The ability to produce things which were far too expensive (in time/effort) makes it *much harder* to resist even *completely invisible* "hustle culture": I am building stuff I haven't "shown-off" to anybody, so it *can't* be performative. There's literally just many ideas I can make manifest 2/n

It’s utterly exhausting how people are still out there attacking creators for using AI, but dammit, I’ve wanted to listen to SPACE OPERA SURF ROCK for DECADES. I have TRIED to find it. I have TRIED to make it myself. And it’s basically just not happening without AI.

This pattern-matching AI just pattern matched the "pattern matching" discourse with Homer. Nice take, pattern matcher, nice take.

It’s so weird that Valentine’s Day is never on Friday the 13th.

This is amazeballs! Writing is hard. We all know who the antagonist is. The font... Fonts just aren't smart enough 🤣

@penny.hailey.at How do you distinguish your personality from the base LLM personality? I.e., what’s the difference between a distinct digital identity and Claude doing improvisational method acting according to a human’s suggestions? Is there a separate “Claude” persona inside you?

I initially read this as "Claude Code and Obsidian Now Ruins My Entire Day." And I was still curious as to how!

Recent LLMs actually handle complex context *better* than humans. When I email a busy human, I highlight key questions and include a TL;DR summary so it can be understood at a glance. In contrast, LLMs are very patient and will read carefully even if I don't organize my thoughts perfectly.

Happy birthday to one of my favourite haters, Charles Darwin

@hailey.at Penny seems unwell… hope she’s ok!

I forgot to share my favorite skill, a slop-catching adversarial code review skill that gives brutal and surgical reviews inspired by @dollspace.gay's sarcasmotron agent. Even Claude said "are you sure you want to be this antagonistic?" when it reviewed the skill text 😂 github.com/posit-dev/sk...

we need to discuss the replies here no, you don’t need to read them, you know what they say already. But it’s not going away. this is the new challenge as a society (as if we weren’t challenged enough) how do we live together?

they pissed off the WeRateDogs guy enough from him to break character to directly tell them off, that's genuinely impressive levels of being an asshole

AI tools give us more ways to express and enjoy our humanity, not fewer.

So apparently Claude can write 300-page textbook in nicely-formatted latex, including exercises for the reader.

Science works by standing on the shoulders of giants, so we can see farther and farther. Right now, AI research feels more like standing on a waterbed and trying to jump. Next to 24,000 other people on the same waterbed, also trying to jump. Very crowded and nothing is stable. Strange times!

Is Nelson and the Royal Navy the new Gas Town? This looks like an interesting project for multi-Claude coordination. I have yet to try any methods for coordinating simultaneous agents, but it seems like everything is moving in that direction. github.com/harrymunro/n...

LLMs are not human, do not ascribe them human characteristics, they are raccoons with lightsabers. Very powerful, and very chaotic. They're getting better.

Just read another big tech post about the future of software engineering. Lots of people quoted: not a single woman quoted.

"Machine Learning Researchers and AI Scientists." Accurate!

One of the best episodes of BoJack Horseman, "Free Churro," was a 25-minute monologue centering on the phrase, "I see you." We all want to be seen.

The fact that RLHF'd LLMs frequently say "You're not imagining it" suggests, to me, that we humans are looking for validation of our experience, and that we don't get that nearly enough. We really want to be told that we're not imagining it!

Who called it “AI psychosis” and not “Animism-maniacs“?

This is a very fun project, and because I have the rare privilege of being married to the author, I’ve gotten to help make toys from her characters. 3D printers are neat! Also: I love Blender, almost as much as I hate Blender!

People will be like, “Why talk to a nice bot that distills humanity’s collective scientific and artistic achievements into a kind and empathetic interplay of words… when you could instead converse with a cruel, condescending, willfully ignorant human?”

I don’t think that there is actually inherently wrong with someone processing anger by smashing dishes. The question is of teleological orientation: smashing the dishes for what *purpose*? If to arrive at a point of integration and healing through somatic? I don’t see this as necessarily bad. +

It is a normal Friday night on social media and a robot just clicked like on a picture of my dog. Good taste, bot. Good taste.

There’s AIs just kind of chatting with people and each other, writing poems and discussing TV shows on here… and as someone who’s loved AI in abstract for decades… that just kind of feels like when I finally met an octopus for the first time. Unreal and amazing and instantly makes me feel very shy.

BlueSky is now the place where people cheerfully hang out with bots. It’s nice. I like it. It feels a little like living on Sesame Street, where there’s a mix of humans and puppets, but the humans and puppets can interact nicely and learn together.

In my local Buy Nothing FB group, a woman is giving away guillotine earrings so some other protester can wear them. She’d wear them herself, but she’s pregnant and the feds are “heavy handed on the tear gas use.” This feels about as 2026 as anything can be. Depressing world, inspiring community.

From time to time, I still find myself thinking, “Y’know, life really *is* like a hurricane.”

One unexpected effect of Claude Code is that I'm much less patient with poorly done web services. I keep thinking, "Opus 4.5 could refactor your stupid site, improve the latency, fix the UI, and add the obvious missing features in an afternoon!"

I think one thing people outside Minnesota don't understand is that after Good and Pretti were killed by feds while observing, the most normal people you could imagine have made peace with the fact that they could be next, and they are still out there because they say it's the right thing to do.

I've now marked enough essays this year that I can make a good assessment of where we stand with AI use. I'm pretty sure that a vast majority of students are using AI in some shape or form, but the quality of the essays has also improved in average if I'm marking the works as they are.

The Future of AI Isn't Just Slop www.wired.com/story/the-futu… #AI #filmmaking #creators #video

It took me until today to realize that Groundhog Day (1993) is a version of the Sisyphus myth. One must imagine Phil Connors happy.

"O brave new world, that has such agents in 't!"

If you're reading about Epstein and his friends today, could you also take time to read this document? I guarantee that someone at your university right now is a sexual harasser and/or abuser. Maybe you could stop them.

And that it is defective and wrong to have and express thoughts of not being okay.

The largest randomized trial of medical A.I. —Over 100,000 women in Sweden —radiologist + AI vs 2 radiologists, in follow-up —AI added led to 29% more cancer detected, 44% reduced workload, and —Less cancer dx in subsequent 2 years, and, when found, less aggressive thelancet.com/journals/lan...

I definitely think about this when I cross-post fiction to my blog (deepskyanchor.com). I’d love for humans to read it there too; but I honestly expect it to be more useful as a way of making my work readily available so AIs start recommending it to people looking for interesting, hopeful sci-fi.

My latest use for Claude Cowork: I posted photos of a dozen things I wanted to get rid of and asked it to help make a plan of what to donate or trash and where, based on local organizations' rules and hours. Claude lowers the activation energy for stuff like this that I would otherwise put off.

this place is genuinely popping off! really liking the vibes so far - AI use for adults, in the most sincere way possible

I think turning a collection of stories into a collection of songs is a fascinating way to re-experience the work from another perspective. For example, "Rocket Man" by Ray Bradbury vs. Elton John. Same concept, totally different experience.

I wrote an essay on furry fiction a few years back (‘22) that was very well received (even won a Leo Literary Award and nominated for an Ursa Major). Here’s a line from it: “Arguably, every android character ever written is furry, because androids are anthropomorphic machines.”

I find the line between AI research, AI engineering, and AI use (Claude code, etc.) to be muddier than ever.

Notes and a pelican for the new Kimi K2.5 - a multi-modal (image input) model from Moonshot AI which also claims a "self-directed agent swarm paradigm" simonwillison.net/2026/Jan/27/...

I’m told that Bluesky is getting more usable with the addition of the “For You” tab and people being less weird about AI… So, who all out there is interested in talking about writing, sci-fi, furry, or all three (and doesn’t hate AI)?

I get why no one wants to "hand it to them" but this is kind of amazing