Category: Writing

Posted on
by

I recently added “Your LLM was trained on my blog” to the intro on my home page. I’m sort of being facetious, but it’s also true: my old site (hyperborea.org) has been in the Common Crawl data set for years, this blog has been online since 2002, and the major LLMs have used Common Crawl’s data in their training.

Way back when I switched to WordPress in 2003, one of the things I found annoying about it was that it would auto-convert punctuation from ASCII to “fancy” characters. Curly quotes, curly apostrophes, em-dashes instead of double-hyphens, ellipses…These were things that bugged the heck out of me when I saw them in broken emails and Word documents. Over the next year, seeing the difference it made in legibility, I came to appreciate real typography. Since then I’ve been happy to let WordPress, and later ClassicPress and Eleventy convert my plain typing.

Side note: I still have trouble using some apps with code snippets (ex: Apple Notes and Stickies), ’cause I haven’t figured out how to stop them from converting ASCII punctuation that means something in the code to Unicode punctuation that means something more specific in printed human language. Also, notice in that sentence that ClassicPress was smart enough to convert a straight up and down apostrophe used on a contraction to a curly apostrophe, and not to a curly single quote like it does with ‘here.’

Anyway, WordPress powers a huge amount of the web. I’ve seen stats bandied about such as 42% of all websites, 60% of content management systems, >50% of the top 10,000 websites (by some measure or another). The exact numbers aren’t important, it’s just important to know that a whole lot of the web uses WordPress to convert what people write into HTML pages for the web.

Now let’s get back to LLMs:

  • People think they can identify LLM-generated text because it tends to include em-dashes.
  • LLMs were trained on writing by actual people on the web.
  • A whole lot of that writing by actual people is presented by WordPress.
  • WordPress has auto-converted typography including em-dashes since 2003.
  • That means a large chunk of LLMs’ training data includes text by real people that uses real em-dashes.

“AI” doesn’t use em-dashes because real people don’t, it uses em-dashes because real people do [note: lots of swearing] (via).

The only place I can imagine it would be even a semi-reliable indicator would be in social media posts, on platforms where that kind of auto-conversion doesn’t happen. Even then, I can easily imagine someone writing a draft in Word, or copying a sentence or two from their WordPress post, or just hitting option+shift+hyphen on a Mac (or whatever the key combo is on Windows).

Posted on
by

A story’s been making the rounds about a software project that enforced a no-LLM-use policy by using prompt injection to delete itself. An “AI” agent-using coder filed a bug report (understandable), but filled it with a bunch of long-winded, clearly LLM-generated comments.

I looked at those comments. I can’t say I read them, because my eyes started glazing over a couple of paragraphs in. The contrast with the posts by the maintainer and other commenters is…stark.

Though I did notice the bit about how nobody reads the docs, which seems rather telling.

One of the problems with letting an “AI” write for you: If you aren’t reading it, and you assume the person at the other end is just going to summarize it anyway, there’s no motivation to make it readable. And no motivation to think about it and narrow down what’s important. And if you’re rewriting the prompt to focus on what matters most, consider that the prompt would get the idea across more effectively.

Posted on
by

Humble Bundle is offering 30 books* by Ursula K. Le Guin supporting the Literary Arts charity, including all of Earthsea, several Hainish novels, Catwings, short stories, Gifts/Voices/Powers, nonfiction writing…

I’ve read the Earthsea series (good-to-great) and most of the Hainish novels (some great, some good, some OK), plus Lathe of Heaven (great), and I’ve got copies of several more on my to-read pile, but I’m totally grabbing this for the nonfiction and short stories. Plus I get another format for my Earthsea reread that’s not the giant collected hardcover tome or my old fragile paperbacks.

I’m going to have a lot of reading material next year.

*As ebooks through Kobo

Update: Unfortunately they aren’t DRM-Free. Kobo does have some of her books without DRM, but as far as I can tell, the only one in this bundle is The Unreal and the Real.

The others I’ve found just now are: Worlds of Exile and Illusion, The Eye of the Heron, The Word for World is Forest, and The Beginning Place. Tor seems to be big on requesting non-DRM formats. These may be all of them currently available without DRM: eBooks.com, where I previously bought Forest, has a filter on their search for DRM-Free, and only turns up the same four books.

It seems kind of ironic to sell The Dispossessed (of all books!) encumbered with DRM.

Posted on
by

For today’s daily blog, I was thinking about posting something about getting out there and voting in the midterm election tomorrow. Aside from the fact that I already posted about elections a few days ago, I realized another problem: who’s going to read it?

I mean, K-Squared Ramblings doesn’t seem to have many regular readers as near as I can tell. And Key Smash! is for testing – I’m not really expecting to pick up many regular readers there. So a call to action isn’t going to reach many people on either of these blogs.

Of course, that brings up the question of why am I blogging these days? Who is my target audience?

  • How-to and troubleshooting posts are intended for other people with similar problems. They don’t have to be blog entries, it’s just easier to write and edit that way.
  • Recurring themes (like funny signs, or solar halos) are intended for anyone who lands on a related article or photo. It’s easier to tie things together on one site than across several.
  • Some of it is soapbox material that I just feel like I need to get it out there. The kind of stuff you shout into the void on Twitter or Mastodon, but maybe it was too long for that, or maybe I did post it on one of those sites and decided that it was worth hanging onto. (Aside from lack of control and limited search functionality, sometimes social networks shut down.)
  • Some of it is stuff that I find interesting and want to share. I usually end up sharing links over social media, and then if I have more that I want to say I’ll do it on a blog, because it’s much better-suited for that. I don’t have anyone specific in mind, just whoever might find the same stuff interesting.
  • And some of it is for me. I don’t think many people read my convention reports (these days people just go for the photos), but I’ll go back myself and check, “Which year was that, again? Was that WonderCon or SDCC?”

Posted on
by

I don’t have the time or ideas for Nanowrimo this year. It’s actually been a decade since I last did it, now that I think about it. But I’ve done NaBloPoMo a few times, and I think I can manage a month of posting one blog entry a day.

Plus it’ll be a good way to test Pterotype federation and other stuff.

I’ve already posted two days in a row at my main blog K-Squared Ramblings. Yesterday I posted a couple of Halloween photos, and today I posted about a hosting problem that took out this blog for a few days.

I do feel like I should try to put some effort in rather than just taking a tweet/toot and calling it a blog entry. But what makes something suitable for blogging vs a social media post?

  • Long-form writing that doesn’t fit in a short note.
  • Image(s) with long descriptions or extended commentary.
  • I think I can probably take something that fits a series and justify it as a blog entry even if it’s short or just a photo with a caption.

So between Key Smash! and K2R, I’ll try to post at least once a day this month. If there are days I end up only posting here, I’ll probably move the posts over when I’m done.

And somewhere along the line I’ll turn on Pterotype on K2R!

Update: I tried for a couple of days but dropped out. Though I did manage to make 19 posts across 15 days of the month between the two blogs – probably more, since I didn’t import all of the Key Smash! posts before taking it offline.

Posted on
by

I don’t like Twitter threads.

In most cases, if something takes more space than one or two tweets to say, it’s easier to read as an article. It’s especially bad with very long threads, and those that aren’t crafted to make each tweet a unit. When sentences continue on from one tweet to the next as if they’re only line breaks, it makes it hard to pick out a statement to highlight, or where to start.

On the plus side: Tweets are more likely to be seen, more easily shared, and people can interact as they’re posted, like a live conversation. A thread that’s crafted to fit in 140-character chunks has a rhythm to it, like a daily comic strip collection vs. a comic book. And an unplanned tweetstorm gives both the writer a chance to get their ideas down quickly and the reader a chance to see them unfiltered.

Compared to an article, a tweetstorm provides immediacy, and any Twitter thread provides reach. But they’re still a pain to read.

Originally written (of course) as a Twitter thread.

Posted on
by

Well, I dropped out of the 30-day blogging challenge after 20 days. Ironically, I think part of the problem was that I tried to get ahead. Most days I would write a post in the evening, either just after the kid went to bed or just before I did. On day 20, I posted at lunch. That broke my routine, and I forgot to post the next evening.

So…

I finished a few posts I’ve been working on for a while, so that worked out. There were several things I noticed that turned into what I hope made interesting commentary. That’s two wins. But then there were the days of filler, too. It’s not like I need to keep “I hate the 405” on its own page, after all. (Besides, that’s a given.)

Things I Learned

I really don’t like writing about current events or politics these days. On Friday the 13th, I couldn’t think of anything else to write about, but couldn’t start. I ended up reading Tumblr for some idea to jump off of, and found the Epi-Pen infographic. Then I went a whole week deliberately not writing about Paris, Beirut, or refugees.

If I wanted to keep the volume up without regard to focus, I think up plenty of ideas during the course of a day. I just have to (1) write down the topic so I remember and (2) find the time/inclination/brainpower to do the writing.

I wrote more about my kid than I expected. Probably more than I’ve written publicly in quite a while. This isn’t a parenting blog by a long shot, but in addition to costuming and art, I used several family conversations as jumping off points.

When I obsessively listen to a show for a week, I come up with lots of things I’d like to write about it. I followed Chess with a Les Misérables binge, and now I’ve got about five Les Mis-related topics on my list.

If you can’t think of anything to write about, ask your spouse for a topic.

I’m actually more comfortable sitting on the couch and writing on my tablet than firing up my computer and sitting at my desk. That surprised me, but the Galaxy Tab S2 is sooooo much faster than the Nexus 7 it’s ridiculous. And hey, couch!

Tally

6 Random observations.
4 Amusing photos that fit within one of the site’s long-running categories.
3 Articles I’ve been working on for a while and finally got around to finishing.
3 Follow-ups to older posts or series.
2 Reaction pieces.
1 Interesting photo that I cross-posted from Instagram/Flickr to satisfy the goal.
1 Total filler that I would delete if it weren’t for keeping a record of how far I got in the challenge.

List

  1. Mixed Emotions/Halloween/Comikaze
  2. Fruit Basket Case
  3. So Much for SudaSudafed
  4. I “Liked” Twitter Favorites
  5. Phone vs. Camera
  6. WiFi is the New Color TV
  7. Gloom ahead, blue sky behind
  8. So that’s where Vaporware comes from!
  9. Aaaargh!
  10. Genre TV Update
  11. Chess (Musical)
  12. Les Mis/Chess: Cosette vs. Florence (and Eponine)
  13. Epi-Pen How-To
  14. Symbolic Costuming: Stage vs. Screen
  15. Coin Slot For Your Cell Phone
  16. AAAAAARP!
  17. Panorama Fail: Chalk Edition
  18. Is the ransom note look obsolete?
  19. Too Many Notifications
  20. Oddly Specific Advertising

»All pages site-wide with this tag