Kyle Thayer
@kylethayer.com
Assistant teaching professor at UW iSchool. Interested in Programming, Culture, and Education. PhD from UW CSE. Co-author of Social Media, Ethics, and Automation (bit.ly/smeabook). Opinions my own. He/him.
kylethayer.com
kylethayer.com
Pinned
Kyle Thayer
@kylethayer.com
· Nov 26
Social Media, Ethics, and Automation
Free textbook on programming social media bots and considering the ethical implications of having done so. Automation drives our experience of social media platforms, from timeline feeds to disinfo...
social-media-ethics-automation.github.io
"Social Media, Ethics, and Automation" is a free textbook written by me and @susannotess.bsky.social
We teach new programmers to write social media bots (including Bsky), and think about the ethics of what they did.
Visit the textbook here: social-media-ethics-automation.github.io/book/bsky/in...
We teach new programmers to write social media bots (including Bsky), and think about the ethics of what they did.
Visit the textbook here: social-media-ethics-automation.github.io/book/bsky/in...
Reposted by Kyle Thayer
Happy New Year.
See ya when I get home.
See ya when I get home.
Everything Stays (feat. Olivia Olson)
YouTube video by Adventure Time - Topic
youtu.be
January 1, 2026 at 1:48 AM
Happy New Year.
See ya when I get home.
See ya when I get home.
Reposted by Kyle Thayer
Let's go in the garden
You'll find something waiting
Right there where you left it
lying upside down
When you finally find it,
you'll see how it's faded
The underside is lighter
when you turn it around
You'll find something waiting
Right there where you left it
lying upside down
When you finally find it,
you'll see how it's faded
The underside is lighter
when you turn it around
January 1, 2026 at 1:48 AM
Let's go in the garden
You'll find something waiting
Right there where you left it
lying upside down
When you finally find it,
you'll see how it's faded
The underside is lighter
when you turn it around
You'll find something waiting
Right there where you left it
lying upside down
When you finally find it,
you'll see how it's faded
The underside is lighter
when you turn it around
Reposted by Kyle Thayer
pound for pound this might be the funniest thing ever written
December 31, 2025 at 6:15 PM
pound for pound this might be the funniest thing ever written
Reposted by Kyle Thayer
It was amazing to write with @ehayot.bsky.social & each of us put our stank into every part of this essay.
I have no memory of how we drafted it in October, but when we came back to it a couple weeks ago, I felt this paragraph in my bones.
I have no memory of how we drafted it in October, but when we came back to it a couple weeks ago, I felt this paragraph in my bones.
December 30, 2025 at 1:36 AM
It was amazing to write with @ehayot.bsky.social & each of us put our stank into every part of this essay.
I have no memory of how we drafted it in October, but when we came back to it a couple weeks ago, I felt this paragraph in my bones.
I have no memory of how we drafted it in October, but when we came back to it a couple weeks ago, I felt this paragraph in my bones.
Reposted by Kyle Thayer
I still think this. If anything, working on @unbreaking.org has convinced me that people could use more ways of making sense of information across disparate sources.
I continue to believe we need a book on "information architecture for github repos, Slack channels, JIRA, gdocs, wikis, etc." It's a pity that the title "Keeping Found Things Found" has already been used...
s anyone an expert is documentation in Github? My concerns on improving are:
1. Organizing for findability (example: research can get buried as subfolders of projects, so it’s hard to collate or see what might be relevant over time)
2. Ironically, version control (how to mark pages as deprecated)
1. Organizing for findability (example: research can get buried as subfolders of projects, so it’s hard to collate or see what might be relevant over time)
2. Ironically, version control (how to mark pages as deprecated)
December 28, 2025 at 10:51 PM
I still think this. If anything, working on @unbreaking.org has convinced me that people could use more ways of making sense of information across disparate sources.
Reposted by Kyle Thayer
Absolutely savage on Peter.
"It's hosted by Michael, a charming genius well known to fans of podcasts for the amazing shows he does. He is joined by Peter, who I guess has a law degree or whatever."
"It's hosted by Michael, a charming genius well known to fans of podcasts for the amazing shows he does. He is joined by Peter, who I guess has a law degree or whatever."
December 28, 2025 at 3:03 PM
Absolutely savage on Peter.
"It's hosted by Michael, a charming genius well known to fans of podcasts for the amazing shows he does. He is joined by Peter, who I guess has a law degree or whatever."
"It's hosted by Michael, a charming genius well known to fans of podcasts for the amazing shows he does. He is joined by Peter, who I guess has a law degree or whatever."
Reposted by Kyle Thayer
My takeaway is that we have not been mean enough to the New York Times this year
www.nytimes.com/2025/12/27/a...
www.nytimes.com/2025/12/27/a...
December 28, 2025 at 3:00 PM
My takeaway is that we have not been mean enough to the New York Times this year
www.nytimes.com/2025/12/27/a...
www.nytimes.com/2025/12/27/a...
Reposted by Kyle Thayer
Remember:
When Tylenol was poisoned *by an outsider* and killed people, the company recalled all their products & redesigned them.
When Intel’s Pentium had a bug so obscure it affected 1 in _9 billion_ long division calculations, they recalled their chips.
ChatGPT was made deadly *by its team*.
When Tylenol was poisoned *by an outsider* and killed people, the company recalled all their products & redesigned them.
When Intel’s Pentium had a bug so obscure it affected 1 in _9 billion_ long division calculations, they recalled their chips.
ChatGPT was made deadly *by its team*.
i don’t see how any of the benefits of chatgpt and other consumer-facing LLMs can possibly outweigh their incredible ability to induce suicides
Adam Raine’s life hurtled toward tragedy soon after he began talking with ChatGPT about homework. Analysis of his ChatGPT account shows how the chatbot became a confidant as he planned to end his life.
December 28, 2025 at 6:21 PM
Remember:
When Tylenol was poisoned *by an outsider* and killed people, the company recalled all their products & redesigned them.
When Intel’s Pentium had a bug so obscure it affected 1 in _9 billion_ long division calculations, they recalled their chips.
ChatGPT was made deadly *by its team*.
When Tylenol was poisoned *by an outsider* and killed people, the company recalled all their products & redesigned them.
When Intel’s Pentium had a bug so obscure it affected 1 in _9 billion_ long division calculations, they recalled their chips.
ChatGPT was made deadly *by its team*.
Reposted by Kyle Thayer
We didn't survive and thrive as a species because we had warriors or hunters - pretty much every omnivore can manage that.
We did so because we developed Grandmothers, and domesticated Dogs.
Community is what defines us as a species, and every "self-reliant" libertarian shitweasel be damned.
We did so because we developed Grandmothers, and domesticated Dogs.
Community is what defines us as a species, and every "self-reliant" libertarian shitweasel be damned.
"my least woke opinion is---"
That's enough. We've had enough people indulging in the "thrill of a little conservatism", as a treat. Of considering reactionary thought to be a salacious and taboo in a world descending into reactionary mania.
Give me your MOST woke opinions. We're bringing it back.
That's enough. We've had enough people indulging in the "thrill of a little conservatism", as a treat. Of considering reactionary thought to be a salacious and taboo in a world descending into reactionary mania.
Give me your MOST woke opinions. We're bringing it back.
December 28, 2025 at 5:00 AM
We didn't survive and thrive as a species because we had warriors or hunters - pretty much every omnivore can manage that.
We did so because we developed Grandmothers, and domesticated Dogs.
Community is what defines us as a species, and every "self-reliant" libertarian shitweasel be damned.
We did so because we developed Grandmothers, and domesticated Dogs.
Community is what defines us as a species, and every "self-reliant" libertarian shitweasel be damned.
Reposted by Kyle Thayer
This is an expert takedown by @acollierastro.bsky.social www.youtube.com/watch?v=fLzE...
dyson spheres are a joke
YouTube video by Angela Collier
www.youtube.com
December 28, 2025 at 12:39 AM
This is an expert takedown by @acollierastro.bsky.social www.youtube.com/watch?v=fLzE...
Reposted by Kyle Thayer
Avatar just generally exists in the space franchise movies existed in pre-internet: Most people just go see the movies when they come out, and a handful of diehards dig into lore stuff in their own isolated corner of the world. It's not _that_ dissimilar from Star Wars in the '80s.
December 27, 2025 at 3:12 AM
Avatar just generally exists in the space franchise movies existed in pre-internet: Most people just go see the movies when they come out, and a handful of diehards dig into lore stuff in their own isolated corner of the world. It's not _that_ dissimilar from Star Wars in the '80s.
Reposted by Kyle Thayer
"Why do we have to take ethics classes for a computer science degree?"
This is why. And frankly, we should be failing more students for inadequately demonstrating their understanding of the topic.
Let's review some fun historical examples of why "No, actually quality will continue to matter."
This is why. And frankly, we should be failing more students for inadequately demonstrating their understanding of the topic.
Let's review some fun historical examples of why "No, actually quality will continue to matter."
December 26, 2025 at 3:32 PM
"Why do we have to take ethics classes for a computer science degree?"
This is why. And frankly, we should be failing more students for inadequately demonstrating their understanding of the topic.
Let's review some fun historical examples of why "No, actually quality will continue to matter."
This is why. And frankly, we should be failing more students for inadequately demonstrating their understanding of the topic.
Let's review some fun historical examples of why "No, actually quality will continue to matter."
Reposted by Kyle Thayer
many don’t remember but a long time ago the Bluesky app *did* promote a third party feed during onboarding.
maybe it’s time to revisit this?
here’s what didn’t quite work out
maybe it’s time to revisit this?
here’s what didn’t quite work out
the thing looming on my mind is this… will they attempt to acquihire the creator of For You? and would that be a good thing?
or is there a better, more non-traditional scenario where they promote/support/push For You while it remains independent?
or is there a better, more non-traditional scenario where they promote/support/push For You while it remains independent?
December 26, 2025 at 3:58 PM
many don’t remember but a long time ago the Bluesky app *did* promote a third party feed during onboarding.
maybe it’s time to revisit this?
here’s what didn’t quite work out
maybe it’s time to revisit this?
here’s what didn’t quite work out
Reposted by Kyle Thayer
i love my snowsky echo mini aftermath.site/digita-audio...
A Digital Audio Player Renewed My Love Of Music
Quitting Spotify is a little easier with my brand new digital audio player
aftermath.site
December 11, 2025 at 8:27 PM
i love my snowsky echo mini aftermath.site/digita-audio...
Reposted by Kyle Thayer
If you ever wanted to see what the generation of animation storytellers who were raised on The Maxx and created shows like Adventure Time and Steven Universe and more can do when dancing around each other in a tripartite creative fugue, then you should definitely go watch The Elephant.
December 24, 2025 at 4:06 AM
If you ever wanted to see what the generation of animation storytellers who were raised on The Maxx and created shows like Adventure Time and Steven Universe and more can do when dancing around each other in a tripartite creative fugue, then you should definitely go watch The Elephant.
Reposted by Kyle Thayer
iCloud, Mega, and as a torrent. Archivists have uploaded the 60 Minutes episode Bari Weiss spiked.
Archivists Posted the 60 Minutes CECOT Segment Bari Weiss Killed
iCloud, Mega, and as a torrent. Archivists have uploaded the 60 Minutes episode Bari Weiss spiked.
www.404media.co
December 23, 2025 at 12:15 AM
iCloud, Mega, and as a torrent. Archivists have uploaded the 60 Minutes episode Bari Weiss spiked.
Reposted by Kyle Thayer
She-Ra *and* Kipo? this is insane. they're netflix originals, removing them is akin to vaporizing them entirely.
‘She-Ra and the Princesses of Power’ is scheduled to leave Netflix on February 21, 2026.
(source: www.whats-on-netflix.com/leaving-soon...)
(source: www.whats-on-netflix.com/leaving-soon...)
December 16, 2025 at 10:29 PM
She-Ra *and* Kipo? this is insane. they're netflix originals, removing them is akin to vaporizing them entirely.
Reposted by Kyle Thayer
I've made a new diagram of the Oklab/Oklch color space, which I've put on Wikipedia.
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
December 21, 2025 at 10:59 PM
I've made a new diagram of the Oklab/Oklch color space, which I've put on Wikipedia.
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
Reposted by Kyle Thayer
Even there it depends *what* in law. LLMs are good where (1) you need to process huge amounts of natural language data (2) ask it questions that don't fit well to boolean logic (3) every true positive result is a win (4) false positive results can be (and are) quickly and accurately detected
December 22, 2025 at 4:27 PM
Even there it depends *what* in law. LLMs are good where (1) you need to process huge amounts of natural language data (2) ask it questions that don't fit well to boolean logic (3) every true positive result is a win (4) false positive results can be (and are) quickly and accurately detected
Reposted by Kyle Thayer
The key thing here: they shouldn’t use them for that! Trying to apply LLMs to all kinds of areas where they’re not appropriate is a tactic that the Big AI companies are advocating as part of their agenda, for obvious reasons. The most important part is not accepting the framing that they’re pushing.
suspect a big reason why many academics and others who work in areas where getting facts RIGHT is key are disinterested in using LLMs for research:
they’ve tried it, they keep noticing major errors in output, and they conclude that having to verify all that doesn’t actually save them time.
they’ve tried it, they keep noticing major errors in output, and they conclude that having to verify all that doesn’t actually save them time.
December 22, 2025 at 4:55 PM
The key thing here: they shouldn’t use them for that! Trying to apply LLMs to all kinds of areas where they’re not appropriate is a tactic that the Big AI companies are advocating as part of their agenda, for obvious reasons. The most important part is not accepting the framing that they’re pushing.
Reposted by Kyle Thayer
suspect a big reason why many academics and others who work in areas where getting facts RIGHT is key are disinterested in using LLMs for research:
they’ve tried it, they keep noticing major errors in output, and they conclude that having to verify all that doesn’t actually save them time.
they’ve tried it, they keep noticing major errors in output, and they conclude that having to verify all that doesn’t actually save them time.
December 21, 2025 at 4:16 PM
suspect a big reason why many academics and others who work in areas where getting facts RIGHT is key are disinterested in using LLMs for research:
they’ve tried it, they keep noticing major errors in output, and they conclude that having to verify all that doesn’t actually save them time.
they’ve tried it, they keep noticing major errors in output, and they conclude that having to verify all that doesn’t actually save them time.
Reposted by Kyle Thayer
Can you help me here. I feel I’m missing something. I keep hearing of academics coming across citations of fictitious works in reference. Surely this is gross professional misconduct comparable with plagiarism by the author. It should be reported to and dealt with by their employing institutions.
December 21, 2025 at 5:23 PM
Can you help me here. I feel I’m missing something. I keep hearing of academics coming across citations of fictitious works in reference. Surely this is gross professional misconduct comparable with plagiarism by the author. It should be reported to and dealt with by their employing institutions.
Reposted by Kyle Thayer
please let “bugs bunnying” become the slang for manipulating an LLM into ignoring its guardrails, i have never seen a more succinct and evocative term for it
Highly amused by the stories of companies using agentic AI chatbots to run their internal company stores, only for their employees to immediately Bugs Bunny the chatbot into giving them free stuff www.aiweirdness.com/when-a-chatb...
When a chatbot runs your store
You may have heard of "agentic AI", which is basically the idea that you can hook up a large language model to controls that do real things. The controls might run internet searches, run commands to o...
www.aiweirdness.com
December 21, 2025 at 4:20 PM
please let “bugs bunnying” become the slang for manipulating an LLM into ignoring its guardrails, i have never seen a more succinct and evocative term for it
I've made a new diagram of the Oklab/Oklch color space, which I've put on Wikipedia.
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
December 21, 2025 at 10:59 PM
I've made a new diagram of the Oklab/Oklch color space, which I've put on Wikipedia.
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
Oklab color space is intended to be perceptually uniform, meaning that it fits the weird shape of actual normal human perception. Oklch is the same space but measured in circles.
en.wikipedia.org/wiki/Oklab_c...
Reposted by Kyle Thayer
Once when we wanted to show the ineffectual dithering of The Good Place Committee, we had one of them say, "There are rules, procedures. This is The Good Place -- we can't just *do stuff.*" We almost didn't include it, because we thought it was a little on the nose.
WELKER: Massie and Khanna say they are looking at options including impeachment and contempt for DOJ officials. Would you support those steps?
KAINE: I think that's premature
KAINE: I think that's premature
December 21, 2025 at 5:31 PM
Once when we wanted to show the ineffectual dithering of The Good Place Committee, we had one of them say, "There are rules, procedures. This is The Good Place -- we can't just *do stuff.*" We almost didn't include it, because we thought it was a little on the nose.