The futility of one million words

I've had a word count tracker for this site for quite some time. It's nothing special, just counts every "word" in posts and in comments by me, defining word as a string of symbols not including a space. A bit of logic filters out quote blocks or code blocks. (I had previously forgotten to filter code blocks, but duplicated the logic for span blocks so that should be good now...) Some quotes/code blocks probably should be counted (I think there might be some cringe poetry in pre tags...) but whatever, accurate enough.

As of today, ignoring this post, there are 482,117 words.

In 2021 I decided to count how many words would be added if I added up my HN comments, which is the other place I write the most at. That gave me another 400,000+ words, naively counted with wc -w, essentially the same logic as for my blog. Though some of those words are also quotes from others, or code, and I didn't try filtering them out. And every comment is separated by a bunch of dashes and wc counts those as words too, adding one extra word per post. In other words, it's an overestimate.

I also added my private journal, same wc method, which in 2021 had almost 50k words.

I've not added twitter, since I don't have that many posts and most of them are crap anyway, plus the character limit limits words anyway. Nor have I added any documents (e.g. google docs or libreOffice docs) I've written, or emails I've written, or code documentation. Nor any of the same of that sort that I did for work. Chat/instant messaging words are also not included, even though there's been a lot of that. No youtube comments. (And youtube deleted them anyway, my backup before the deletion is older than 2021 and they got deleted in 2022 or 2023, I forget, when I migrated my old pre-google account to be tied as my primary google-tied youtube account...)

Well, at some point between 2021 and now, my total words from what I have tracked exceeded one million. Today I decided to update my HN and journal numbers. About 530k and ~80k respectively, so in grand total, 1,150,637 words.

A long time ago I read the advice that to get decent at writing, you should write your first million words as soon as you can. And consider most of them as not very good. This is similar advice to learning board games like chess or go that say something like play (or often, lose) your first 50 games as fast as you can.

Well, one million words later, and... I actually think I might be worse at writing than I was in, say, 2011. So much for the practice.

I do often go over my writing in multiple passes to make edits or reword (sometimes I even look up a word in the dictionary that I'm sure I already know, just to check if I'm using it in a way that conveys my meaning as exactly as I want compared to some other word) or drop things. Sometimes this editing happens right after "posting" too, as I see things differently in the final presentation, though this is mostly on HN where I haven't yet coded up a 'preview' button. The editing certainly improves my writing over not bothering (sometimes I don't bother) but there's still a lot to be desired in the results I think.

My stylistic habits have only further cemented, I'm sure. My love of parentheticals is still strong. I like to start sentences with And or But. I get carried away with ellipses meant as just a softening or trailing off voice. I sometimes make way too long sentences and like to uses semicolons and double dashes -- incidentally if you see me use a — symbol (emdash or endash?) then you can assume I got replaced by an AI. Sometimes I use a … symbol over ... but mostly on twitter since it'll count for 2 instead of 3 chars but then I often just use .. instead. Drop subject markers a lot in short form like IMs or twitter, but it shows up here and elsewhere too. (Maybe can blame JP learning a bit since that language goes full cave man with its dropping things and assuming context.) I sometimes wonder if subconsciously I've ever used the joking thought "'I'-heavy text is often annoying to read, suggests narcissism, but if you just omit the 'I' you'll be fine!" as actual justification. I probably insert "that" into my phrasing more than necessary, for that I blame French education somewhat. (Sometimes I confuse myself if maybe I should use "which" instead of "that", my rule of thumb is that if I can drop the subclause and the overall meaning is mostly retained, then use "which".)

I have started to include AI analysis/critique/commentary on some of my writing. Usually only for some HN posts or emails though, not so much for blog posts. It's helped a bit in pointing out organization issues or places I could be clearer, or will sometimes tell me I'm saying something false or not so well supported. Sometimes it will suggest specific rewording or revisions, I pretty much never use them verbatim since it destroys my own style (not so bad if trying to mask my style I guess). If I'm intentionally being sharp/a bit snarky/aggressive it will smooth that out too. A lot of its word choices aren't as good so I keep my own. It also tries not to assume too much about an audience even if I tell it one so will oversimplify something. It is helpful that way though when trying to be diplomatic or avoid conflict or see alternatives about how to better word something. A few times I've been surprised that, by telling the AI that I'm uncertain about even posting something (e.g. I don't think it really adds to the conversation in a useful way) and it's ok to tell me to not bother, it has actually told me to not bother. Without that context it will just default to trying to be supportive and improve the writing in various ways it thinks are possible improvements.

Teenage me would be appalled at the number of spelling mistakes / typos I tend to make and not always notice. Firefox's spellcheck in textarea fields can be partially blamed since it kind of breaks sometimes~ Most typos are probably in instant messaging contexts (youtube live chat is related) where I do notice them but only after sending, and editing isn't always worth it. Or in youtube live chats you can't edit anyway, you have to delete and repost, and it's kinda pointless anyway since it could have already scrolled off screen or been read by the streamer.

I'm still amused/partially horrified by the one time I sent an email to HR. She told me in person that she was sharing it around the HR team and complimenting how carefully worded it was. Yes, that was by design.

I tend to add too much information a lot of the time though. Some of this is probably "internet autism", some comes from a desire to be better understood; though like code, more isn't always better. More words can increase the chance of unintentional irony that people can find annoying or upsetting.

Well, just wanted to reflect and ramble a bit. Again, so much for writing a lot, it hasn't really made me better. If you're writing professionally, just the act of writing every day is going to improve you more than something like reaching a particular amount. Intensity over time matters more than pure time. Like, you can actually get kind of far in learning a foreign language if you sign up for a 3 month course that has you working at least a few hours each day on it, but you won't get very far if you just do some anki studying once in a blue moon, and over 10 years later you might have more hours than that 3 month course would have put you through but still know a lot less.

Then again, the advice is "reach one million words as fast as you can", so it does take speed or intensity over time into account. I've had this blog for 16 years, and the blog itself has only hit about half a million, that's not very fast. I just don't have that much to write about.

---

Posted on 2025-07-19 by Jach

Tags: personal, thought

Permalink: https://www.thejach.com/view/id/445

Trackback URL: https://www.thejach.com/view/2025/7/the_futility_of_one_million_words

Back to the top