#posse

There's also a ton of technical work that goes into publishing each edition of this newsletter. Things have moved on somewhat since I published an in-depth process-post in 2021, though I'm still totally reliant on Loren Kohnfelder's python scripts that help me turn the XML file I compose every day into files that are (nearly) ready to publish:

https://pluralistic.net/2021/01/13/two-decades/#hfbd

Much of the technical work is down to the fact that I'm still completely wed to the idea of "POSSE" (Post Own Site, Syndicate Everywhere):

https://pluralistic.net/2022/02/19/now-we-are-two/#two-much-posse

This means that after I write the day's post, I reformat it and republish it as a text-only newsletter, a Medium post, a Tumblr post, a Twitter thread and a Mastodon thread. This involves a ton of manual work, because none of the services I post to are designed to facilitate this, so I'm always wrestling with them. This year, all of them got worse (incredibly).

Medium – where I used to have a paid column – has dropped its free-flag for my account, which now limits me to how many posts I can schedule. This doesn't come up often, but when I do schedule a post, it's generally because I'm going to be on a plane or a stage and won't be able to do it manually. There's no way I'm going to pay for this feature: I'm happy to give Medium my work gratis, but I will not and do not pay anyone to publish my work, and I never will.

Tumblr did something to its post-composing text editor that completely broke it and I've given up on fixing it. I can't even type into a new post field! I have to paste in some styled text, then delete it, then start typing. It's ghastly. So now I just have a text file full of formatted HTML snippets and I work exclusively in the Tumblr HTML editor, pasting in blobs of preformatted HTML (including the florid, verbose HTML Tumblr uses for its own formatting) and then laboriously flip back and forth to the "visual" editor to see the parts that went wrong. Here's how busted that visual editor is: searching for a word then double-clicking on it does not select it. You have to click once, wait about 1.5 seconds, click again, wait again, and then you can select the word.

Twitter has entered a period of terminal technical decline. I know, I know, we always talk about how fucked Twitter's content moderation is, for obvious and good reasons, but from a technical perspective, Twitter just sucks. If I make a post with an image and alt text in anticipation of later using it to start a thread, it often goes "stale" and will not publish until I delete the image and re-attach it and re-paste the alt text. Meanwhile, the thread editor is also decaying into uselessness. Fill in a 25-post thread and hit publish and, the majority of times, the thread publication will die midway through, displaying lots of weird failure modes (phantom empty posts at the end of the thread that need to be individually selected and deleted are a common one, but not the only one). The old Twitter's ability to add a new thread to an existing one has been dead for at least a year, so every post after the 25th stanza has to be manually tacked on to the previous one, which is made far harder by the fact that Twitter no longer reliably shows you the post you just made after it publishes.

Mastodon still lacks a decent thread editor, one that has even the minimal functionality of Twitter circa 2020. Meanwhile, the Fediverse HOA continues to surface from time to time, with someone who's had a Masto account for ten seconds scolding me for posting threads – from my account whose bio starts "I post long threads." It's genuinely tedious to be shouted at for "using Mastodon wrong" by someone who started using Mastodon yesterday (I opened my first Mastodon account in 2018!), and even worse when they double down after I point them to the essay I've written to explain why I post the way I do, and what to do if you want to read my work somewhere that's not your Mastodon timeline ("Can you believe this asshole wrote a whole essay to explain why he posts his stupid Mastodon threads?"):

https://pluralistic.net/2023/04/16/how-to-make-the-least-worst-mastodon-threads/

Then there's email: I continue to love email, but email doesn't love me back. After years of being blackholed by AT&T and then Google, this turns out to be the year that Microsoft bounces thousands of messages to its Hotmail and Outlook users because they have arbitrarily and without warning added my mail-server to a blacklist. Thank you to the Fediverse friends who escalated my trouble ticket – but man, this is a headache I could certainly do without:

https://pluralistic.net/2021/10/10/dead-letters/

My sysadmin, the incomparable and tireless Ken Snider, tells me that he's got the long-overdue new hardware installed at the colo and he's nearly ready to stand up my long-anticipated personal Mastodon server, which will let me solve all kinds of problems. He's also going to stand up my own Bluesky server, at which point I will part ways with Twitter. I wish I could have used the regular Bluesky service while I waited, but just setting up an account permanently binds you to totally unacceptable and dangerous terms of service:

https://pluralistic.net/2025/08/15/dogs-breakfast/#by-clicking-this-you-agree-on-behalf-of-your-employer-to-release-me-from-all-obligations-and-waivers-arising-from-any-and-all-NON-NEGOTIATED-agreements

What's the point of a service that has account- and data-portability if signing up for it makes you permanently surrender your rights, even if you switch servers? This might be the stupidest social media unforced error of the post-zuckermuskian era.

There is one technology that has made my POSSE life better, and it might surprise you. This year, I installed Ollama – an open-source LLM – on my laptop. It runs pretty well, even without a GPU. Every day, before I run Loren's python publication scripts, I run the text through Ollama as a typo-catcher (my prompt is "find typos"). Ollama always spots three or four of these, usually stuff like missing punctuation, or forgotten words, or double words ("the the next thing") or typos that are still valid words ("of top of everything else").

The reason this is so valuable to me is that errors magnify through each stage of POSSE. Errors that make it through the python publication script take 10x the time to fix that they would if I caught them beforehand. Errors that I catch after running the scripts and publishing the posts take 10x time more. Errors that I have to fix later on – once I've closed all the relevant tabs and editors – take 10x again more time. Some POSSE channels (email, Twitter) can't be fixed at all.

So catching these typos at the start of the process is a huge time-saver. I have some very generous readers who have the proofreader's gene and are very helpful in catching my typos (hi, Gregory and 9o6!), and I feel bad about depriving them of their fun, but there's still the odd error that slips through, and they always catch it.

Ollama is a pretty good typo-catcher. Probably half of the "errors" it points out are false positives, which is better than the false positive rate for Google Docs' grammar-checker. As someone who uses a lot of jargon, made up words, etc in his prose, I'm used to overriding my text-editor. I wouldn't simply trust an LLM's edits any more than I would accept every suggestion from a spell-checker. Hell, yesterday I sent back a professionally copyedited manuscript (the intro for the paperback of Enshittification) and marked "STET" on about a third of the queries.

Doubtless some of you are affronted by my modest use of an LLM. You think that LLMs are "fruits of the poisoned tree" and must be eschewed because they are saturated with the sin of their origins. I think this is a very bad take, the kind of rathole that purity culture always ends up in.

Let's start with some context. If you don't want to use technology that was created under immoral circumstances or that sprang from an immoral mind, then you are totally fucked. I mean, all the way down to the silicon chips in your device, which can never be fully disentangled from the odious, paranoid racist William Shockley, who won the Nobel Prize for co-inventing the silicon transistor:

https://pluralistic.net/2021/10/24/the-traitorous-eight-and-the-battle-of-germanium-valley/

Further, we wouldn't have the packet-switched network that delivered these words to you without the contributions of the literal war-criminals at the RAND corporation:

https://en.wikipedia.org/wiki/ARPANET

Refusing to use a technology because the people who developed it were indefensible creeps is a self-owning dead-end. You know what's better than refusing to use a technology because you hate its creators? Seizing that technology and making it your own. Don't like the fact that a convicted monopolist has a death-grip on networking? Steal its protocol, release a free software version of it, and leave it in your dust:

https://www.eff.org/deeplinks/2019/07/samba-versus-smb-adversarial-interoperability-judo-network-effects

That's how we make good tech: not by insisting that all its inputs be free from sin, but by purging that wickedness by liberating the technology from its monstrous forebears and making free and open versions of it:

https://pluralistic.net/2025/01/14/contesting-popularity/#everybody-samba

Purity culture is such an obvious trap, an artifact of the neoliberal ideology that insists that the solution to all our problems is to shop very carefully, thus reducing all politics to personal consumption choices:

https://pluralistic.net/2025/07/31/unsatisfying-answers/#systemic-problems

I mean, it was extraordinarily stupid for the Nazis to refuse Einstein's work because it was "Jewish science," but not merely because antisemitism is stupid. It was also a major self-limiting move because Einstein was right:

https://www.scientificamerican.com/article/how-2-pro-nazi-nobelists-attacked-einstein-s-jewish-science-excerpt1/

Refusing to run an LLM on your laptop because you don't like Sam Altman is as foolish as refusing to get monoclonal antibodies because James Watson was a racist nutjob:

https://www.statnews.com/2025/11/07/james-watson-remembrance-from-dna-pioneer-to-pariah/

Or to refuse to communicate via satellite because they were launched into space on a descendant of a rocket designed by the Nazi Wernher von Braun and built by slaves in a death camp:

https://wsmrmuseum.com/2020/07/27/von-braun-the-v-2-and-slave-labor/4/

The AI bubble sucks. AI itself is a normal technology:

https://knightcolumbia.org/content/ai-as-normal-technology

It's not "unethical" to scrape the web in order to create and analyze data-sets. That's just "a search engine":

https://pluralistic.net/2023/09/17/how-to-think-about-scraping/

There's plenty of useful things people can do with AI. There's plenty of useful things people will do with AI. AI is bad because it's an economic bubble and a grift, but not because we've created a bunch of utilities that would – under normal circumstances – be called "plug-ins":

https://pluralistic.net/2025/12/05/pop-that-bubble/#u-washington

I started blogging 25 years ago, just before the dotcom bubble popped. That bubble-pop inflicted a lot of pain on people who didn't deserve it, including the normie investors who'd been suckered into blowing their life's savings on dogshit stocks, and everyday workers who found themselves out of a job. But the world was better off. So was the web. With the bubble popped, real, good stuff could access talent, servers and office space.

In the six years I've been doing this, I've seen several bubbles come and go: crypto, web3, metaverse. Now it's AI. But those bubbles were like Enron, frauds that left nothing good behind. AI is like the dotcom bubble, awash in sin and inflicting untold misery, but it will leave something useful behind:

https://pluralistic.net/2023/12/19/bubblenomics/#pop