#AndroidAppRain at apt.izzysoft.de/fdroid today with 13 updated and 2 added apps:

* ADB over WiFi: toggle ADB between USB and WiFi mode (root)
* Locale Changer: change the default system-wide locale

Over at F-Droid, 32 updated and 4 added:

* UpVPN - Serverless VPN: VPN based on WireGuard
* Threema Libre: Secure messaging
* BendyStraw: manage NewPipe databases
* LibreBible: Latin Vulgate Bible reader

Enjoy your #free #Android #apps with #FDroid and the #IzzySoftRepo :awesome:

The #ext4 data corruption issue[1] in #Linux #kernel v6.1.64 and v6.1.65 that was fixed with #LinuxKernel 6.1.66[2] apparently hit #Debian 12.3 bookworm point release[3]. Fixes are in the works, but preparing them will take a bit[4].

[1] lore.kernel.org/stable/2023120…

[2] git.kernel.org/pub/scm/linux/k…

[3] bugs.debian.org/cgi-bin/bugrep…

[4] fulda.social/@Ganneff/11155162…

This entry was edited (2 years ago)

V nedávném odhalení bylo zjištěno, že vlády sledují uživatele smartphonů pomocí push notifikací od Apple a Google.
To zdůrazňuje potřebu alternativních řešení, jako je ntfy.arch-linux.cz/
Ntfy je open-source nástroj, který umožňuje odesílání push notifikací bez závislosti na infrastruktuře velkých společností.
👉 Soukromí: NTFY zajišťuje, že vaše data neputují přes servery třetích stran, čímž zvyšuje ochranu osobních informací. 🧵 1/2
in reply to Archos

@dosmanak Další zdroje:

bleepingcomputer.com/news/secu…

letemsvetemapplem.eu/2023/12/0…

This entry was edited (2 years ago)

Big Brother is watching you - now also via Google's & Apple's push. 🤬

That's why we don't use Google Push. When using Tuta, Google sees nothing! 💪

✅ Zero tracking
✅ Fully encrypted
✅ Maximum privacy

The data they don't get, they can't hand out to authorities! Go secure now: tuta.com/create-email-account?…

Here's more info on why we don't use Google Push and do not send any info via Apple notifications: tuta.com/blog/open-source-emai…
#security #privacy #google #apple #notifications #surveillance #tracking

To nevypadá špatně. Jede i na slabém stroji. Pro někoho kdo chce mít malou instanci super
#activitypub #gotosocial #fediverse
docs.gotosocial.org/en/latest/
github.com/superseriousbusines…

"Your Monthly 30km"❤️💙 Do jsem se Teda nechal unést a překonal jsem rovnou 3km. Teda, jednou jsem si musel odpočinout, asi 3 minut...

🎶NWOBHM🎶
🎸🎤SAMSON🎸🥁
už se zase nenápadně přibližuju k mým oblíbeným Iron Maiden. V Samson totiž kdysi zpíval Bruce Dickinson 🎤

in reply to Sean Randall

So much of our family whatsapp is visual now too. I'm loving the generated description's of images, and never thought I'd enjoy all the memes but I really do. I know there's plenty of ethical concern about LLM's, their water use etc and I hope we can find ways of balancing all that. But I also never believed we'd be where we ar with tech describing images at this level. Whatever next?
in reply to Sean Randall

What next? I hope the next part is figuring out how to make this stuff a lot more efficient, like Llava for example, and moving it out of the greedy hands of big corp into the caring hands of everyone else. And, naturally, everyone who contributed should have a say in how that's done. Like llava, or llama, are still trained on huge amounts of text it shouldn't have been, and I think it's important we figure out how everyone walks away happy from the whole thing. Because I agree, being able to get descriptions of things in the way this tech provides is very, very big and I don't have to tell people not to send me pictures anymore. Because they can, now.
in reply to Talon

IMO a lot of the environmental concerns are questionable. Sure it costs electricity to run an AI tool, but so does sending a photo to someone. If the grid is using reasonable energy sources it's not as much of a problem anyway. And of course if we can run these things locally there's no significant water use, since we don't cool computers at home with water.

I don't think it's reasonable to forbid people to do analysis on text or images, and I think that sort of copyright maximalism does not end well.

in reply to Talon

I've been thinking a lot about this because I do think, and have even experienced, how AI is used for evil. To undermine and diminish the work of professionals, to impersonate people, and of course how AI is trained on lots and lots of data that either the person owning the data didn't want, or that the company just straight up shouldn't have access to. But on the other side, you have these multi modal models, even just the small ones, which can give impressive and detailed descriptions of images. And I totally understand every single problem people have with AI, and I stand with you. But also it hurts if you tell me that I shouldn't have access to images. Like you can see them. It doesn't matter to you. But to someone who can't, this is really, really big. It's a world that's previously been either closed off, or only accessible through the kindest souls who caption things. But even that isn't always possible. Maybe the medium doesn't allow it, maybe the artist or owner just didn't do it, maybe there's a language barrier, maybe I took the photo myself. There are many reasons why images might be uncaptioned. And maybe there's a better algorithm, better data, whatever, to get better descriptions, but we don't have it. And very few people talk about it, because usually, either very rudimentary categorization is more than enough, and we've been able to do that for a very long time, or the use case just isn't apparent to many people. But we are a big part of that use case. It makes us more independent. And yes, these models lie, but people aren't always accurate either. Not even intentionally. It's a big problem and I don't know what the solution is. But I'm sure that ableism isn't it.

modulux reshared this.

in reply to Talon

So cars are both good and bad? I guess you have a point there. I don't know. All I know is, a lot of the tech we use every-day has some unintended negative consequences yet we have to use them for survival in a modern world. It's not something I'm saying we should take lightly but still, most average folks don't have the power to stop stuff like climate destruction when its the responsibility of large corps to figure it out which in practice, they're sort of understanding but not fast enough.
in reply to Nick's world

And we get there with discourse. For example the voices who are strongly against AI are just as important as those of us who now have access to entirely new worlds because of its existence. Just like how there are very important discussions to be had about cars and their impact vs. those of us who need them. We can't just say something is good for me so it must be good for everyone else, because that's not how it really works. And exactly this nuance is important in figuring out a better overall situation for everyone. When someone says AI is bad, then they have valid reasons for having this opinion, and proof exists that they're just as right as everyone who does use it for good, like us and image descriptions for example. So it's very tricky.
in reply to Talon

Also for example if the solution to the image description problem is something that isn't generative pretrained transformers containing lots and lots of questionably acquired data, then I'm all for it. Just like how, for a lot of people, public transport could potentially be an option. The important thing is that we don't get stuck in one particular way and instead look at what problems we're trying to solve and how we actually solve them.
in reply to Talon

Also I feel like I should say that I'm not just going to dump random photos you send me into OpenAI's hands. I keep talking about local first models for a reason. They're definitely less accurate than OpenAI's models, but also I really don't think they should know exactly what it is I'm getting described. So if I can, I'll do that instead. And I think it's important for this tech to get to a point where it's easier and accurate enough for everyone to do that too. I want image descriptions but I also want to be able to do it privately. Also i don't want OpenAI to be able to tell me what kinds of pictures my innocent clean little mind should be able to comprehend.
in reply to Talon

people talk about the evil of AI, and image descriptions is one good thing, but there's other good things it's done, like using AI to reproduce someone's voice after a stroke so her husband can now have actual conversations with his wife again after many years. there's medical good things AI can do, but everyone's focused on doom and gloom and it's sad. yes image descs for the blind is good, but there's other good from it too.
in reply to WingedUnicorn🧙🏻‍♀️🌕🔥

But we need especially as white people to be aware about the harms, especially to black and brown folks, racial profiling, using AI to "flag criminals" and the harm this has done. Yes it's helpful, and yes it is causing amazing amounts of harm. Both things are true But as white people we will never be harmed by AI like other marginalized people are harmed.
in reply to WingedUnicorn🧙🏻‍♀️🌕🔥

in the beginning of the thread I do say that ableism isn't the way to solve it, as in, criticizing the people who make use of it for accessibility. I do agree. But both of those things can be true. We benefit from them, and they cause real harm. And I think it's important to not forget it or diminish that fact. Work is nowhere close to being done. Inf act it's probably just starting.
in reply to Talon

I get they're problems. but how many times are you going to bring up the fact this is a problem. sometimes one can say something a bit too much then it becomes nagging or hounding in some people's view. would you want to hear about the danger of something every day, multiple times a day? I don't because personally I get it, but not everyone can do something about that particular something.
in reply to Talon

There was some research done at MIT on a new neural network architecture to replace the transformer that resulted in I think up to 100x reduction in resource usage, which brings it squarely into the embedded devices region. Their test case was an LLM. The thing is, after GPT did it, everybody said no you don't get to scrape us! And thus the reddit and X API lockdowns. It is literally designed, whether intentionally or not, to ensure that no equal competitor can emerge.

Jestli vás omrzel klasický košík a bidon, tak tohle video se vám bude líbit.
invidious.cz/watch?v=pbyyP3oDC…
This entry was edited (2 years ago)

The government is looking into whether the ongoing energy crisis fulfils the legal requirements for the measure, Ramaphosa said in closing remarks at the ANC’s strategy meeting on Monday.

“Work is already underway within govt to establish whether the legal requirements of a national state of disaster are met and what specific actions we would be empowered to undertake,” he said.

#CarteBlanche #SouthAfrica #Loadshedding #News

mybroadband.co.za/news/energy/…

I've long believed that all apps (at least on Android) should start supporting UnifiedPush[1]. Tools like ntfy could serve as a central hub for notification distribution, allowing anyone to have their own notification server to connect apps and mobile devices. This would ensure greater privacy.
This is also why I developed NotiMail: we shouldn't have to rely on the 'big players' for our notifications and data.

[1] unifiedpush.org/

#UnifiedPush #AndroidApps #PrivacyMatters #NotiMail #Google #Apple #Android #iOS

,,Smutná čísla umrzlých lidí, se kterými se setkáváme rok co rok, nás staví před zásadní otázku: Jaká je hodnota lidského života? A jaká je cena za individuální selhání?Ve skutečně spravedlivé společnosti by odpověď na druhou otázku nikdy neměla být: ohrožení smrtí. Lidé bez domova potřebují pomoc, ne odsouzení, ne spekulace nad tím, jak se do své situace dostali a že si za ni mohou sami."
#bezdomovectví
irozhlas.cz/komentare/mraz-bez…

10 extensions that turbocharge Windows Explorer pcworld.com/article/2161758/yo…

reshared this

Apparently the BBC (UK) has had 144 complaints about a recent episode of #DoctorWho because it contained an openly #trans character.

I'm going to make a complaint to the BBC that there weren't enough #transgender characters in Doctor Who. I would love if 144 other people did the same thing. Here's the link: bbc.co.uk/contact/complaints/m…

(For your easy reference: "The Star Beast" aired on 25/11/2023, BBC One, and the trans character is called Rose.)

Reply when you've done it, so we can count us all. When we get to 145, everyone... gets a free BLÅHAJ? 🦈

Edit: 2 days later, mission successful: 159 complaints/comments submitted, and still climbing! A *huge* thank you everyone - stay tuned for hopefully some freedom of information requests!

This entry was edited (2 years ago)

For anyone in band who played clarinet and had your #band teacher tell you how bad the alto clarinet is, this is one of my favorite clips that just shows just how decent the sound of the clarinet is when played correctly. Disclaimer: Not a clarinet player. youtube.com/watch?v=T_IA8K7xK8…
#band
This entry was edited (2 years ago)

New street food in China: noodles topped with mashed potatoes, probably inspired by another Chongqing potato noodle dish

youtu.be/CjH0ZOpgy9o

#Food #ChineseFood

There are days when this whole LLM talk gets kinda boring... I mean, for how long can you really talk to your local LLM? And what would you really talk to it about? OpenHermes-2.5: This Local LLM Is All You Need anakin.ai/blog/openhermes-2-5/
in reply to The Continent

Also in this week's issue:
🇳🇬 Nigeria: Another military “mishap” kills 88 people
🇮🇱 Israel: Importing more Africans to fill the labour gap of youth sent to war
🇹🇿 Tanzania: The death penalty is alive and well
🇸🇴 Somalia: What “loss and damage” actually looks like
🇨🇺 Cuba: Kenya kills controversial doctor exchange programme
🎬 Review: The movie about the missing movie

Get this and more in this week's issue.

thecontinent.org

This entry was edited (2 years ago)
in reply to The Continent

In last year’s reader survey, one reader neatly summed up The Continent: “It’s good.” This newspaper likes concise. We also value your thoughts on how we’re doing, and what else we can do. Please give us a few minutes of your time and complete this year’s survey. We use your answers to make our journalism better, and to convince funders to give us the money to keep doing it. This means we can pay journalists for their work, and keep The Continent free for all. forms.gle/UMQhyg1vTzBbL7ru8