Baldur Bjarnason
@baldurbjarnason.com
Writer, web developer and consultant based in Hveragerði, Iceland. Lapsed Interactive Media Academic. Webby Tech Stuff and webby book stuff.
https://www.baldurbjarnason.com/
https://softwarecrisis.dev/
https://www.baldurbjarnason.com/
https://softwarecrisis.dev/
Pinned
Baldur Bjarnason
@baldurbjarnason.com
· Oct 27
I figure since there's been a recent influx of people here on Bluesky, it might be worthwhile to do "highlights reels" of sorts for my newsletter, which is at www.baldurbjarnason.com
The biggest topic on my newsletter is software and web development, usually from a bit of a systemic perspective.
The biggest topic on my newsletter is software and web development, usually from a bit of a systemic perspective.
So this is happening. Any day now. Written and illustrated by @uglyreykjavik.bsky.social
November 10, 2025 at 4:20 PM
So this is happening. Any day now. Written and illustrated by @uglyreykjavik.bsky.social
Reposted by Baldur Bjarnason
Over these past decades, software has coached us into thinking like a computer. Not necessarily on purpose, but because the software wasn't designed to be usable by thinking like a human.
This side effect was a net positive for libertarians, who leapt from anti-human software to an anti-human world
This side effect was a net positive for libertarians, who leapt from anti-human software to an anti-human world
Paulina Borsook "warned that tech libertarians wanted an anti-human world that worked more like a computer. From Cyberselfish: A Critical Romp Through The Terribly Libertarian Culture of High Tech, a book based on her 1990s writing" www.thenerdreich.com/she-warned-a...
November 10, 2025 at 10:39 AM
Over these past decades, software has coached us into thinking like a computer. Not necessarily on purpose, but because the software wasn't designed to be usable by thinking like a human.
This side effect was a net positive for libertarians, who leapt from anti-human software to an anti-human world
This side effect was a net positive for libertarians, who leapt from anti-human software to an anti-human world
Reposted by Baldur Bjarnason
Generative AI is the apotheosis of this process. It's praised for doing better than a human on certain tasks, but those tasks were *made for computer logic* in the first place.
The new world will try to convince you that this computer logic is more valuable than your human thought. Defy it.
The new world will try to convince you that this computer logic is more valuable than your human thought. Defy it.
November 10, 2025 at 10:42 AM
Generative AI is the apotheosis of this process. It's praised for doing better than a human on certain tasks, but those tasks were *made for computer logic* in the first place.
The new world will try to convince you that this computer logic is more valuable than your human thought. Defy it.
The new world will try to convince you that this computer logic is more valuable than your human thought. Defy it.
Reposted by Baldur Bjarnason
After the Rubygems drama, I felt frustrated that I can’t escape evil people trying to dominate my life.
I found myself mourning the future I thought we would have; but also hopeful that it’s not too late to start choosing kindness.
Join me in building a better future:
okayfail.com/2025/in-prai...
I found myself mourning the future I thought we would have; but also hopeful that it’s not too late to start choosing kindness.
Join me in building a better future:
okayfail.com/2025/in-prai...
In Praise of dhh
A reflection on Ruby's past, present, and future.
okayfail.com
November 9, 2025 at 4:52 PM
After the Rubygems drama, I felt frustrated that I can’t escape evil people trying to dominate my life.
I found myself mourning the future I thought we would have; but also hopeful that it’s not too late to start choosing kindness.
Join me in building a better future:
okayfail.com/2025/in-prai...
I found myself mourning the future I thought we would have; but also hopeful that it’s not too late to start choosing kindness.
Join me in building a better future:
okayfail.com/2025/in-prai...
“Brussels knifes privacy to feed the AI boom – POLITICO”
https://www.politico.eu/article/brussels-knifes-privacy-to-feed-the-ai-boom-gdpr-digital-omnibus/
Well, this isn't good
https://www.politico.eu/article/brussels-knifes-privacy-to-feed-the-ai-boom-gdpr-digital-omnibus/
Well, this isn't good
November 10, 2025 at 10:46 AM
“Brussels knifes privacy to feed the AI boom – POLITICO”
https://www.politico.eu/article/brussels-knifes-privacy-to-feed-the-ai-boom-gdpr-digital-omnibus/
Well, this isn't good
https://www.politico.eu/article/brussels-knifes-privacy-to-feed-the-ai-boom-gdpr-digital-omnibus/
Well, this isn't good
Reposted by Baldur Bjarnason
I spotted this on Mastodon and I find it horrible, not least for the speed with which this has happened.
November 4, 2025 at 9:56 AM
I spotted this on Mastodon and I find it horrible, not least for the speed with which this has happened.
Reposted by Baldur Bjarnason
It is shocking how audio transcription, translation etc have gotten SO MUCH worse in the past year, and there is really no way else but down for the companies
November 10, 2025 at 8:18 AM
It is shocking how audio transcription, translation etc have gotten SO MUCH worse in the past year, and there is really no way else but down for the companies
Reposted by Baldur Bjarnason
On the topic of speech-to-text: Norway's national broadcaster NRK is cutting pay for subtitlers bc they're replacing manual translation w/ machine outputs
Absolutely inevitable trajectory: results that get worse and worse and worse as companies have to exponentially enshittify with worsening slop
Absolutely inevitable trajectory: results that get worse and worse and worse as companies have to exponentially enshittify with worsening slop
NRK kutter i honoraret til tekstere – skal bruke mer KI
NRK skal effektivisere tekstingen ved hjelp av kunstig intelligens (KI). For teksterne betyr det et heftig kutt i honorarene.
www.kampanje.com
November 10, 2025 at 8:17 AM
On the topic of speech-to-text: Norway's national broadcaster NRK is cutting pay for subtitlers bc they're replacing manual translation w/ machine outputs
Absolutely inevitable trajectory: results that get worse and worse and worse as companies have to exponentially enshittify with worsening slop
Absolutely inevitable trajectory: results that get worse and worse and worse as companies have to exponentially enshittify with worsening slop
Reposted by Baldur Bjarnason
David Gerard is a sophisticated techno-skeptic and an engaging, often funny, writer. For years he's been documenting cryptocurrency scams (and the ur-scam that is crypto itself). Now he's onto AI & its hype. I highly recommend the work of this curmudgeonly contrarian. He's the real deal.
AI gets 45% of news wrong — but readers still trust it
no plans to fix it, though
www.youtube.com/watch?v=_dOP... - video
pivottoai.libsyn.com/20251105-ai-... - podcast
time: 4 min 58 sec
no plans to fix it, though
www.youtube.com/watch?v=_dOP... - video
pivottoai.libsyn.com/20251105-ai-... - podcast
time: 4 min 58 sec
November 5, 2025 at 10:55 PM
David Gerard is a sophisticated techno-skeptic and an engaging, often funny, writer. For years he's been documenting cryptocurrency scams (and the ur-scam that is crypto itself). Now he's onto AI & its hype. I highly recommend the work of this curmudgeonly contrarian. He's the real deal.
Reposted by Baldur Bjarnason
Instead, "AI" is completely in the thrall of billionaires and TESCREALists -- how else to explain anyone taking the paper-shaped objects out of corporate labs seriously?
>>
>>
November 9, 2025 at 3:03 PM
Instead, "AI" is completely in the thrall of billionaires and TESCREALists -- how else to explain anyone taking the paper-shaped objects out of corporate labs seriously?
>>
>>
Reposted by Baldur Bjarnason
At NeurIPS is important -- it's not like we were publishing in some other venue, or just posting commentary on the internet.
But "AI" isn't a scientific, academic field anymore.
>>
But "AI" isn't a scientific, academic field anymore.
>>
November 9, 2025 at 3:01 PM
At NeurIPS is important -- it's not like we were publishing in some other venue, or just posting commentary on the internet.
But "AI" isn't a scientific, academic field anymore.
>>
But "AI" isn't a scientific, academic field anymore.
>>
Reposted by Baldur Bjarnason
Five years ago, at NeurIPS, in Raji et al 2021 we told them so:
datasets-benchmarks-proceedings.neurips.cc/paper/2021/h...
datasets-benchmarks-proceedings.neurips.cc/paper/2021/h...
November 9, 2025 at 3:01 PM
Five years ago, at NeurIPS, in Raji et al 2021 we told them so:
datasets-benchmarks-proceedings.neurips.cc/paper/2021/h...
datasets-benchmarks-proceedings.neurips.cc/paper/2021/h...
Reposted by Baldur Bjarnason
I appreciate the work of these authors to show that this problem not only is still here but has grown:
www.nbcnews.com/tech/tech-ne...
But it is also quite frustrating 🧵>>
www.nbcnews.com/tech/tech-ne...
But it is also quite frustrating 🧵>>
AI's capabilities may be exaggerated by flawed tests, according to new study
A study from the Oxford Internet Institute analyzed 445 tests used to evaluate AI models.
www.nbcnews.com
November 9, 2025 at 2:59 PM
I appreciate the work of these authors to show that this problem not only is still here but has grown:
www.nbcnews.com/tech/tech-ne...
But it is also quite frustrating 🧵>>
www.nbcnews.com/tech/tech-ne...
But it is also quite frustrating 🧵>>
✋🏻 Never used any of those chatbots or copilots. That they would be more detrimental than not and that they were cognitive hazards was pretty clear from the outset
Hands up if you've never used Chat GPT ✋
(I feel like Dozer and Tank in The Matrix right now - at first I didn't use it because, rather ironically, I'm lazy and stubborn (peak Taurus energy there) - literally no, don't make me use the new thing I don't wanna. Now I'm glad I didn't 😅)
(I feel like Dozer and Tank in The Matrix right now - at first I didn't use it because, rather ironically, I'm lazy and stubborn (peak Taurus energy there) - literally no, don't make me use the new thing I don't wanna. Now I'm glad I didn't 😅)
Probably a good way to tell right now if the job you’re applying for is run by absolute dumbfucks is to ask if they’re using AI.
November 9, 2025 at 12:29 PM
✋🏻 Never used any of those chatbots or copilots. That they would be more detrimental than not and that they were cognitive hazards was pretty clear from the outset
Reposted by Baldur Bjarnason
Science will never succeed until we recognize the humanity of all scientists and break the dehumanization cultures that crush too many people.
November 8, 2025 at 7:23 PM
Science will never succeed until we recognize the humanity of all scientists and break the dehumanization cultures that crush too many people.
“Bullet Points: November 7, 2025 edition - by Dave Karpf”
https://davekarpf.substack.com/p/bullet-points-november-7-2025-edition
The predictions he wrote last year are depressing to read today
https://davekarpf.substack.com/p/bullet-points-november-7-2025-edition
The predictions he wrote last year are depressing to read today
November 9, 2025 at 11:02 AM
“Bullet Points: November 7, 2025 edition - by Dave Karpf”
https://davekarpf.substack.com/p/bullet-points-november-7-2025-edition
The predictions he wrote last year are depressing to read today
https://davekarpf.substack.com/p/bullet-points-november-7-2025-edition
The predictions he wrote last year are depressing to read today
Reposted by Baldur Bjarnason
Remarkable that every boss talks about how amazing AI is, all the way up until that AI's outputs hit people whose jobs don't depend on agreeing, at which point they inevitably have to conceal that they are using AI.
The shitty Coke ad studio. Games that use AI generated assets. Support chatbots.
The shitty Coke ad studio. Games that use AI generated assets. Support chatbots.
The studio that made the shitty coca-cola AI ad getting their teeth kicked in on socials bc of how icky bad it is, then the studio CEO coming out insulting and dismissing folks kicking their teeth in, and THEN releasing the worst ever cover up “hey we actually made this by hand” scam vid
🧍♀️
🧍♀️
November 9, 2025 at 3:24 AM
Remarkable that every boss talks about how amazing AI is, all the way up until that AI's outputs hit people whose jobs don't depend on agreeing, at which point they inevitably have to conceal that they are using AI.
The shitty Coke ad studio. Games that use AI generated assets. Support chatbots.
The shitty Coke ad studio. Games that use AI generated assets. Support chatbots.
Reposted by Baldur Bjarnason
Tech guys six months ago: haha yes we’re cutting all this WASTEFUL spending by eliminating medical research and USAID
Tech guys now: yes I think taxpayers will be excited to bailout my non consensual pornography machine
Tech guys now: yes I think taxpayers will be excited to bailout my non consensual pornography machine
November 8, 2025 at 11:02 PM
Tech guys six months ago: haha yes we’re cutting all this WASTEFUL spending by eliminating medical research and USAID
Tech guys now: yes I think taxpayers will be excited to bailout my non consensual pornography machine
Tech guys now: yes I think taxpayers will be excited to bailout my non consensual pornography machine
Reposted by Baldur Bjarnason
LLMs are trained on public sources of code examples. The vast majority of those examples are poorly-written. This is why LLMs are so good at generating poorly-written code.
So, when someone says they're *impressed* by LLM-generated code...
So, when someone says they're *impressed* by LLM-generated code...
November 9, 2025 at 6:46 AM
LLMs are trained on public sources of code examples. The vast majority of those examples are poorly-written. This is why LLMs are so good at generating poorly-written code.
So, when someone says they're *impressed* by LLM-generated code...
So, when someone says they're *impressed* by LLM-generated code...
Reposted by Baldur Bjarnason
if you terfs actually spent time reading radical feminists instead of selectively quoting them you’d realise, very quickly, that ‘gender critical’ is just bigotry rooted in patriarchy.
here is catherine mackinnon:
here is catherine mackinnon:
![Much of the current debate has centered on (endlessly obsessed over, actually) whether trans women are women. Honestly, seeing “women” as a turf to be defended, as opposed to a set of imperatives and limitations to be criticized, challenged, changed, or transcended, has been pretty startling. One might think that trans women—assigned male at birth, leaving masculinity behind, drawn to and embracing womanhood for themselves—would be welcomed. Yet a group of philosophers purporting feminism slide sloppily from “female sex” through “feminine gender” straight to “women” as if no move has been made,[4] eventually reverting to the dictionary: a woman is an “adult human female.”[5] Defining women by biology—adult is biological age, human is biological species, female is biological sex—used to be criticized as biological essentialism. Those winging to the Right are thrilled by this putatively feminist reduction of women to female body parts, preferably chromosomes and reproductive apparatus, qualities chosen so that whatever is considered definitive of sex is not only physical but cannot be physically changed into.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:bcgtpezwhu5ynanpiezkmfim/bafkreigw3lotshdxaqfr6b42ttj2keuc4dywkwtbjag7fvnxlm5lonzr54@jpeg)
December 28, 2024 at 12:13 PM
if you terfs actually spent time reading radical feminists instead of selectively quoting them you’d realise, very quickly, that ‘gender critical’ is just bigotry rooted in patriarchy.
here is catherine mackinnon:
here is catherine mackinnon:
Reposted by Baldur Bjarnason
But if you pay attention to any artist, they're like "oh yeah I chose that color because I was out of green that day" or "yeah that stunt was actually a mistake we kept in" or "the actor misspoke and we rolled with it."
Recalcitrance is not a virtue and it does not make you a worthier artist
Recalcitrance is not a virtue and it does not make you a worthier artist
November 9, 2025 at 2:35 AM
But if you pay attention to any artist, they're like "oh yeah I chose that color because I was out of green that day" or "yeah that stunt was actually a mistake we kept in" or "the actor misspoke and we rolled with it."
Recalcitrance is not a virtue and it does not make you a worthier artist
Recalcitrance is not a virtue and it does not make you a worthier artist
Reposted by Baldur Bjarnason
It's always weird to me when people are like "I'm an uncompromising artist" and then the famous story about any great artist is how they compromised. Like Akira Kurosawa being asked about a shot and going "well, there was a highway on one side and a factory on the other, so that's all I could shoot"
November 9, 2025 at 2:31 AM
It's always weird to me when people are like "I'm an uncompromising artist" and then the famous story about any great artist is how they compromised. Like Akira Kurosawa being asked about a shot and going "well, there was a highway on one side and a factory on the other, so that's all I could shoot"
Reposted by Baldur Bjarnason
The worst thing about this is that Coca-Cola is a rich company. Its market cap is £300bn, it has nearly £11bn cash on hand. The cost of real animators wouldn't even be a rounding error. They have literally no reason to use AI to make their Christmas ad.
Mirroring the video here if you don't want to even humor going onto Twitter, but when it gets into the "process" section, there's so many red flags of them lying about the work they put into the commercial that reveal how much of this video is them trying and lying after the fact.
November 9, 2025 at 9:10 AM
The worst thing about this is that Coca-Cola is a rich company. Its market cap is £300bn, it has nearly £11bn cash on hand. The cost of real animators wouldn't even be a rounding error. They have literally no reason to use AI to make their Christmas ad.
Reposted by Baldur Bjarnason
Perhaps try reading this, written by Sharon Begley in 2021 before she died, as a counter to all the "Oh, Watson wasn't really sexist or racist. His denigration of Franklin was a bit to show that biology was fun, and hey, he went to Africa!" hagiographies.
www.statnews.com/2025/11/07/j...
www.statnews.com/2025/11/07/j...
James Watson, dead at 97, was a scientific legend and a pariah among his peers
James Watson, the co-discoverer of the structure of DNA who died Thursday at 97, was a scientific legend and a pariah among his peers.
www.statnews.com
November 9, 2025 at 9:27 AM
Perhaps try reading this, written by Sharon Begley in 2021 before she died, as a counter to all the "Oh, Watson wasn't really sexist or racist. His denigration of Franklin was a bit to show that biology was fun, and hey, he went to Africa!" hagiographies.
www.statnews.com/2025/11/07/j...
www.statnews.com/2025/11/07/j...
Reposted by Baldur Bjarnason
by the way, I've written a couple of things on the "western civilization is killed by suicidal empathy" line of reactionary thought. start with the Three Body Problem fans.
madeinchinajournal.com/2023/12/11/t...
madeinchinajournal.com/2023/12/11/t...
November 8, 2025 at 12:49 PM
by the way, I've written a couple of things on the "western civilization is killed by suicidal empathy" line of reactionary thought. start with the Three Body Problem fans.
madeinchinajournal.com/2023/12/11/t...
madeinchinajournal.com/2023/12/11/t...