed as a co-author, though none have taken me up on it yet. This is just the tip of the iceberg and doesn't include the memos I write with coworkers at work, the essays edited by my teachers in school, and the articles I wrote for my college newspaper that the editors spent hours helping me refine.

3) Reach & perspective: Many people influence my thinking, but that reach is limited to the group of individuals I know personally. Some of the most valuable input comes from my friends who are most different from me. They influence my writing process in ways I don't expect, because their background and context and expertise is different from mine. I would love to incorporate perspectives beyond my "natural" social circle. I've had some success meeting more varying people through Twitter, and in some cases I've invited them to be part of the writing process, but it isn't a natural thing to do.

4) Critical reading: The "audience" during the editing process is more engaged than the audience after the post is published. This is because an editor isn't a passive reader but rather an active participant. When friends invite me to give feedback on their writing, I engage with the material in a far deeper way than most things I read, because I'm a partner in its development.

5) Continuous evolution: Writing is just a snapshot in time of the author's thoughts, and it can get out of date as they learn more. Some of the biggest updates I've had in my worldview occurred after (and in some cases because!) I published an essay on the topic.

Unfortunately, these dimensions of the writing process is hidden from the reader. We only see the result. My drafts are encrusted with comments from friends, and it's at least as interesting to see how these conversations, questions, and disagreements inform the concept as it is to read the published piece itself. But in the end, they are resolved into oblivion, and the final version contains no trace of the commentary.




The studies for Michelangelo's Creation of Adam are as fascinating to me as the final product. So interesting to see how he developed familiarity for human physiology before tackling the final painting.


# II. The current state of publishing tools
A lot of work happens before the writer hits "publish", and it continues afterwards too, as the author continues to learn about the subject and refine their understanding. However, that process is invisible. We lack visibility into each of the 5 categories mentioned before:

1) Revision: Wisdom is not just a series of atomic conclusions. It is something closer to an algorithm, a process through which a person makes sense of the world. If you wish to learn from someone, you should understand how they got to the result.

Right now, most writing leaves readers only with the results of the author's thought process rather than the understanding of how they got there. If the reader doesn't agree with that conclusion, they’ll just shut it out and never understand how the author arrived there. If they do agree, they’ll only pay attention to the fact goals are shared and not look more deeply into the assumptions behind the author’s views or their own.

I've yet to find a tool that treats the revision phase of writing with due respect. Google Docs comes the closest, but comments feel perilously ephemeral, like they might disappear at any moment. I'd like comments to stick around, as an artifact alongside the essay itself. This would would allow future readers to follow along with your thought process.

This could look something like a "GitHub for writing". One of my favorite aspects of the Pull Request model is that it leaves a paper trail of comments that allow future codebase archeologists to understand why decisions were made. I discovered the hard way that Git is not the right model for prose, but a lot of ideas could be borrowed from how Pull Requests work.

"We know our software sucks. But it's shipping! Next time we'll do better, but even then it will be shitty. The only software that's perfect is one you're dreaming about. Real software crashes, loses data, is hard to learn and hard to use. But it's a process. We'll make it less shitty. Just watch!"
— Dave Winer, We Make Shitty Software

2) Credit: My writing is a potpourri of others' inspiration and input, but the byline does not reflect this teamwork. I always thank them at the bottom of a post, but that gratitude rings hollow. Few readers see it, and none know what the thanked person did.

Why does credit matter? There is a cost (in time and effort) to editing, and there is little reward despite its immense value. We might have nice friends who are willing to take a look and sometimes even give serious feedback to our writing, but they’re doing us a favor, and it’s hard to systemize it. In short, editors are not appropriately incentivized. This results in a shortage that's observable along several dimensions:
  • Volume: There are just not enough people in the "market". There is more need for good editors than there is supply. Of course there is an imbalance—writers get all of the credit while reviewers get close to none!
  • Continuity: It feels rude to ask even the most generous friends to revisit your piece again and again. But if you ever talk with a great author about their relationship with their editor, they always cite a strong, continuous relationship as being crucial to the development of their ideas. We’re not all authors of bestselling books, but we could all benefit from a good editor. Giving a reviewer credit could change the ask from a handout to a mutually-beneficial collaboration.
  • Quality: Your friends might glance at the piece, maybe point out a weakness here or there, but they often don’t really dig into it. Even if they want to, they have other priorities, and they have no reason to put this at the top of their list except out of the goodness of their hearts. If they got credit for their contribution, they might be more motivated. I've had the luxury of many friends who give extremely high quality feedback, but I think I am lucky in this case. Even so, I feel guilty asking friends for help when they don't get much out of it, so I throttle myself before they even have a chance to say no. I would feel less guilty asking if they got more out of it.

Books are better than many mediums in this regard, since many have dedications at the beginning. However, these Dedications are often an afterthought, they are isolated from the context, and they are at best a reconstruction of the process through the memory filter of the author. No wonder most readers skip over them. (I do recommend at least skimming dedications chapters, because I have found some real gems about many a writer's process that have later informed my own.)

Academic papers are somewhat stronger on this dimension too, due to a deep culture of citations. Peer review is very broken though, in large part because there's little incentive for reviewers to do a good job. It would be awesome to have some level of context on who reviewers were in academic papers, for the sake of giving them skin in the game.

3) Reach & perspective: My request for help is also limited to a narrow ring of people I'm close to, but I would love to expand that group to people with more dissimilar views. I for one would love to help an interesting stranger revise their writing on a topic I know well. The problem is there is no way to find one another.

How cool would it be to embed the reflections of an expert in your writing or to get the perspective of another student from a continent you've never visited? Right now those people have close to zero reason to help you, but that might change if there were a better system to surface their contributions.

The web is beginning to solve this problem by connecting people who otherwise wouldn't meet. I want to take this one step further and connect people who otherwise wouldn't meet and who can help one another improve their ideas.

4) Critical reading: In traditional publishing, the audience is not invited to participate in the evolution of the author's idea. It is handed to them as an artifact to look at, not to modify or improve. Comments sections fix this somewhat, but I want a tool that really gets into the guts of a thing. Imagine if you could open a pull request on a New York Times column or open an issue with your favorite author!

5) Continuous evolution: The current publishing paradigm presupposes that a book or an article is "done" when it is published. If there is a particularly egregious error, a newspaper will add an errata section to the back pages of the next issue. That is about the extent to which a piece of writing is expected to evolve.

This is a shame, because it leads us to regard an author's thoughts as written in stone, something that won't change. However, that could not (or at least should not) be farther from the truth! We are all learning all of the time, and our ideas evolve with it.

"We should think of the details of our political views as analogous to betting on a slightly crooked roulette wheel, designed to land on the number seven more than a proportionate amount of the time… We should choose the course that is most likely to be correct, but at the end of the day we are more likely wrong than right."
— Tyler Cowen

Instead of treating a piece of writing as something that is immutable, we could look at a sequence of versions as snapshots in time. A reader could scan backwards and forwards on that timeline to see how the author's views have evolved. This would have a few major benefits:
  • It would enable readers to understand how, when, and why the author changed their interpretation and articulation of the concept.
  • It would encourage a more realistic view of epistemics, one that emphasizes that great thinking is a process of iteration rather than a lightbulb moment. This would make more people feel like writing and learning in public is something that they can participate in too, in an active way rather than just as a passive consumer of information.
  • It would reward the author for epistemic honesty rather than appear as a mark of shame, which Errata sections of newspapers. Right now admitting that you had something wrong is taboo; we should celebrate that and make it easier.

# III. Why is process divorced from the final product?
If the writing process is so fertile with insight, why don't we have tools to cultivate and harvest it? My guess is it has to do with the constraints of print media. Again, back to our 5 categories:

1) Revision: When you typeset each individual character in the printing press by hand, you don't have much patience to revise a draft. You're certainly not going to typeset the meta writing that is the revision just for the sake of sharing it with the readers! It's just too darn costly.

2) Credit: You may have been able to list the people who played a role in drafting, revising, and editing a piece, but it would be unreasonable to track each and every contribution they made. Again, it was just too darn costly.

3) Reach & perspective: Before the digital age, it was hard to connect with people outside of your local community. Physical proximity dominated who could interact with one another. Some intellectuals could stay in contact through letters, but mail is slow. This low bandwidth meant that even in the best case scenario, the feedback loop was very loose.

4) Critical thinking: The static nature of a piece of writing extended to a writer's relationship with their audience, too. Readers could only participate by mailing in a letter to the editor. There was limited capacity for a "conversation" between author and audience. As a result, readers couldn't reshape the writing alongside the author, they could only passively consume it.

5) Continuous evolution: When we were restricted to a printing press imposed deadline, we were forced to think in terms of completion, because once the ink was on the paper, that was it. Quotes could be redacted, and authors could disavow past work, but these updates were basically hopeless.

In other words, surfacing the process was impractical, and those constraints calcified into assumptions about what writing was. Technological progress has lifted these constraints, but we have yet to revisit some long-held assumptions about what the writing process should be.

# IV. What now?
It is time for this to change. We're no longer constrained by the shortcomings of print media. We can make changes to our work at virtually no cost, and we can receive readers' input at the speed of a fiber optic network. Back to the 5 categories one last time:

1) Revision: It's now cheap to track changes throughout the process. It requires no extra work on the part of the author to do so. If it's digital, the text editor or publishing tool can manage all of that on the backend without anyone thinking about it.

2) Credit: It is now trivial to trace many forms of contributions to a piece. When someone makes a suggestion or fixes a typo, we can track that easily too and give them the credit they deserve.

3) Reach & perspective: It is now trivial to connect to people across the globe, at least from a technical perspective. There are still deficiencies in the marketplace design, where in practice we aren't seeing these people connecting to collaborate. That said, information technology makes it technically possible, so I see this as an opportunity.

4) Critical thinking: There is no reason readers have to be passive consumers of writing anymore. They can actively participate in a piece as a conversation with the author. We're already starting to see this with Twitter, comments sections, and more, but I'd like to see tools that encourage more constructive cooperation and more granularity. For instance, it would be cool if I could open a Pull Request on a paragraph of your essay, perhaps to recommend how it could be better phrased. This would be much more useful than my just commenting at the bottom of the post "You are so inarticulate, you must be an idiot!".

5) Continuous evolution: We have had version control for code for decades now, and developers have understood for nearly as long that it is not just a productivity tool but a deep way to explore the history of an idea (in this case captured in code). We understand that most software is never truly finished—there are always improvements to be made, extensions to be built, and bugs to be squashed. The same is true for writing. Thoughts are never finished, only shared.

~

I realize this is all quite abstract. This essay is more of a conceptual sketch than a specific plan, but I hope someone takes it as a call-to-action. I have a lot more specific ideas of how these things could work, so if anyone is interested to explore it further I'd love to connect.

Our tools reflect and shape how we approach creative work. We finally have the capability to build tools that encourage continuous improvement, so let's go build new tools!