That's definitely true for me. My journey as a writer also follow steps you mention, maybe not in the same order: write fun stuff without worrying -> start to worry and write less -> realize that a crappy first draft can end up in a drawer or your hard drive -> write crappy first drafts so that you have clear ideas for the rewrite of the actual post.

As for why writing is a superpower, I have another hypothesis, in addition to the working memory: writing forces you to commit your ideas to a specific form. While they're in your head, it's really easy to adapt them whenever something slightly wrong appears. Same in a conversation: you can explain away, or in a different way, so that in the end it seems the other one has understood you. But paper is ruthless, and writing something down means you can't get away with changing it without a thought. If it's wrong, you need to change the words. And that forces you to go further intellectually.

100% Agree.  

Similarly, I often force myself to write summaries of books I have read.  I often "feel" like I learned something even though I can't articulate what I learned or what I would do with that knowledge.  Summaries help ensure that I'm not just taking that feeling of learning something at face value.

It sounds like you don't keep an idea book or dairy or anything similar. I'm totally on board with the idea that lowering the bar for publishing posts could actually increase average quality, but another thing you could try to get yourself out of the catch-22 is to keep something like that -- if you know you're writing for yourself, maybe the monkey in the brain would not be so tempted to think it's pointless if you don't publish. (Keeping it in a handwritten notebook or something could help tell the monkey that you're not making a blog post draft, since it would not be easy to copy/paste into LW.)

This could also help inspire more ideas for blog posts.

For some reason "writing to think" never worked well for me. I can only figure out stuff with nonverbal thinking and imagination, then try to put it in words.

I'd like to be able to look through my list of posts and feel content that each and every one is something that I put into the world because I am really proud of it and it deserves to be there, but that mindset just leads me to the catch-22.

Another reason to be less strict about quality before publishing: you're not a perfect judge of the quality of your own work.  Sometimes your writing is better than you think it is, and filtering too hard means that some good writing won't be published.  If you don't lose any of your bets, you're not taking enough risks.

Don't cringe when Looking at old Work

Actually, I think it leads to a second catch-22 as well. When I look back at my old posts, I'm horrified by a lot of them, despite the fact that I tried to hold myself to this high standard for publishing.

You know, the funny thing is that to get good at writing, writing a lot is actually a good strategy. Holding yourself to a very high can be actively harmful to the goal of writing a lot.

(I don't expect that you need this advice anymore, but maybe somebody else sees it.)

I think one way to frame this productively might be to have the realization that it is necessary to write a bunch of bad blog posts, in order to get good. Each of these bad blog posts, that you can look back upon, should be seen as a success. They were a necessary step along the way. It might not be apparent how you got better at writing from one blog post to the next, but ultimately the skill that you have reached now is the result of very many of these (at least most of the time) tiny steps.

A Stream model of Content Publishing

Writing to think makes sense. But what if the end result still turns out crappy? What if it's meh? What if it's good but not great? Should you publish it to the world? I'm someone who leans towards saying no. I like to make sure it's pretty refined and high quality.

But that leads me to a catch-22: most thoughts I want to explore don't seem promising enough where I'd end up publishing them. Or, rather, they usually seem like they'd take way too much time to refine. And if I'm not going to publish them, well, why write them up in the first place?

I feel like there is a conceptually (but possibly technically challenging) solution to this. You don't want to push away readers by writing bad posts, and you do not want them to update towards you being dumb. You also don't want to push away readers by writing about topics that they are not interested in. This would likely happen by default if you write about a wide range of topics.

IIRC somewhere in the Arbital postmortem [LW · GW] there is a simple solution to this. Instead of having only one channel to publish your content, create multiple ones.

You could devise a system where you can assign a tag to a post. Readers can then decide for which tags they want to get notifications. This also gives you dynamic filtering of posts. I expect that to be good for exploring the content, e.g. finding older posts that you are likely interested in.

Additionally, you could, for each post, attach the amount of time that you have invested into the post per word written. This is probably a reasonable proxy for quality.

You could also sort each post into 3 bins based on how good you think that post turned out. Or have categories that classify how exploratory some piece of writing is. Is this writing the result of you trying to get a better understanding of something, or does the writing aim to provide the best possible explanation for an important topic that you understand well?

And I am sure if there is a lot more that you could do. It would probably be important to have a few good default configurations that people can choose between, to not overwhelm them with the available options. E.g. there could be one stream that is about topics you expect LessWrongers to be interested in, from the high-quality post tag category.

Re writing for mental health, the Self Authoring program does this. Essentially it makes you spend quite a few hours simply writing about your past, present, and future under various headings. Apparently this is clinically effective at fixing various mental health issues.

(I tried it, and though I don't have mental health issues, it provided many insights into myself and I can see how others would find it very useful.)

Thanks for writing this post! I agree with much of what you said and want to add a few thoughts.

I think writing is so powerful because of how it functions as a tool and shapes / impacts those who utilize it, plus, the act of writing generates a product, an object in reality that can be interacted with by oneself and/or others: writing is a cognitive prosthetic that serves as the best method for the recursive improvement of human thoughts over time!

From Cary Wolfe's "What is Posthumanism": "In Luhmann as in Derrida, writing takes center stage as the paradigm of communication, but only because it exemplifies a deeper “trace” structure (the grammè of the program, as it were) of meaning— a paradigm whose essential logic is for Luhmann only intensified by the sorts of later technical developments, beginning with printing, in which we have already seen Derrida himself keenly interested in texts like Without Alibi and Archive Fever. In this light, the problem with “oral speech,” as Luhmann describes it, is that it threatens to collapse the difference between information and utterance, performatively subordi- nating information to utterance and presuming their simultaneity— “leaving literally no time for doubt,” as Luhmann puts it40—in precisely the manner analyzed in Derrida’s early critique of the subordination of writing to speaking. But if the value of language is that it is “the medium that increases the understandability of communication be- yond the sphere of perception” (160), then writing is its full realization. “Only writing,” Luhmann observes, “enforces the clear distinction between information and utterance,” and “only writing and printing suggest communicative processes that react, not to the unity of, but to the difference between utterance and information. . . . Writing and printing enforce an experience of the difference that constitutes com- munication: they are, in this precise sense, more communicative forms of communication” (162–63)."

I think the gist of that quote, with a bit more context added and some of those thoughts translated into LW terms, is that:

(1) There are maps and maps are not the territory.

(2) Humans use language to make communicable maps of the territory. The "trace" or grammè are "deeper" non-language maps that can be extracted from language as "felt" or intuited maps of the territory that capture more of it than writing does on its own. You might think of trace and/or grammè as "the underlying meaning" or "analyzed or intuited meaning" that can typically be extracted from written text if one applies a bit of thinking against said text.

(3) Communication involves both information and the transmission, or, utterance, of said information.

(4) Writing and the published word are the most powerful and thus, most communicable forms of communication because other language mediums collapse the difference between information and its utterance or lose too much of the deeper structures (i.e. lost the trace or grammè) whereas writing preserves those structures the most and explicitly preserves the difference between information and its utterance.

(5) To be able to reflect on, apply thought to, etc. communications, one must be able to separate information and utterance, with more deeper thought only possible if there's a very clear separation between information and utterance in whatever communications one is receiving. As Lumann says in that quote, oral speech doesn't separate those two things and in fact often totally collapses the difference between them, thus making longer-term access to the information and the deeper structures communicated not possible via oral speech. Writing on the other hand, preserves such things nicely, and allows for unlimited time for reflection and thought to be applied against them, provided the written text survives. Thus, the written and published word is the most communicative form of communication, the only form of communication that explicitly preserves for up to unlimited amounts of time (we have books from 2000 years ago, e.g.) the possibility of thinking about what was communicated.

When you have up to an unlimited amount of time to reflect on, think about, etc. ideas or thoughts, and can channel your thoughts into the written and published word where yourself later on and/or others can apply reflection and thinking on what you had produced, well, now you have a recursive pattern of improvement where each product of thought can generate through human thought applied to it / against it, a new and better product of thought!

Thus, writing is truly humanity's best tool for improving thinking, because it's a cognitive prosthetic allowing for unlimited recursive improvement of human thought over time.

I believe that recorded oral speech via music or videos is now allowing other language mediums to have more power and grant access to similar kinds of recursive improvement, but writing still seems to be in a category uniquely its own for how powerful it is. For now, anyways.

I'm very curious to see how improvements in image + video searching, music search, improvements in VR tech, etc. might improve on those mediums ability to share in writing's power.

Notes: my thoughts on writing are currently very influenced by Cary Wolfe's "What is Posthumanism", but I've been very influenced by reading many other different products of thought targeted at writing over the years such as Umberto Eco's "How to Write a Thesis", hundred of blog posts about writing, A.G. Sertillanges' "The Intellectual Life: Its Spirit, Conditions, Methods", casual discussions with friends about writing, Cyril Connolly's "Enemies of Promise" (this book broke my ability to write for a semester in college, such was its power), and more.