Videos prove, the Russian army stays online in Ukraine using network equipment from American company Ubiquiti.

Worried customers have been discussing this in Ubiquiti’s online forums, but threads have been shut down for “violating community guidelines”.

Many companies see their products sold to Russia, but Ubiquiti products are actually online, which means Ubiquiti could trace the illegal use and intervene — so why don’t they?

hntrbrk.com/ubiquiti/

reshared this

Holy shit. TIL that Janet Jackson is the only Grammy-winning artist with a CVE.

CVE-2022-38392 indicates that playing Rhythm Nation near certain hard drives will cause a crash, because the song contains a resonate frequency with a 5400RPM spinning disk of a certain diameter and construction.

Neat.

#music #infosec

reshared this

Are you interested in testing your USB cables? Then I have a blog post for you:

blog.literarily-starved.com/20…

Be aware: You might discover that your cable is fooling your PC

This entry was edited (1 day ago)

As it came up in a few conversations during "FOSDEM week", here's a link to the OpenSSF blog post about why the idea of "attestation for open source projects" is, in my opinion, and others, a bad idea:

openssf.org/blog/2026/01/21/pr…

Yes, FOSS foundations and projects need ways of getting funding, that is very important, but thinking that "attestation is how we will get that money!" might not be such a good idea given the risks involved, and the past experience for those that have attempted it.

@privacybrowser on privacy browser 3.19.3 fdroid android. when i use share url from another browser or from other app, it overrides my existing tab. ex. pbrower open tab is ddg.com. i am reading stoutner.com/ on fennec. when i click share button on fennec and choose pbrowser. ddg.com url is replaced with stouter.com is there anyway to change this behaviour

Iron Blue Intention, Castlevania Bloodlines

Sensitive content

Onj 🎶 reshared this.

I read a great anti-AI screed today (not linking to avoid a pile-on on the author). I found it to be an almost perfect distillation of the typical anti-GenAI arguments I read. The basic form is:

1. *Lists various ethical problems with LLMs*
2. Therefore a consumer boycott is the only solution
3. Fuck you, you're a monster, etc if you don't agree with me

Do #2 and #3 follow, though? Let's think about it.

#3 #2
Unknown parent

mastodon - Link to source

Justin Macleod

The thing that I don't see regulation and all that other great stuff solving is the cognitive effects of AI, which, so short a time after it taking off, have already been demonstrated by studies. People use Chat GPT and that has a negative effect on their problem-solving skills. Also the environmental impact.

But anyway, I think consumer boycots are a way forward but I'm totally against the scolding. Encouraging a boycot is fine,

Unknown parent

mastodon - Link to source

Justin Macleod

saying you're evil and part of the problem if you don't abandon all evil big tech is just self-righteous, holier than thou barrier-building. And we need to work together. Look hard enough at anyone and you will see their double standards, inconsistencies and downright compromise of principles for the sake of avoiding risk or retaining comfort in a difficult life. Is that ideal? No. Is it to be accepted placidly and with resignation? No. But condemnation isn't the answer.

Part of me wants to make the native VLAN, that is, the VLAN that provides access to devices who don't talk 802.1Q VLAN-speak, one that has Internet-only access, and if any devices want access to the local network, they have to speak to a certain VLAN. Of course, that means that anything requiring local access would need to be 802.1Q compatible to speak VLAN, and be properly configured. The other part of me says that my local network is simply not that big of a security target, and I'd just be creating extra headaches for myself. But the geek in me still insists, LOL. So the decision is not yet decided.
This entry was edited (11 hours ago)

GrapheneOS version 2026012800 released:

grapheneos.org/releases#202601…

See the linked release notes for a summary of the improvements over the previous release.

Forum discussion thread:

discuss.grapheneos.org/d/31269…

#GrapheneOS #privacy #security

RE: tldr.nettime.org/@dk/115962588…

when I wrote my critiques of the centralization of the internet, it wasn't because I hated that internet, I love the internet! Same now, I don't hate AI, there is actually a lot I love about it.

What I don't love is centralization, proprietary software, bad dev practices and exploitation.


I love conversational interfaces, and I am a big proponent of automation. The trouble is when these tools are not sold as productivity software to empower people who are doing work, with skill and effort, but as ways to create value without skill and effort.
8/11

in reply to SuspiciousDuck

To by si měli nějak sjednotit, aby bylo jasné, co se tím myslí. Já bych od obědového menu čekal salát, jídlo, aspoň vodu a kávu. Ale přijde mi k ničemu, když recenzent A pod tím vidí, přeženu, samotné hranolky, zatímco recenzent B pětichodové menu se spárovaným vínem.

(To je samozřejmě kritika těch recenzí, zejména teda asi Googlu, ne tebe ani restaurace.)

I was so excited about the new feature the Polish government added to our government app recently, the ability to sign documents using a "qualified digital signature", *for free*.

For the non-Europeans in the audience, a qualified signature is the European equivalent of DocuSign, except far less intuitive, far more expensive, and, by extension, not accepted anywhere near as widely. By regulation, it's the only kind that has the same power as a classic paper signature.

This feature requires an electronic ID card, and as my old ID was soon to expire, I recently got one of these anyway.

I tried doing this with two different providers, and it turns out that both of them just return a very helpful "Server error, without even an explanation of what's wrong."

As they say, disappointed but entirely not surprised.

Wow! #MapToPoster really does generate some great city posters.

github.com/originalankur/mapto…

Ondřej Caletka reshared this.

Wow, so flashing Graphene OS on the Pixel 6 Pro was actually quite easy, just plugged the device in and followed the guide at grapheneos.org/install/web. By now I'll just update this thread with any new discoverysI make, and yes I do plan to write a detailed blog post about that as soon as I can and know more, I really can't wait to explore this honestly.
#GrapheneOS
in reply to Jonathan

Holy moly dude, installing e-speak on there took me about half an hour, longer than even flashing the thing. Had to guide my non techy mother through the stuff, firstly tried to get it through the F-Droid site, but according to my mom the browser wouldn't let her select a download location so I pulled out an USB C stick and put the APK on there. And after surely 10 minutes of idling and apparently loading voice data, E-Speak finally came to life and started speaking. Now it's time to customise this. Not sure whether a simpler option would've been the installation through ADB, but I think you have to enable all that first so the stick route was probably the easiest.

We are cooking @fedora 44, and we could use some help testing it before it is ready for prime time!

From February 11 to 13 we will have the Fedora @gnome@floss.social 50 Desktop Test Days!

Check fedoraproject.org/wiki/Test_Da… to find out how you can participate. #fedora #gnome

This entry was edited (12 hours ago)

Buying a google pixel and installing grapheneOS on it is a much better move for your privacy and mental health than buying any of these crappy e-reader phones or some flip phone that only works over unencrypted communication channels. The rise of 'dumb phones' has lead to an influx of poorly build, insecure phones built by people with no business developing complex tech products. What I find 90% of the time is that people buy these products, realize they actually do rely on features like maps, good cameras, banking apps and all the other smart conveinences of the modern world. Ultimately they end up throwing the device away or in a junk drawer and it just becomes another e-waste waste product. If you want to actually break free from your phone addiction use a product that gices you thr conveinence you want but doesn't try to guide your towards addiction every day. A stock grapheneOS phone is basically a dumb smart phone and thats how it should be.

#FOSDEM was absolutely WILD! 😻

🤯🤯🤯 Got a hand-made #XMPP developers post card with a personal mention!

🤯🤯 Got a "Thanks Ge0rG, keep up the good work!" while walking the food mile!

🤯 Met an IRC contact from 25 years ago, @olbohlen ("Apparently I'm the only operator of #OpenFire on #Solaris") and connected him to @guusdk who brought stickers and hugs!

🤯 Accidentally met an IRC contact from 20 years ago

🤯 Accidentally met an #FDroid member iRL

🤯 Met @sven (#fail0verflow) after 20 years

in reply to Bogomil Shopov - Бого #FOSDEM

and you ran away before I could hand you a bunch of the DI.DAY flyers! 😱 It was a pleasure meeting you and thanks for your talk! We really need to unite our efforts!

Also wanted to ask you what specific messenger "N" stood for in your slides 😉

(P.S.: I've omitted from OP all the people I *intended* to meet, only recounted the unexpected things that happened!)

This entry was edited (12 hours ago)
in reply to Jakob Rosin

@jakobrosin Would you count this? I used the copilot website to design internal spaces for a game I play. There is the possibility, within that game to create floors, walls, lighting and ceilings. Not knowing anything about interior design, colour coordination and lighting, I turned to copilot. I found the responses enlightening and they elevated my creation. Not saying this offsets negatives, but I regard it as my biggest AI success. @jscholes

Logic Pro 12 will not work offline!

Apple will not give me a 11.2.2 version of Logic Pro that I had before. It's obvious that they want to move to subscription model, but I don't want to. So Apple blames the victim, saying "you should have had a backup"? A reputable company would give me the previous version. No warning about this before udate.

This is not cool, Apple. Too busy swanning around the White House, Tim? Let's have some support.

#MacOS #apple #LogicPro

This entry was edited (13 hours ago)

Peter Vágner reshared this.

If you take just one thing away from #FOSDEM this is it.
👇🏽
“If we lose our democracies, Open Source is irrelevant and goes away”

or putting it another way:

If we don’t apply, enforce, adopt and embrace Open Source in our public and government institutions we risk to lose our democracies.
#OpenSource #SaveDemocracy
Boost if you agree!
👍🏾

This entry was edited (1 day ago)

Hi Folks,

I am looking for screen reader users to test a site taking approximately one hour. This is a paid gig - $50 via a gift card. Please ping me for more details.

Please boost for reach.

Thanks!

#a11y #accessibility #ScreenReader

This entry was edited (13 hours ago)

reshared this

in reply to Quin

@TheQuinbox Apparently Claude Code's system prompt is like 16K tokens. That's insane! I also struggle with how much detail to go into. I feel like too much yields odd results, but I don't have much experience with Claude, since I've mostly been using Gemini-CLI, which in one go, managed to convert Echo TTS on Apple Silicon using MPS rather than Cuda.
@Quin
in reply to Brandon Cross

So the thing about Claude is that you have to treat it like a very intelligent 6-year-old who knows how to code. Never make it search for stuff, give it exact line numbers even if you can, searching wastes tons of tokens. What really throws Claude off is too many disjointed thoughts in one prompt, like if I drop one sentence about a future feature I want to add in the middle of the prompt, Claude will think that is a very important detail and try to design for that feature, even when it doesn't make any sense. It's also far from perfect, the experience with Claude is I "/superpowers:brainstorm fix #301", and it will either produce something that's incredible and fixes it on the first try, or make me cringe and wish I just did it myself.

There's a new infestation on the block. Automatic meeting summariser AI-s. They join virtual meetings on Zoom or Teams on behalf of people, unannounced.
So, as an organiser, you may see something called "Firefly" or "read.ai" join. They proceed to spam the chat with "helpful" snippets of the meeting. Then, they will send everyone, who was invited to the meeting emails with "helpful" summaries, which are hidden behind a link, for you to klick and get sucked into their marketing database.
From a recent meeting, I received no less than 5 emails from 5 different services.
Personally I just kick them out as an organiser, just as I would do with a physical meeting, when a random bloke nobody knows about would barge in and sit in the corner. If they would harass everyone after the meeting in the parking lot in their cars with "Tell me your credit card number and I tell you 10 super cool things of the last meeting", I would cut their owners out of my comms completely. In fact, I am not really shure why we find it acceptable in virtual spaces right now.

Zach Bennoui reshared this.

I missed this but Gentoo Linux allows you to bootstrap the rustc compiler using a regular C++ compiler via mrustc. You should be able to build the entire Rust toolchain from source without having to rely on a rustc binary distribution if you wish so.

wiki.gentoo.org/wiki/User:Tric…

#Gentoo #RustLang

Wow… Google fined a whopping $68 MILLION DOLLARS for eavesdropping on people’s spoken conversations. That’ll teach them! It’s going to take Google a whole hour and a half to make that money back. I guess they won’t be pulling shit like this anymore after feeling that sort of pain. We really showed ‘em this time, eh?

Correction: it’s not even a fine; it’s a settlement with no admission of wrongdoing. So even worse.

mastodon.social/@JulianOliver/…

This entry was edited (9 hours ago)

reshared this