You can remove tailwind-merge, clsx, cva, etc by moving more of your styling to data attributes. Add data-slot=”componentName" to all your UI components, so you can override default styles with *:data-[slot=componentName]:p-8. For things like button variants, you can add data-variant={props.variant}
December 24, 2025 at 4:58 PM
You can remove tailwind-merge, clsx, cva, etc by moving more of your styling to data attributes. Add data-slot=”componentName" to all your UI components, so you can override default styles with *:data-[slot=componentName]:p-8. For things like button variants, you can add data-variant={props.variant}
Reposted by Eva
This is the basic flaw in almost every generative AI use case. If quality matters at all, the quality control makes it not worth the trip.
December 23, 2025 at 5:46 AM
This is the basic flaw in almost every generative AI use case. If quality matters at all, the quality control makes it not worth the trip.
Reposted by Eva
by far the funniest use case for AI is “I use it for powerpoints because I hate making them and nobody looks at them” because it kinda sounds like the real workflow savings would be to stop making powerpoints
December 23, 2025 at 2:10 AM
by far the funniest use case for AI is “I use it for powerpoints because I hate making them and nobody looks at them” because it kinda sounds like the real workflow savings would be to stop making powerpoints
Reposted by Eva
The reverse correlation of how mgmt thinks of people vs tech amazes me.
The tech can make a ton of mistakes, but that's OK, it will magically get better someday. To hire a person they must prove they have solved your exact problem before and 100% guaranteed to be successful on day one.
The tech can make a ton of mistakes, but that's OK, it will magically get better someday. To hire a person they must prove they have solved your exact problem before and 100% guaranteed to be successful on day one.
December 23, 2025 at 5:34 PM
The reverse correlation of how mgmt thinks of people vs tech amazes me.
The tech can make a ton of mistakes, but that's OK, it will magically get better someday. To hire a person they must prove they have solved your exact problem before and 100% guaranteed to be successful on day one.
The tech can make a ton of mistakes, but that's OK, it will magically get better someday. To hire a person they must prove they have solved your exact problem before and 100% guaranteed to be successful on day one.
Reposted by Eva
New from me: The storied fiddler Ashley MacIsaac wanted to play a show at a Nova Scotia First Nation. It got cancelled because Google's AI search summary incorrectly called him a sex offender. Now he's worried for his safety www.theglobeandmail.com/culture/arti...
Fiddler Ashley MacIsaac has show cancelled over Google AI-generated misinformation
Cape Breton musician was set to perform in Nova Scotia before a search summary incorrectly described him as a sex offender
www.theglobeandmail.com
December 23, 2025 at 4:27 PM
New from me: The storied fiddler Ashley MacIsaac wanted to play a show at a Nova Scotia First Nation. It got cancelled because Google's AI search summary incorrectly called him a sex offender. Now he's worried for his safety www.theglobeandmail.com/culture/arti...
It's prefers-reduced-motion not prefers-no-animation. Do not globally remove animations for prefers-reduced-motion. Instead selectively apply transform animations when prefers-reduced-motion is not set. When it is set, use an animation that doesn't depend on motion, e.g. a fade
December 23, 2025 at 4:05 PM
It's prefers-reduced-motion not prefers-no-animation. Do not globally remove animations for prefers-reduced-motion. Instead selectively apply transform animations when prefers-reduced-motion is not set. When it is set, use an animation that doesn't depend on motion, e.g. a fade
Reposted by Eva
It's not really an AI issue, it's a training database issue (a classic). Which means the bias is actually human--the people who made the training database for the AI failed to provide enough samples across race (2/?)
December 21, 2025 at 12:02 AM
It's not really an AI issue, it's a training database issue (a classic). Which means the bias is actually human--the people who made the training database for the AI failed to provide enough samples across race (2/?)
Reposted by Eva
Ok so
The headline is a bit misleading. AI shows racist bias in that psychiatric model (not surprising)
The slides issue is more complex--the AI detects markers that are correlated with race and then can't diagnose because it's been trained on too many white ppl (1/?)
futurism.com/health-medic...
The headline is a bit misleading. AI shows racist bias in that psychiatric model (not surprising)
The slides issue is more complex--the AI detects markers that are correlated with race and then can't diagnose because it's been trained on too many white ppl (1/?)
futurism.com/health-medic...
Doctors Catch Cancer-Diagnosing AI Extracting Patients' Race Data and Being Racist With It
A new study reveals an alarming bias present in four leading AI cancer screening tools, related to race, gender, and age.
futurism.com
December 21, 2025 at 12:01 AM
Ok so
The headline is a bit misleading. AI shows racist bias in that psychiatric model (not surprising)
The slides issue is more complex--the AI detects markers that are correlated with race and then can't diagnose because it's been trained on too many white ppl (1/?)
futurism.com/health-medic...
The headline is a bit misleading. AI shows racist bias in that psychiatric model (not surprising)
The slides issue is more complex--the AI detects markers that are correlated with race and then can't diagnose because it's been trained on too many white ppl (1/?)
futurism.com/health-medic...
Reposted by Eva
I think this post nails the actual problem, for researchers at least—AI hallucinations would simply not be a problem in academic work if we’d not normalized citation-as-signaling rather than actual engagement—you can only cite a fake paper if you’re not in the habit of reading the papers you cite
Do not cite an academic paper unless you’ve read it
AI Is Inventing Academic Papers That Don't Exist -- And They're Being Cited in Real Journals
Academic articles from authors using large language model are creating an ecosystem of fake research that threatens human knowledge itself.
www.rollingstone.com
December 19, 2025 at 7:01 PM
I think this post nails the actual problem, for researchers at least—AI hallucinations would simply not be a problem in academic work if we’d not normalized citation-as-signaling rather than actual engagement—you can only cite a fake paper if you’re not in the habit of reading the papers you cite
Reposted by Eva
There are always empty words about "human in the loop", that you should always check the work of an LLM. But the entire value of a summary is as a replacement for reading. Nobody uses an LLM to generate a "summary" *and* independently verifies.
December 21, 2025 at 3:35 AM
There are always empty words about "human in the loop", that you should always check the work of an LLM. But the entire value of a summary is as a replacement for reading. Nobody uses an LLM to generate a "summary" *and* independently verifies.
It tickles me every time I think of the HTML5 semantic names for the bold, italics, and underline elements. Bring attention to, Idiomatic text, and Unarticulated annotation. True literary genius 🤌🏿
December 21, 2025 at 4:55 PM
It tickles me every time I think of the HTML5 semantic names for the bold, italics, and underline elements. Bring attention to, Idiomatic text, and Unarticulated annotation. True literary genius 🤌🏿
Reposted by Eva
❓"When can we expect this feature"
Please refrain from asking this question towards open source maintainers. You generally shouldn't "expect" anything. I don't often agree with DHH but his take on OSS being a gift, not something you're entitled to, is something I relate with.
Please refrain from asking this question towards open source maintainers. You generally shouldn't "expect" anything. I don't often agree with DHH but his take on OSS being a gift, not something you're entitled to, is something I relate with.
December 20, 2025 at 8:42 AM
❓"When can we expect this feature"
Please refrain from asking this question towards open source maintainers. You generally shouldn't "expect" anything. I don't often agree with DHH but his take on OSS being a gift, not something you're entitled to, is something I relate with.
Please refrain from asking this question towards open source maintainers. You generally shouldn't "expect" anything. I don't often agree with DHH but his take on OSS being a gift, not something you're entitled to, is something I relate with.
Every time a USAmerican compares genAI to cars or the highway system, it only serves to further convince me that gen AI is a net negative. I want trains not ChatGPT.
December 20, 2025 at 3:33 PM
Every time a USAmerican compares genAI to cars or the highway system, it only serves to further convince me that gen AI is a net negative. I want trains not ChatGPT.
Reposted by Eva
You know what would be cool? If some of these guys tried to replace the internet with a simple-answers machine where they control what's inside, how you access it, and what it's allowed to say. I can't imagine any problems with that! Maybe we should use their output to build all future software?
The 6 richest men in the world are all Trump allies.
They control your algorithms and data on X, TikTok, Twitch, Instagram, Facebook, WhatsApp, Threads, and YouTube.
Just in case you’re wondering why more isn’t done to stop right-wing bots and misinformation on social media.
They control your algorithms and data on X, TikTok, Twitch, Instagram, Facebook, WhatsApp, Threads, and YouTube.
Just in case you’re wondering why more isn’t done to stop right-wing bots and misinformation on social media.
December 19, 2025 at 8:26 PM
You know what would be cool? If some of these guys tried to replace the internet with a simple-answers machine where they control what's inside, how you access it, and what it's allowed to say. I can't imagine any problems with that! Maybe we should use their output to build all future software?
Reposted by Eva
I don't mean to come across as snide, but I'm tired of hearing what AI *could* do. I want to hear what it's doing that makes it worth keeping around.
Steve Jobs didn't need to go around telling people what Apple would be in 5 or 10 years. He just said what awesome thing it was doing now.
Steve Jobs didn't need to go around telling people what Apple would be in 5 or 10 years. He just said what awesome thing it was doing now.
December 19, 2025 at 10:15 PM
I don't mean to come across as snide, but I'm tired of hearing what AI *could* do. I want to hear what it's doing that makes it worth keeping around.
Steve Jobs didn't need to go around telling people what Apple would be in 5 or 10 years. He just said what awesome thing it was doing now.
Steve Jobs didn't need to go around telling people what Apple would be in 5 or 10 years. He just said what awesome thing it was doing now.
Reposted by Eva
“Love it or hate it it’s here to stay” is a crazy thing to keep saying about a technology. If people don’t like using something it doesn’t have to stay we’re not talking about bad weather here
December 19, 2025 at 7:37 AM
“Love it or hate it it’s here to stay” is a crazy thing to keep saying about a technology. If people don’t like using something it doesn’t have to stay we’re not talking about bad weather here
Reposted by Eva
also designing and writing code is the fun part. make an AI that goes chasing down half a dozen stakeholders and gets them to agree on requirements or something.
December 19, 2025 at 11:51 PM
also designing and writing code is the fun part. make an AI that goes chasing down half a dozen stakeholders and gets them to agree on requirements or something.
Reposted by Eva
As I said, LLM proponents tend to be dismissive whenever it's suggested that they try to measure their own productivity with vs. without LLM assistance
Thanks for the illustration, I guess?
Thanks for the illustration, I guess?
December 19, 2025 at 11:48 PM
As I said, LLM proponents tend to be dismissive whenever it's suggested that they try to measure their own productivity with vs. without LLM assistance
Thanks for the illustration, I guess?
Thanks for the illustration, I guess?
Reposted by Eva
the conversation we *should* be having with every business/institution leveraging AI to screw labor is: where's MY discount, now that you're cutting costs?
why isn't tuition coming down? why aren't products cheaper? video games? movies? doctors? lawyers?
push HARD on this, relentlessly!
why isn't tuition coming down? why aren't products cheaper? video games? movies? doctors? lawyers?
push HARD on this, relentlessly!
The Information newsletter: Can UCLA Replace Teaching Assistants with AI?
no but they're gonna anyway
what your tuition buys
no but they're gonna anyway
what your tuition buys
![Miller’s team also recently used software from startup StackAI to develop an AI-powered app that writes letters of recommendation, saving faculty members time. Faculty type basic details about a student who has requested a letter, such as their grades and accomplishments, and the app writes a draft of the full letter.
AI is “one of those things that you might worry could dehumanize the process of writing recommendation letters, but faculty also say that process [of manually writing the letters] is very labor intensive,” Miller said. “So far they’ve gotten a lot out of” the new app.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:2aiipqemoobrrilfx4jbm6pb/bafkreidwkhezhxghoctd5kb7nbbqqkhfhylpmnfbyjtycjtuagzi5dodqu@jpeg)
December 19, 2025 at 2:23 PM
the conversation we *should* be having with every business/institution leveraging AI to screw labor is: where's MY discount, now that you're cutting costs?
why isn't tuition coming down? why aren't products cheaper? video games? movies? doctors? lawyers?
push HARD on this, relentlessly!
why isn't tuition coming down? why aren't products cheaper? video games? movies? doctors? lawyers?
push HARD on this, relentlessly!
Do not write negated test names. Negated test names are too broad. Test names should include the behavior you expect, not the behavior you don’t expect
December 19, 2025 at 3:33 PM
Do not write negated test names. Negated test names are too broad. Test names should include the behavior you expect, not the behavior you don’t expect
Reposted by Eva
By rebranding, these projects end up feeding into the ongoing "AI hype," that, as Whittaker notes, presents AI as a multitool of efficiency which can be applied across a variety of domains. This rebranding to survive has the effect of enabling and maintaining the conditions of AI hype.
December 18, 2025 at 12:03 AM
By rebranding, these projects end up feeding into the ongoing "AI hype," that, as Whittaker notes, presents AI as a multitool of efficiency which can be applied across a variety of domains. This rebranding to survive has the effect of enabling and maintaining the conditions of AI hype.
Reposted by Eva
sheesh, the only thing worse than getting kicked out of the polycule is finding out via jira ticket 😣
![Screenshot of an email:
Subject: [PIPEXP-1192] Login Access Request
To: me
Good morning Kimberly! =^uwu^=;;
I’m sorry to hear that you’re having trouble logging into Jellyfin! It looks like your SSO account was indeed deactivated on 2025 Dec 15.
According to our records,
1. On or about 2025 Nov 15, you were broken up with by ___, who noted this as a “final break” on vis event report sheet. I’m also seeing that Ve ticked the “Unilateral” box, meaning it isn’t contestable.
2. Our Salesforce shows you aren’t close to any other polycule members
Unfortunately, our solo-poly access policy doesn’t allow us to extend resources to unpartnered members after the 30 day grace period following your break. This includes entertainment resources (Jellyfin / *darr / nzbget / Pornhub for Enterprise™ seat), as well as O365™, Adobe CC™, and shell access on the Nix clusters. We’re so sorry for your loss, cutie!! 😭💜
If you have any questions, you can reply to this email to generate another ticket, or you can contact zzzoeee@ from our Partner Infra team (note: three Zs and three Es in their LDAP, we have a lot of Zoes here!)
—-
Regards,
___ (🏳️⚧️she/her, UTC -5)
Partner Infra Platform Experiences team
As a courtesy, a copy of this email has been sent to your manager listed in Workday™
New name or pronouns? Dating someone new? Just visit go/update-me to update your account!
Include this line in your email to comment on the ticket: 250e084a-b937-47a6-948d-b519112ad35a](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:rhtazgvloflondzburvubdft/bafkreids254lgwtdyd62paodq5ynb5tatkemp5l77ynw6udhshedj6pab4@jpeg)
December 17, 2025 at 2:49 PM
sheesh, the only thing worse than getting kicked out of the polycule is finding out via jira ticket 😣
Reposted by Eva
Don't come at me with that "software engineer" bullshit until I hear you calling apps "software engines"
December 18, 2025 at 3:02 PM
Don't come at me with that "software engineer" bullshit until I hear you calling apps "software engines"
Reposted by Eva
Wow, if you've ever gotten upset about having to support some browsers that are a couple years out of date - Google search apparently _just_ started rejecting the Dreamcast browser!
www.timeextension.com/news/2025/12...
www.timeextension.com/news/2025/12...
After 25 Years, Google Has Finally Killed Dreamcast Web Browser Support
Sad times
www.timeextension.com
December 18, 2025 at 2:08 PM
Wow, if you've ever gotten upset about having to support some browsers that are a couple years out of date - Google search apparently _just_ started rejecting the Dreamcast browser!
www.timeextension.com/news/2025/12...
www.timeextension.com/news/2025/12...
Reposted by Eva
I don't know how many times this needs to be said, but apparently a lot: If you are not in a position to verify the accuracy of some synthetic text, the synthetic text is not useful/has a high potential to be misleading.
This is one of those cases.
This is one of those cases.
The rationale I was given for AI Summaries *alongside* ordinary abstracts:
* The AI-generated summaries are intended to be more accessible to non-experts.
(1/3)
* The AI-generated summaries are intended to be more accessible to non-experts.
(1/3)
December 18, 2025 at 3:39 AM
I don't know how many times this needs to be said, but apparently a lot: If you are not in a position to verify the accuracy of some synthetic text, the synthetic text is not useful/has a high potential to be misleading.
This is one of those cases.
This is one of those cases.