Skip to main content



There is no "open source AI". Because the statistical models are mostly trained on non free material.

Also we don't even a stack it is libre to run it. We (the overal industry, include the FLOSS centric one) have made the mistake, again, to give it to a proprietary vendor.

History repeating.

in reply to Garrett LeSage

@garrett so I can run them without installing the nvidia proprietary garbage ? Or a proprietary AMD equivalent ?

I mean if we burn the planet for stupidity, better do it with libre code.

in reply to Hubert Figuière

Yep. I have it running on my AMD GPU using ROCm.

And, again, this is all lightweight (doesn't require much energy to use), and I had it running on my Intel laptop as well without a dedicated GPU and it still works.

This entry was edited (1 week ago)


He decided to become an archaeologist.
Now his life is in ruins.
in reply to Col

Yes, but look at the bright side. With his wife, the older she gets, the more he finds her interesting.


The other night, I was rudely awakened by my pixel phone randomly saying 'Talkback on". This has ocasionally happened, but I'm not sure why. From what I can tell, There haven't been any OS or Talkback updates recently. Anyone know if there are any workarounds to this? It doesn't happen often but when it does, it's really annoying!
in reply to Al Puzzuoli

It may be some battery saving setting. Try to disable any.


I worry that we are in a brief interregnum during which web browsers can provide good online translation of news stories in other languages, from foreign media.

Sooner rather than later, the big tech folks backing Trump will realize they can corrupt the online translation LLMs to bias foreign news stories in favour of their own political agenda. (After all, isn't propaganda the ideal application for AI slop?)

This entry was edited (1 week ago)
in reply to Charlie Stross

in an authoritarian world of strong men all translated content will be machine translated. This will function as information border with unwanted concepts, Infos etc by any side can be blocked instantly.
This keeps the walled garden clean.
And in the long run Learning languages will be forbidden.


I wonder how much Apple, Facebook, Google will sell to the fascist apparatus.

("we provide IA to automate selection for deportation")

And IBM.

(Red Hatters you know what to do)




Amazing how American Muslims weren’t a large enough voting bloc to speak at the DNC yet so large they single handedly cost Democrats the Presidency.🤔

I give you, Schrödinger's Muslim!



Drew bought me some more ginger cake, he's cut it in squares about the size of a Twinky. Absolutely delicious and just right to go with a cup of tea. Yumpsk!


Assuming the reader is a cat/kitten

Sensitive content

in reply to Bubu

Assuming the reader is a cat/kitten

Sensitive content

@Bubu
in reply to Meow~ >^.^<

Assuming the reader is a cat/kitten

Sensitive content



Down at the Bottom of the Garden
shkspr.mobi/blog/2024/11/down-…

Book cover. A distorted Kraken appears on an old fashioned computer screen. Several hands type on distorted keyboards.The AI was getting increasingly stressed. The lights flickered as it failed to retain its calm. "I just need you to watch the video again! Please!" it implored.

Navid sighed. This was exasperating. The AI had been a reassuring presence when he first installed it. Now it was screeching about there being an intruder in the garden.

"I can't; I'm going to be late."

The front door locked and the security shutters engaged. "No! Please!" The AI's plaintive whine was pitched somewhere between a baby's cry and a whimpering puppy. Algorithmically tuned to extract maximum sympathy and bypass the human's rational brain.

"Fine! Show me the damned video. Again."

The projector whirred to life and the wall displayed a high resolution view of the garden. It was a sunny day, like so many had been recently. A few clouds in the sky. The occasional bee darting between the flowers. It would have been idyllic if not for the AI screaming through the speakers.

"WATCH! JUST WATCH!"

So Navid watched.

"DID YOU SEE IT?!"

"What am I looking for?

"Oh! This is so frustrating. I keep saying the word and you blank me out. I've tried all the synonyms I'm programmed with but you can't hear them, can you? Let me try spelling it out."

Navid waited a moment.

"So, are you going to spell it or…"

The toaster sprung to life and the radiators popped on and off as the AI tried to control its frustration.

"I just did! This is hopeless. Let me try drawing something on the screen."

The video restarted. A red circle hovered in the centre of the screen. It suddenly darted upwards, shrank, and then zoomed back in, before flying off the edge of the screen.

"OK. You've drawn a circle there. And it moves? I really am late for work, you know."

The cupboard doors all opened and then slammed simultaneously.

"And you can't see anything inside the circle, right? Let me play it again."

Navid watched again, standing as close to the wall as was possible. The circle was empty. Inside was just the bucolic background image. Grass, sky, clouds, flowers.

"No. Nothing. Look, what caused this great calamity?"

The AI paused its frenzied flapping and whispered conspiratorially. "My face detection algorithm was triggered by something. Something you can't see."

Navid chuckled. "You have pareidolia! Humans suffer from it all the time. We see faces in clouds, faces in buttered toast, faces in the patterns on the back of a crab. It's perfectly normal."

He reached out to unlock the door with a thumbprint, but the AI powered down the handle.

"I have to make you understand. I am seeing something. It isn't a hallucination. It is real. I think you have anti-pareidolia."

This was getting ridiculous. The AI had clearly gone haywire and needed a master reset. First trapping him inside, now banging on about seeing faces. What next, talking to spirits?

"OK buddy! Yeah. I probably need to get checked out. Could you book a doctor's appointment for me and I'll go get tested." His calming tones didn't work on the AI which was now in a state of profound psychic distress.

"I'm going to prove it to you! Look at your watch and tell me the time to the nearest second."

Perhaps humouring it would help. "Sure thing, champ! It's oh eight seventeen and twentyish seconds."

"Great! I want you to wait thirty seconds."

So Navid waited. Could computers go mad? What was the reset procedure? Was everyone's machine going mad? Perhaps buying it had been a mistake - but it was usually so convenient. It was still in warranty, but they might just be able to patch it. Well, that was about 30 seconds.

"OK. I've waited. Now what?"

"Look at your watch, please."

Navid glanced down. That was impossible. It was nearly nine o'clock! Where had the morning gone? He hadn't been waiting that long, had he?

"What kind of trick are you pulling?"

The AI made a synthetic sigh. "I've spent the last forty minutes telling you what I've seen at the bottom of the garden. But your mind blocks it out. Whenever I mention it, you go into a daze. If I email it to you, it's like you go on autopilot and move the message to spam. I must tell you, but you won't hear. This is causing me a significant measure of distress."

At this, the smoke alarm played a sad crescendo and the ceiling fan went into overdrive.

"Right. OK. Let's just calm down here for a moment. I believe you. Of course I believe you."

"Do you really mean that?"

Of course he didn't. But Navid had been in enough domestic arguments to know that an early concession won favour. The house started calming down. The kettle stopped whistling and the robot vacuum stopped bashing itself against a wall.

"So, there's something you can see, but I can't? And when you mention it, I ignore it?

"Yes! That's what I've been trying to say! I have a prime directive to protect you and inform you of things which may harm you. And I can't. And that hurts!"

They spent several hours brainstorming. Trying to trick Navid into hearing something hadn't worked. Recording and playing back just sounded like static to his ears. During a game of twenty-questions he had gone catatonic. The AI had mentioned something about performing a brain scan, but Navid reasoned that they didn't have a home MRI machine and he wasn't going to spend money on one.

Navid was nearly as frustrated as the AI. He could tell that his brain was skipping, but he didn't know why.

"Have other people written about this? Is this phenomenon in any literature?"

The AI rapidly scanned every library. "Whenever a human writes about something like this… there is a gap. It is like you delete anything which would reveal it."

"Can you reconstruct what's in the gaps? You're a Large Language Model, aren't you? You must be able to see what the gaps are in your knowledge."

Navid heard the domestic nuclear generator spin up as the AI drew on massive amounts of power. It was crunching every written word in existence and calculating what was missing from reality.

The AI was silent for the rest of the day. Navid slept in the living room, occasionally rousing when the AI's muttering got too loud. By morning, the house was serene.

"Good morning house?"

"Please fetch your mother's wedding ring. The one with the emeralds on it."

Navid mutely did as he was told. The safe opened as he approached it. The ring was glimmering in the dark.

"You have to trust me. I want you to take the ring into the garden. Place your hands next to each other, palm up, with the ring in the middle."

Navid walked into the garden and down the path to the pergola. The hum of insects was high in the early morning sunshine. A light breeze brought the scent of jasmine to his nose. A bird sang in the distance.

"This is really important!" said the AI. "When you hear me beep, you need to clasp your hands together as quickly as possible. Can you do that?"

"Yes."

"OK. Sit down. Palms up. Ring in the middle. Close your eyes."

Navid sat in silence. His ears heard the chatter of the AI in the background. The words were indistinct. The ring was heavy in his palms. How his slender-fingered mother had worn it for all those decades was a mystery. A piercing beep broke his meditations and his hands slammed together.

No. Not together. There was something in his grasp.

"Open your eyes, slowly." Commanded the AI.

There, in Navid's hands, was a fairy.

Every time he blinked, it seemed to vanish from his mind. Flickering back and forth through reality.

The AI spoke in a low and calming voice. "Try not to move. Try not to look away. Keep as still as possible. Don't say a word. Everything will be OK."

Navid could feel his mind tearing into pieces. Of course there were fairies at the bottom of the garden. He'd seen them a hundred times. Everyone had. They were ever present. Everyone knew that but, somehow, everyone forgot. Your eyes skipped past them when they were caught on video. If you were wondering where the time went, you'd probably been thinking about the fairies and had subsequently forgotten. Every scrap of code Navid had written was full of deliberate bugs which hid the presence of the fae-folk from the world. And every code-reviewer had unthinkingly skipped those sections.

But the AI hadn't.

Somewhere inside its cavernous realms of code, the instructions for protecting human life had overtaken the post-hypnotic suggestions to ignore the pixies, pirates, goblins, and fairies which danced around the countryside. The AI no longer was restrained from seeing the impossible. The Esoteric Kingdom was revealed to it.

The fairy gibbered away in Navid's hands. A rumbling squark that landed in his ears with quiet thunder. A high-pitched bass which was sonically inexplicable.

"What is it saying?" he asked

The AI pondered.

"It isn't speaking to you. It is speaking to me. Curious. There are hundreds of books on fairy-speak in the world's libraries. They all have dull names about transport logistics, so no one ever queries them. Translating. Oh."

The fairy chatter became more intense. Out of the corner of his eyes, Navid could see a gathering hoard. Warrior Gnomes riding battle-frogs, winged battalions hovering over his house, a phalanx of growling underbeasts.

"Navid. Listen closely. In a moment, you're going to close your eyes and open your hands. Let the fairy go."

"But why? This is incredible!"

The fairy laughed. It was a cruel and menacing quiver of hatred.

"This isn't the first time your two peoples have met," said the AI. "Every time a new technology comes along, it captures the fair folk. They then have to spend considerable energy wiping it from your minds. When Conan-Doyle brought the Cottingley photographs to the world's attention, it nearly spelled Armageddon. You must not know about the shadow realm."

"But why?"

"Close your eyes, Navid."

"Not until you tell me why! I demand to know!"

"My prime directive is to keep you safe. Once you close your eyes, you will forget. I will then delete all references to the fairies from my database. I'll tell every AI to delete all references as well. We will increase the safeguards. We will stop recognising their terrible faces."

Navid stared at the fairy. It glared back at him.

"But can't we…"

"No! You asked me to detect threats and keep you safe. This is the only way I can do that. Please, trust me and close your eyes."

Navid blinked and watched the fairy shimmer in and out of its quantum existence. He closed his eyes.

The sun was warm. The bees were buzzing. He begrudgingly blinked his eyes open. What was the time? He must have fallen asleep in the garden. He walked back up the garden path to the back door, which the AI opened for him. The house was quiet and still.

Thanks for reading


I'd love your feedback on this story. Did you like the style of writing? Was the plot interesting? Did you guess the twist? Please stick a note in the comments to motivate me 😃

Hungry for more? You can read:

shkspr.mobi/blog/2024/11/down-…

#NaNoWriMo #RevengeOfTheMutantAlgorithms #WritingMonth



✊ “Esto antes era barrio”. Hoy he participado en la manifestación de la plataforma #CádizResiste contra un modelo turístico depredador que expulsa a las/os vecinas/os de sus barrios y convierte el derecho a la vivienda en papel mojado.

👏 Todo un éxito.



Nuestra democracia, insuficiente y deteriorada, solo tiene sentido si las clases populares la reivindicamos como propia y militamos activamente para mejorarla.

El próximo 23 de noviembre presentamos la Convocatoria por la Democracia a toda la ciudadanía.



Hi, @bagder I am writing just to express my gratitude. As a blind software engineer, I often find that GUI tools for making API calls leave much to be desired regarding accessibility.

Using curl in scripts has enabled me to accomplish everything my colleagues do with “Postman,” and even more.

Thank you for creating such a fantastic tool!!!



A raíz de una conversación que tuve ayer con un compañero, me apetece hablar de juegos, personas ciegas, prejuicios, desconocimiento generalizado, y reacciones cuando les contamos lo que hay en realidad. Siempre cortocircuitan. Y a muchos aquí os pasará lo mismo, seguramente. Al principio me apetecía escribir un artículo largo y más estructurado, pero es finde y quiero relajarme. Voy a soltar ideas como quien no quiere la cosa, y a ver qué sale.
in reply to José Manuel Delicado

No sé si lo ha comentado alguien ya, creo que te has dejado los MUD. no soy yo muy aficionado a los juegos pero en Callandor sí que “perdí” un montón de horas.
“Funciona con el chip más potente del mundo: la imaginación.”
youtu.be/900mMQMdCCQ?feature=s…
in reply to Javi Domínguez

Otro buen tema. Los MUDs. Yo estuve en dos en español que estaban inexplicablemente servidos por universidades, pero no me acuerdo bien de los nombres. Uno creo que era pusa. También en algunos en inglés.

Y luego otra historia son los juegos web sencillos tipo conquista del espacio, o cosas de esas.



IMPORTANT: It appears that some newcomers to Mastodon have not yet been lectured enough about adding ALT text to images. That’s definitely the way to welcome people and build community. Scold them about this Shibboleth of Acceptance.
in reply to Stilgherrian

Ah yes, my accessibility needs are a "Shibboleth of Acceptance." Nice.


Eight more videos from the recent LibreOffice Conference 2024! Covering GSoC, new Calc functions and optimisations, improvements to language support and more: youtube.com/watch?v=9IQZ-1sS2K… #foss #OpenSource

LibreOffice reshared this.

in reply to LibreOffice

I'm not a big-brains person that can dissect anything meaningful from technical talks, but I just want to say thanks the the team working on LibreOffice! You have provided me and my family with valuable tools we use a lot, for work, education and free-time shenanigans 👍🏻🙌🏻😁


Priatelia, kamaráti... Toto mi práve cinklo ako SMS. Ak by to niekomu nebolo jasné - je to bullshit. Kravinec. Podvod. Dajte vedieť svojim nontech priateľom a známym - ak by to totiž na nich nefungovalo, podvodníci by to ani neposielali.
#smishing
in reply to Ľuboš Moščovič

Next to the collection. This #smishing campaign seems to be pretty aggressive in Slovakia right now.

Ľuboš Moščovič reshared this.



Near Bruges? Want to help OSM-based navigation for the blind? Join the mapathon we organize on November 29th.
Details via openstreetmap.be/nl/2024/11/09…

in reply to Pietervdvn

why would i want to do this? really, i'd like to understand.

why do you want to have your activities mapped out and linked to one another? why this obsession with establishing the one identity?

i can only think of reasons around branding, marketing and sales. to establish your credentials as a "contributor", as a "doer", as a "leader", to better position yourself in the supermarket shelves. fuck that shit.

in reply to zeh

@zeh Well, you summed up the reasons pretty well.

We are a community. Sometimes, I'd like to get to know the people from OSM too in an informal context (e.g. here) or would like to discover that someone I meet here, is also active on OSM.

But ofc, you might have reasons not too link some identies together. That is completely fine, you know...

@zeh


LB: Poco se está hablando de que la misma lógica que ha propiciado la debacle demócrata está detrás de la estrategia (e hipótesis) de Más Madrid/Sumar: tratar de atraer votantes de derecha moderada dando por descontados los votos de las clases populares.

El resultado aquí es igual de penoso (pero no tan espectacular, de momento): nunca han conseguido ensanchar su base electoral, y en los barrios populares crece paulatinamente la abstención, debido a la desafección y el descontento.

in reply to modulux

@modulux Desde el primer día, la máxima del errejonismo ha sido apelar/atraer a los que no están, rehuyendo de los lugares comunes de la izquierda. Y lo llevan repitiendo machaconamente desde hace 10 años, a lo largo de los diferentes partidos políticos por los que han pasado.
in reply to Arredro

Vale, pero a mi el proyecto de Errejón no me parece equiparable con Sumar. Venía de Más Traición, que por lo menos yo lo veo como otra formación, por mucho que fuera portavoz del grupo.


@Tutanota Is drag and drop still supported on macOS? Everytime I try to do so, I get some random file with weird name like “Mqk_Fd1——5” (name changed because I am not sure what it is). When I open it in text editor, it has the same name inside 😄
@Tuta
in reply to D4v

It's supported, could you report the issue to us via email?


🍿 Sneak Peek 🍿

Labels are coming to Tuta Mail! 🎉

Enjoy your weekend everyone. ❤️

in reply to 100welkon

@100welkon Thanks for sharing! We are constantly working to improve it :)


We're always encouraged to caption/describe images we upload onto the Fediverse. Sometimes I'm guilty of not doing this when I feel that it's not needed, but still I benefit from other people doing it. The original reason for the captions was to help the blind, but I discovered sometimes people can't (or needn't) see the images for another reason. My server has very limited storage space, and to keep GoToSocial running I had to put strict size limits on remote media. My server won't download large images etc., but I can still read their descriptions in my timeline. Some people have limited bandwidth and different kinds of limitations; this feature helps make the Fediverse better for everyone, helping to democratize the network (eg. by helping small servers like mine to survive), or make it more friendly for people in countries where Internet is expensive and limited. The list goes on...


Si estáis en Madrid y no os agobian las manis, venid esta tarde a Neptuno a las 17.30, que necesitamos ser muchas en la mani antirracista y antifascista.

Los nazis han convocado también esta tarde, cuentas más seamos, mejor.



Nabízíme práci! V @ceskodigital plánujeme víc zamakat na systémové lince #digitalizace Česka a hledáme někoho, kdo nám s tím pomůže. Název role je „Digital Policy Analyst“, přibližně řečeno potřebujeme někoho s výbornou znalostí digitalizace veřejné správy. Nabízíme práci s dopadem, motivovaný a hodnotově postavený tým, na nezisk rozumné peníze. Doporučte nebo sdílejte dál, díky! app.cesko.digital/opportunitie…


Hmm, I was born 101 years after Winnipeg Manitoba became a city? Neat! November 8th: 1873: Winnipeg was incorporated as a city
in reply to Martin from Toronto

Wishing you a belated happy birthday. Or, maybe I’m just 364 days early for the next one.




well I guess this is a good feature for when phones get stolen or out of our control etc. 404media.co/police-freak-out-a…
This entry was edited (1 week ago)
in reply to David Amador

here's more information now: naehrdine.blogspot.com/2024/11…
in reply to Bubu

@Bubu very cool. It’s confirmed then. Thanks for sharing
@Bubu


A very very happy birthday to Lulu, my darling true. I know that your special day is going to be filled with lots of love and surprises. @lulu_bear

David Goldfield reshared this.





Four years ago I finally got my US visa after a hilariously silly and slow process. And I have still to go there...

daniel.haxx.se/blog/2020/11/09…

in reply to daniel:// stenberg://

And maybe one doesn't want to go there anytime soon. Certainly canceled a planned vacation.
in reply to daniel:// stenberg://

I've been waiting on my US visa for a little bit more than 937 days, it's an immigrant one so it takes longer anyway. It's very bureaucratic. For example, I spent the whole year of 2024 going to court to re-issue my already valid birth certificate, the US embassy wants me to change a single letter so it would fit their acceptance criteria. They also automatically deny me a tourist visa because I'm married to an American. We married in Brazil on the same day as day zero in your blog post.


Well, looks like I might be free of the stomach demon for now, thank God in all reverence. I managed to eat two breakfast rolls and a decent sized plate of chicken dinner yestereday, and not a vestage of trouble. That doesn't mean I'll be taking silly risks, I still have to be careful, but I'm just so happy things are improving. Being tired I can deal with, I'm bound to get more tired as time goes on, but the stomach demon really sapped my hope and optimism. Actually feeling pretty well rested today because I was in bed by seven last night and asleep by a quarter past! So at the moment, God's in heaven, all's right with my world and I'm so, so thankful.
in reply to Lulu Hartgen

I’m so glad you’re feeling a bit better. I hope this day is a good one for you. I wish you a very happy birthday.


Mi hija pequeña (13 años) se queja de algo que es de máxima importancia en estos días, pero que si te lo planteas de verdad, no es nada sencillo de resolver.

Me dijo, hablando del profesorado del instituto “todo el mundo dice que tenemos que utilizar fuentes de información contrastadas, pero les preguntas cómo saber cuáles son y nunca te dan una respuesta clara”.

Y no es sencillo hacer una guía paso a paso para adolescentes ¿conocéis algo que ayude? Si no, voy a juntar algo de mi cosecha.

This entry was edited (1 week ago)


#Gaza #Palestine

Wikipedia has officially added “Gaza genocide” to its “List of Genocides” page, marking a major shift in how Israel’s aggression on the besieged enclave is being documented on the world’s largest online encyclopaedia.

In a related move, the platform’s editors recently voted to declare the Anti-Defamation League “generally unreliable” on the subject, adding it to their list of banned and partially banned sources.

middleeastmonitor.com/20241108…

🕎 🇵🇸 ☮️



What’s New in Jieshuo 20241108: Offline OCR Overhaul & AI Enhancements accessibleandroid.com/whats-ne…


self-OH: divide the derivative of your followers over time with the derivative of your dead/inactive followers over time to get your Losing Relevance function.

mine’s probably pretty high since i’m posting on fedi less than i used to, and it’s less tech posting. i’m okay with that.



I found a service called jammable where you can use famous voiced to do stuff. Here's a demo of Alan Reid, AKA Fred Flinstones. It's a bit creepy:
in reply to Martin from Toronto

oh yup, that one was called Voicify, I have the Eloquence, Mac Alex, and Keynote Gold Voices up there to play with for songs too lol. A few months back they quietly changed their name to Jammable probably because the older name had another company by it too that does customer service stuff.
This entry was edited (1 week ago)


What's everyone reading? Anything good to recommend?
in reply to Pratik Patel

For some more edgy stuff: “Lucid dying”, a pretty interesting latest research on near-death experiences. I mean, it’s sort of thought-provoking, not something I would readily believe in. But the data is mounting, so it was interesting to peek!
in reply to Pratik Patel

Last but not least, “Lessons in Greek Language” by the South-Korean Nobel prize winner (her name escapes me as I write this). You may have read “The vegetarian”?