tante

Sociotechnologist, writer and speaker working on tech and its social impact. #Communist. #Feminist. #Antifascist. #Luddite. Email: [email protected] | License CC BY-SA-4.0 #noAI

"Ein-Mann-Gegenkultur" (SPIEGEL)

Public Key

npub1dedm8axlwewsysrxlw93kz43j6mgwf4cpync9gu8e95cl5za3s6s4gtq8a

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsxukan7n0hvhgzgpn0hzcmp2cedd58y6uqjfuz5wruj6v06pwccdgpz4mhxue69uhhyetvv9ujuerfw36x7tnsw43q9z66gm

Publishing to

relay.ditto.pub

Author Public Key

npub1dedm8axlwewsysrxlw93kz43j6mgwf4cpync9gu8e95cl5za3s6s4gtq8a

Show more details

Last Notes

2026-07-14 16:45:34 GMT

- reply

@nprofile…3c9j you have to do marginal setup. Look here https://rknight.me/blog/setting-up-mastodon-author-tags/

2026-07-08 08:29:49 GMT

- reply

@nprofile…09mr They don't want to host that stuff (and pay for it). Which is totally legitimate. Microsoft takes all that code. Why does Codeberg need to accept it?

2026-07-07 18:02:00 GMT

If you are a member of Codeberg e.V. please take the time to participate in the poll that was just sent out about banning vibe-coded projects on Codeberg.
Please agree to the proposal. Slop can live on GitHub.

2026-06-18 07:35:21 GMT

There is a massive campaign to delegitimize protest about "AI"/data center water usage as "fake". But I wonder, is there more "substance" that the effective altruist blog post by Andy Masley as a basis?
(Link to post, but be aware it's on the Nazi platform Substack https://blog.andymasley.com/p/the-ai-water-issue-is-fake )

2026-06-02 08:00:20 GMT

"The SpaceX IPO is great for Elon Musk and terrible for you
- The biggest public offering ever is financial nihilism’s final form."
The SpaceX IPO and the way it's being shoved into all indices to force ETFs etc zu buy SpaceX stock is criminal. People will lose their savings.
https://www.theverge.com/ai-artificial-intelligence/940001/elon-musk-spacex-ipo-ai

2026-06-01 08:57:21 GMT

The `rsync` thing really makes me sad.

2026-05-18 13:43:49 GMT

Bei der ZEIT diskutieren Sascha Lobo, @nprofile…54m9 und @nprofile…0qc2 über das Internet. Das meiste ist nicht überraschend, aber ich finde es erfrischend entlarvend, wie der Boomer-Influencer Lobo seine "Privatwirtschaft über alles, weil die meine Talks zahlen" Nummer runternudelt.
Lobo: "Ihr wollt eine Welt ohne Big Tech – aber ich will europäisches Big Tech."
Ja digga, dann bist du halt nicht Teil der Lösung.
https://archive.ph/lisr2#selection-2185.418-2191.39

2026-05-15 17:36:41 GMT

"But you can't deny that it's useful" is what I hear often when I criticize "AI". But that's not the argument because everything is useful to someone for some purpose usually at someone's cost. "Child labor" is _useful_.
Question is: Do the uses bring more social value than the cost?

2026-05-12 13:50:55 GMT

- reply

@nprofile…nv0d @nprofile…k0ra that really depends on which features you use. If you basically just use git hosting and maybe issues it's reasonably easy, there are migration scripts (which are of course never 100% perfect). The behavior of CI pipelines is somewhat differently from when I last looked at things. Also: If you use some of the more project management oriented gitlab features, forgejo lacks some of them.

2026-05-12 07:16:35 GMT

If you're still on GitLab you might want to speed up planning your migration path.

2026-05-07 11:44:57 GMT

"There are no more juniors. There was a funeral for their passing in 2024. Nobody came. The machine does what they do now, but cheaper. Of course, juniors weren't valuable for what they produced, they were valuable for who they would become: the senior engineer who knows where the bodies are buried. We optimized for output, and abolished apprenticeship. A few years from now, we'll wonder where all the seniors are. We shot them. Nobody will remember."
https://www.stvn.sh/writing/programming-still-sucks-fqffhyp

2026-05-07 10:00:05 GMT

Is OpenAI's model showing signs of model collapse?
(Original title: OpenAI ChatGPT goes goblin mode — let none say ‘model collapse’)
https://pivot-to-ai.com/2026/05/06/openai-chatgpt-goes-goblin-mode-let-none-say-model-collapse/

2026-05-05 11:55:40 GMT

- reply

@nprofile…fw8h what have I ever done to you?

2026-05-05 11:53:48 GMT

- reply

@nprofile…fw8h hab ich vergessen / verdrängt

2026-05-05 11:47:42 GMT

- reply

@nprofile…fw8h OMG

2026-04-29 12:58:40 GMT

The Open Source community never learned that big tech is not trustworthy.
Maybe this time?
(based on https://de.wikipedia.org/wiki/Datei:Zuckerberg_-_dumb_fucks.png)
https://tldr.nettime.org/system_nettime/media_attachments/files/116/488/111/320/619/976/original/93f42305253181ca.png

2026-04-29 12:42:13 GMT

- reply

@nprofile…s733 it's kinda insulting TBH. Not even coming up with "we are being attacked" or anything. Just "yeah we created a bad situation and we do not invest into the core business because we are too busy misusing all the code you naively gave us"

2026-04-29 12:32:50 GMT

Microsoft and people working there just are not very trustworthy sources of information these days.
As the GitHub exec's gaslighting shows.
They just think you are dumb or very ill-informed so the can get away with nonsense.

2026-04-27 12:17:10 GMT

We kinda settled on talking about "enshittification" a lot because it maps to our experience of the world: Everything slowly collapsing.
But what we need is a push towards first "unshittification" and then "antishittification": Digging us out of the ditch and then building political and economic infrastructures and policies that make sure it does not happen again.

2026-04-27 10:56:28 GMT

Okay gotta migrate off to stock Debian then I guess.
Microsoft kills its OS with all the copilot garbage users do not want and Canonical has to go: "Yeah, that's where we need to go."
Strategic genius.

2026-04-24 13:32:07 GMT

- reply

@nprofile…c3an musst ihn als Hörer mal drum bitten ;)

2026-04-24 13:23:07 GMT

- reply

@nprofile…c3an Ed ist nicht so politisch, ich glaube er kann mit meinen Argumenten nicht ganz so viel anfangen

2026-04-24 12:43:54 GMT

A lot of the time when people argue that we should look at the opportunities of "AI" instead of always looking at the damages, that "AI" just won't go away I mentally replace "AI" with "child labor" and it really shows how people think about the world.

2026-04-24 12:26:55 GMT

- reply

@nprofile…c3an nope. Der sucht glaube ich eher US Promis

2026-04-24 09:53:55 GMT

- reply

@nprofile…m0rx getting books on is a little finicky: It can connect to a wireless network and then has a web interface to upload things to. It's not a USB mass storage device sadly. But that's the only annoyance I have.

2026-04-24 09:51:48 GMT

- reply

@nprofile…m0rx it has the perfect size
https://tldr.nettime.org/system_nettime/media_attachments/files/116/459/059/788/006/143/original/1eaa65ba7ccf6f3a.jpg

2026-04-24 09:48:53 GMT

- reply

@nprofile…m0rx It is fantastic. I use the Open Source Firmware and it works beautifully.
I don't use it to read long papers of texts that I want to annotate (it does not have that functionality) but I have gotten to spending my lunch breaks reading fiction books (currently going through Ursula K. Le Guin's works) instead of staring at the phone.
The reader is so small that I just have it in my coat pocket all the time. The battery lasts forever.

2026-04-23 11:32:56 GMT

- reply

@nprofile…sksg thank you!

2026-04-22 08:02:04 GMT

- reply

@nprofile…unx4 Well if that is about the recent "Mythos" thing, did you see https://www.flyingpenguin.com/the-boy-that-cried-mythos-verification-is-collapsing-trust-in-anthropic/ ?

2026-04-22 07:38:00 GMT

- reply

Okay I did it https://www.linkedin.com/feed/update/urn:li:activity:7452618832789106689/

2026-04-22 07:16:38 GMT

Wondering if I should also post my "AI is a fascist artifact" thing to LinkedIn or if that would just go right over people's heads there.

2026-04-21 21:45:47 GMT

- reply

@nprofile…rny0 that is exactly _not_ the kind of support I wanted.

2026-04-21 21:39:16 GMT

With my day job being a bit ... in trouble (company is not me as a person) I keep looking at whether I should try to do a PhD.
Which is a massively dumb idea that I should not follow up on cause it would not work and I'd just hate myself.
If I mention it again, please punch me over TCP/IP.

2026-04-21 21:15:29 GMT

- reply

(I care so much about this issue that I actually did even some proofreading!)

2026-04-21 21:14:13 GMT

In which I argue that "AI" as we understand it today is a structurally fascist artifact.
https://tante.cc/2026/04/21/ai-as-a-fascist-artifact/

2026-04-21 21:12:18 GMT

A @nprofile…u7p7 at least tries to keep his promises.
Here it is: https://tante.cc/2026/04/21/ai-as-a-fascist-artifact/

2026-04-21 06:07:29 GMT

This sort of repairability should be legally required.

2026-04-20 17:30:40 GMT

The "agentic" in "agentic AI" refers to the agency the "AI" providers want to take from you.

2026-04-20 16:43:14 GMT

Okay I am a bit late but I promised some people I'd write this. So if all goes well I might get it out of the door today.
https://tldr.nettime.org/system_nettime/media_attachments/files/116/438/033/551/954/717/original/315adc981f0ed2f7.png

2026-04-20 10:11:01 GMT

Personally I mostly tried shifting to "digital agency" ("digitale Handlungsfähigkeit" in German) because it talks about what it's about rather than pointing at the "evil other"

2026-04-20 07:24:19 GMT

I have been warning about the term "digital sovereignty" and how it is right-wing coded and probably can't be salvaged for non-right-wing politics.
The German fascist party AfD now created a European foundation to push their fascist politics further. The name: "Sovereignty Foundation".
Even thought tactically it might feel like it makes sense to use the term to get funding, you are integration right wing politics into your thinking and speaking.

2026-04-16 18:23:20 GMT

Jetzt hab ich auch selber eine Kopie. Mein Beitrag über die korrodierende Wirkung stochastischer Text-Extruder auf demokratische und auch Verwaltungs-Prozesse ist zwar jetzt schon ein paar Monate alt, aber ich kann immer noch voll dahinter stehen.
https://tldr.nettime.org/system_nettime/media_attachments/files/116/415/764/321/555/755/original/899e1a6375051a9c.jpg
https://tldr.nettime.org/system_nettime/media_attachments/files/116/415/767/296/588/239/original/344e6741fc5da434.jpg

2026-04-16 14:47:34 GMT

"Tech" wasn't /taken/ from us. We just never were willing to realize that it did not include us except as powerless users and customers.
Love for "Tech" is often more akin to Stockholm syndrome.

2026-04-08 13:44:39 GMT

"On the acceptance of GenAI"
https://smallsheds.garden/blog/2026/on-the-acceptance-of-genai/
https://tldr.nettime.org/system_nettime/media_attachments/files/116/369/383/660/367/551/original/0c4cfd13cf91f6be.png

2026-04-02 19:54:55 GMT

“I also see that I have two Microsoft Outlooks, and neither one of those are working.”
Honestly. That's like 50% of office workers' experience these days

2026-04-02 14:09:02 GMT

Linux foundation has been complicit with the scammy crypto scene for a long time. "Linux Foundation" sounds great, but they are doing a lot of shady shit that we should not accept.

2026-04-01 15:28:39 GMT

- reply

@nprofile…lkdl both

2026-04-01 15:18:52 GMT

I have been thinking a lot about the term "free and open web" lately and I think that if we want to build an Internet that can live up to some promise beyond "I can order garbage delivered to my house quickly" we need to go back to the drawing board. Hard.

2026-03-31 22:33:26 GMT

- reply

@nprofile…4z9s @nprofile…lkqc I elaborated on that here https://tante.cc/2026/01/15/software-as-fast-fashion/

2026-03-31 21:14:04 GMT

Anthropic's Claude Code's full source code leaked. Claude is seen by many to be the best coding LLM on the market with Anthropic proudly stating that Claude Code itself is mostly written by the LLM.
Now this sounds good as long as nobody can see the code which is quite the trash fire. Detecting "code sentiment" via regular expressions, variable and functions names containing prompt parts trying to influence the bot, a completely intransparent mess of a control flow that makes actual maintenance and debugging functionally impossible and the prompts ... of the prompts. All the begging and pleading to the chatbot not to do this or not to do that or please to do this.
It is fascinating but it is as far away from actual engineering as drunkenly pissing your name in the snow. Dunno what you call the people prompting software at Anthropic but "engineer" is not it.
Now it is fun to look at the currently hyped product striped bare and showing its pathetic quality but that is the future of software if we let those companies continue to undermine every good practice software engineering has tried establishing.
The software we have to use will be bad, insecure, unmaintainable, expensive with nobody having the skills or resources to build something better. As I wrote a few months ago: LLM based software production is equivalent to saying that fast fashion should be the only way to produce clothing. A tragic degeneration of the quality of the artefacts we rely on build for maximum profit on the backs of people in countries from the global majority.

2026-03-30 10:20:06 GMT

"Das Auf und Ab der Tech-Konzerne an der Nasdaq hat ein Ende - nun geht es stetig abwärts. Um ihre Versprechen zu halten, agieren OpenAI und Co zunehmend verzweifelt."
(Original title: Tech-Konzerne an der Nasdaq: Druckabfall in der KI-Blase)
https://www.golem.de/news/tech-konzerne-an-der-nasdaq-druckabfall-in-der-ki-blase-2603-207049.html

2026-03-25 15:00:46 GMT

I personally consider "I asked ChatGPT to generate a response to you" not witty but a form of an insult. Don't do that please. If that is how you want to talk to people at least don't tell me. It's offensive.

2026-03-25 14:58:17 GMT

That would just be one user account. The service connects with each user's account through a browser making it a bit annoying to detect. So you can block the developer but anyone else could use the same tool to scrape your posts and feed it into a slop machine.

2026-03-25 13:38:08 GMT

- reply

@nprofile…efvt yeah. Since it runs in people's browser's IP blocking or headers don't work, but it registers as a specific Application name with a mastodon server which could be blocked.

2026-03-25 13:34:02 GMT

Does anyone know if a Mastodon server can block client applications based on their name?

2026-03-24 21:08:14 GMT

The OpenAI IPO in the fall will be the closest the "AI" bubble has come to the crypto bubble.

2026-03-24 20:14:43 GMT

OpenAI kills their Sora Video generator.
Which isn't that surprising TBH:
- it is massively expensive to run
- the results were not that usable
- it was a lawsuit generator
So in order to keep the numbers looking better till their IPO at the end of this year they are starting to cut shit. Just an indicator that things are not looking good cashflow-wise.

2026-03-23 13:12:46 GMT

The "AI" bubble currently cannot pop because tech has nothing else to land on.
Metaverse was dropped after a few months like a bad habit because everyone could pivot to "AI". But there is no new story available right now.

2026-03-23 13:09:35 GMT

Spending 3 to 5 USD for every USD you make is not exactly what I'd consider good business.

2026-03-23 13:04:39 GMT

"Microsoft 365 has more than 450 million paid commercial seats. After roughly two years on the market, Copilot has converted approximately 15 million of them into paying users. That's a 3.3% conversion rate, at $30 per user per month, generating roughly $5.4 billion in annual revenue. That's less than what Microsoft spent on infrastructure in a single quarter (3)."
None of the numbers make economic sense.
https://finance.yahoo.com/markets/stocks/articles/nadella-paid-650m-recruit-ai-170000361.html?guccounter=1

2026-03-22 14:52:19 GMT

- reply

@nprofile…fvyn I knew her short stories just never had read her Long-Form

2026-03-22 14:30:14 GMT

It was my plan to read some Ursula K. Le Guin and having gone through half of "The Dispossessed" I am so angry for it taking me so long to do it.
Holy fuck.

2026-03-21 21:20:20 GMT

- reply

@nprofile…208x WARUM WERDEN DIE HAFENSÄNGER DANN SO KRASS GUT BEZAHLT?

2026-03-21 20:47:52 GMT

Ich hätte also nachweislich mehrfach deutlich mehr Recht als hochbezahlte PWC Berater*innen.

2026-03-21 07:29:27 GMT

The "parrot" in "stochastic parrot" stands for "memorization".

2026-03-20 12:00:16 GMT

Declaring one's humanity seems to be the thing these days so I also did:
https://tante.cc/human.json (using the standard defined here: https://codeberg.org/robida/human.json)
If we know each other/met and want me to vouch for you, ping me.

2026-03-20 09:41:27 GMT

- reply

@nprofile…5fek he needed to defend himself since my criticism got a bit of traction. I get it. I think he's still arguing against a straw man but I personally didn't feel too attacked TBH.
He used a lot of quotes to basically say "I do not need to care about the source of things" which sadly is kind of the mainstream conservative standpoint.

2026-03-20 09:32:08 GMT

- reply

@nprofile…5fek he is? He kinda wrote a blog post trying to rebuke my criticism (not 100% successfully) but I didn't see more than that.

2026-03-19 17:54:59 GMT

This is garbage. `uv` is the best tool for the job in the Python ecosystem but it being developed and maintained by a startup (with no path to sustainable funding) this was basically inevitable.
People underestimate the bad position open source is in currently and how much of that space is annexed

2026-03-19 11:50:16 GMT

- reply

@nprofile…dv93 sind seine LinkedIn Posts eigentlich offizielle Dokumente oder vibed er da auch nur so rum?

2026-03-10 16:47:54 GMT

- reply

@nprofile…4agh but Microsoft and Amazon have super sovereign EU data centers! Easy. They also have chips!

2026-03-10 16:41:04 GMT

- reply

@nprofile…hhq5 https://showrss.info/

2026-03-10 09:56:27 GMT

- reply

@nprofile…3a4s absolut.

2026-03-10 08:46:10 GMT

- reply

@nprofile…m5v4 Vivaldi ist gut, hängt natürlich auch 100% von Google ab (da ein repackaged Chrome). Ich nutze auch sowas wie Riverside mit Firefox bzw. Zen (einem Firefox derivat). Aber von den Chrome Reskins ist Vivaldi sicher einer der besten.

2026-03-10 08:26:41 GMT

Weil die "KI" Unternehmen alle Hardware kaufen, wird der "du bekommst nur noch ein dummes Terminal und musst alle relevante Infrastruktur mieten" Trend noch weiter angeheizt. Hardware wird nicht nur teuer, sie wird komplett unzugänglich. Da hilft uns dann auch keine Open Source Software mehr.
https://www.golem.de/news/verteuerte-hardware-ki-konzerne-versperren-den-weg-aus-der-cloud-2603-206158.html

2026-03-09 20:01:41 GMT

- reply

@nprofile…m5v4 @nprofile…37xe nur als Hinweis: Brave fällt immer wieder mit shady shit auf (Crypto bullshit, referral codes überschreiben, damit Brave das Geld bekommt, etc) und wurde von einem Typen gegründet und wird weiterhin von ihm geleitet, der wegen seiner Homophobie bei Mozilla rausgeflogen ist. Brave ist ewww

2026-03-09 15:20:05 GMT

Dass Unternehmen sich alle Hardware krallen macht nicht nur Computer teuer: Sie ist ein direkter Angriff auf unsere Fähigkeit, unabhängige Infrastrukturen zu bauen.
Uns werden die "Mittel der Computerisierung" entrissen um uns Zugang zu vermieten.
(Original title: Verteuerte Hardware: KI-Konzerne versperren den Weg aus der Cloud)
https://www.golem.de/news/verteuerte-hardware-ki-konzerne-versperren-den-weg-aus-der-cloud-2603-206158.html

2026-03-09 13:43:00 GMT

- reply

It's also deeply funny how software developers label the thing deskilling them "skills".

2026-03-09 13:42:14 GMT

"AI" skills are basically just documentation and it says something about how documentation only gets resources and respect if it is framed as "technology". Which has gender bias all over it.

2026-03-09 13:00:28 GMT

A few unsorted thoughts about how the "Claude" models differ from what for example OpenAI release. Basically: One is a religious text, the other tries to be more of a commercial product
https://tante.cc/2026/03/09/claude-magic/

2026-03-05 19:04:29 GMT

Oh: The whole "use an LLM to strip away licenses/original authors' expressed preferences" thing is actually happening a lot:
https://toots.nu/@dalke/116175977286467174

2026-03-05 18:54:20 GMT

- reply

@nprofile…j6r5 @nprofile…w2cy "AI content can't be copyrighted" is a bit of an oversimplification. TBH for the matter at hand it is relevant only in the sense that Claude surely was trained on chardet so it supports the "derivative" argument a bit.

2026-03-05 18:52:37 GMT

With all that's going on today was still a good day because Mark Pilgrim is still around. His work was very influential on me as a young engineer and it makes me happy to see him still doing things.

2026-03-05 18:16:10 GMT

- reply

@nprofile…w2cy yes, there are contributors who contributed under LGPL especially Mark Pilgrim who started the whole project under that license.

2026-03-05 18:15:23 GMT

Here's the thing (wrt chardet).
The "clean room" statement is weakened by putting the new "independent" code in the same git repo/tree. I do think that this does imply a level of dependency/relationship that goes beyond pattern matching strings.
He didn't generate chardet-ng or rename the old version "chardet-legacy" and created a new repo for the claude version. I do think that that as a form of communication/statement matters.

2026-03-05 18:13:18 GMT

- reply

@nprofile…w2cy yes. They jumped from 6.0.0 (LGPL) with a huge merge request that changed everything to the Claude generated version and relicensed it. So it is in the same tree. Same name/canonical URL and everything. So the "clean room" argument is at least softened by putting it in direct succession

2026-03-05 18:08:00 GMT

Is that good?

2026-03-05 17:45:43 GMT

- reply

@nprofile…w2cy Before the Claude "clean room" reimplementation chardet was licensed under LGPL.
So the dude used claude (which was probably trained on chardet/LGPL) to generate a new version of chardet with the same API etc but put it under MIT license.

2026-03-05 14:50:42 GMT

This case shows how Open Source will die. With anyone just being able to pipe existing code and tests through an LLM and claiming that to be "clean room" (which is hogwash) no licensing can protect your work from being accumulated and monetized by anyone. The commons are actively being shredded in front of our eyes.
https://github.com/chardet/chardet/pull/322

2026-03-05 10:24:00 GMT

- reply

@nprofile…4agh submissions closed. Maybe next year

2026-03-05 10:02:07 GMT

Okay, my "Democracy instead of just “Open Source”" talk proposal for #rp26 was rejected. Maybe I'll find another home for it.
https://tante.cc/2026/01/19/republica-2026-talk-proposals/

2026-03-05 09:40:33 GMT

- reply

@nprofile…hccw less "on fire" more "massively sleep deprived"

2026-03-05 08:32:37 GMT

- reply

@nprofile…vt6g @nprofile…drgj @nprofile…qxkq German courts have made the same calls.

2026-03-05 07:55:11 GMT

- reply

@nprofile…drgj @nprofile…qxkq with code just adding comments wouldn't suffice.

2026-03-05 00:20:05 GMT

"People proudly saying how much the software they work on is “vibe coded” or “written by Claude” underlining how there is no pride in or responsibility for the work one not only put out but actually forces on people."
(Original title: Artisanal care)
https://tante.cc/2026/03/04/artisanal-care/

2026-03-04 23:47:01 GMT

New episode of Some More News on the damage "AI"s do to our mental health is a good summary
https://www.youtube.com/watch?v=-FPJCnEIfjY&

2026-03-04 23:15:43 GMT

Okay. Here it is. A quick note on that new "AI declaration".
Which is mostly just a bit clueless but which also assembled some of the most vile people on the planet for a bunch of scientists and activists to put there name next to.
If your tent is big enough for fascists, it's no longer a tent but a concentration camp in the making.
https://tante.cc/2026/03/05/nothing-to-declare/

2026-03-04 21:43:41 GMT

- reply

I shouldn't spend time writing about that trash fire. I should not spend time writing ...

2026-03-04 17:47:42 GMT

- reply

@nprofile…69ps WIZO?

2026-03-04 17:00:50 GMT

- reply

@nprofile…5tzc I think that it's bad, yes