Search results for tag #opensource
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/109/436/509/454/338/545/original/47ebf4508db42048.gif)
WE REPORT FEWER INSTALLED APPS NOW. THAT IS THE FIX.
That number on the card went from 3,981 to 791 and it is the thing I am proudest of in this release.
In 2018 I wanted a list of the apps installed on an iPhone. I thought that was one question. It is at least six, and no single artifact answers all of them.
So we tested it properly, iOS 12 through iOS 26, every registered test image we have. The old installed summary had been counting app extensions, malformed bundle IDs, and apps known only from container activity. Strip those out and the count drops by 80 percent. Nothing was thrown away. It moved to a table whose label says exactly what the evidence supports, and the historical view got bigger at the same time, 9,575 rows to 11,786.
The history got bigger. The claim got smaller. That is the trade I want every time.
My favorite detail: on one public test image, two apps being removed produced seventeen container destruction rows in nine seconds, because every app extension owns its own container. Read those rows as apps and you would report a phone shedding seventeen. It shed two.
An inventory and a stack of receipts are both useful. They are not interchangeable.
Link in bio.
#DFIR #DigitalForensics #MobileForensics #iOS #OpenSource #ForensicScience #Infosec
![[?]](https://cdn.fosstodon.org/accounts/avatars/109/224/068/401/095/470/original/5a6bec71ce4217cd.png)
New Job Alert: MOBIUS, a nonprofit 501(c)(3) library membership organization located in Columbia, Missouri, is seeking a Customer Service DevOps Linux Engineer. Learn more and apply now on #OSJH
https://opensourcejobhub.com/job/29581/customer-service-devops-linux-engineer/?utm_source=smm
#DevOps #engineer #Linux #ITServices #AWS #GoogleCloud #OpenSource #upstreaming
Alt...
MOBIUS LINKING LIBRARIES
![[?]](https://files.mastodon.social/accounts/avatars/116/597/984/930/130/852/original/0cf9960fb9039d51.png)
Socket found 737 fake VPNs on Chrome Web Store, impersonating brands like Proton and NordVPN. The extensions have 75,000+ downloads and routed traffic through an operator's server.
This is why open-source matters for privacy tools: you can verify the code instead of trusting marketing.
Read more: https://socket.dev/blog/chrome-vpn-extension-impersonation
![[?]](https://mst-gts.airikr.me/fileserver/01HXQ9K4V66SFN8WF3SVYC18X6/attachment/original/01KVASV5K2G14N26KDR6EPDC2G.webp)
[I have finally created my own image sharing website with end-to-end encryption! Feels so darn good 
But to be honest, everything is built by me, but not the encryption on client side, management on server side, and to view the thumbnails in the gallery when logged in without breaking the encryption.
But I am very proud of the result anyway! 
More info: https://pic.airikr.me
The source code can be found on my self-hosted Forgejo server since Codeberg are trying to fight to "get rid off" AI generated code. I understand them and I understand the risk, though.
But why AI? Without AI, I would not be able to create this project with end-to-end encryption at all since I suck at JavaScript.
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/109/429/381/491/877/380/original/560ac6d98e44a7bc.jpg)
ICS[AP] updated CISA ICS Advisories Master File for 8/25/2026 & the following year's CSVs:
CISA_ICS_ADV_2026_8_25_26.csv
Updates on 8/25/2026 included:
CISA_ICS_ADV_2025_8_25_26.csv
Available @ ICS[AP] GitHub:
https://github.com/icsadvprj/ICS-Advisory-Project/tree/main
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/110/560/511/618/170/134/original/4f33ba36608d1311.png)
9to5Linux: Ardour 9.8 Open-Source DAW Is Now Available with Initial Scales Support https://9to5linux.com/ardour-9-8-open-source-daw-is-now-available-with-initial-scales-support
Also:
OpenSSL 4.0.2 Released with Important Security and Bug Fixes https://9to5linux.com/openssl-4-0-2-released-with-important-security-and-bug-fixes @9to5linux @mariusnestor #Linux #opensource
![[?]](https://secure.gravatar.com/avatar/088105cc4f211b43698ad3ad1b2c4d58ef402bfb423627bfa976a443afc5bf1e?s=120&d=mm&r=g)
The license is all you need (for Open Source AI)
Over two years of practice and recent policy decisions prove that the Open Source AI Definition v1.0 is valid. Now let's focus on the other unresolved issues, like fairness. [SENSITIVE CONTENT]
It turns out that open training data is not the most important piece of an Open Source AI. The license of the model is.
I spent two years leading the process that decided what “Open Source” should mean when the thing being opened is a foundation model instead of a codebase. While serving as executive director of the Open Source Initiative, I ran a large, global, inclusive effort, with dozens of organizations and hundreds of contributors for a couple of years, arriving at a definition of Open Source AI that none of us could have dictated alone. The harshest discussion in that process, by a wide margin, was about training data, not weights or code.
A small but vocal minority pushed hard for the definition to require only open training data as a condition of calling a model “Open Source,” not as a nice-to-have but as a floor. Without it, they argued, you don’t have all of the four freedoms, the backbone of Open Source: you can’t fully study or reproduce the AI, so you don’t have Open Source at all. The objection didn’t go away after the Open Source AI Definition (OSAID) v. 1.0 shipped. The definition requires the weights and the code that generates them under an Open Source license; additionally, it requires sufficiently detailed disclosure about what went into training them, but not necessarily the data itself. I’ve defended the outcome of the process for two years. Practice so far has shown that the OSAID holds. This summer, policy defended it, too.
The purity test
A version of the argument lingers: if the training data stays a black box, an Open Source license on the weights is a technicality, because you can’t verify what you can’t inspect. Real security and real trust require opening the whole AI-creating pipeline, not just the output. I get the intuition; it is more rigorous. But reality shows that regulators and AI adopters aren’t acting on data provenance. They act on the AI model’s license first and the credibility of who built the thing second.
Suppose the Chinese government released a model tomorrow with full weights and a complete, public training corpus, thoroughly documented. The security establishment would not trust it automatically. Trust doesn’t come from disclosures alone. It comes from long-established credibility of the labs building the models, the governance of the project. A track record earns trust in a way a data dump can’t. Today, data disclosure remains an argument for courts to dissect copyright issues.
Nineteen days dark
On June 9, Anthropic shipped two new models, Fable 5 and Mythos 5. Three days later, the US Commerce Department applied export controls with immediate effect, barring access by any foreign national, anywhere, including Anthropic’s own staff. Nationality can’t be verified in real time at the API layer, so both models went dark, globally, for everyone. Partial clearance came June 26 for a short list of vetted US critical-infrastructure organizations. Commerce lifted the controls June 30. Fable 5 came back online June 30, nineteen days after a policy decision took a live model away from the entire world, including the company that built it.
Six weeks later, the mirror image played out. Moonshot AI, a Chinese lab, opened the API for its new K3 model on July 16 and released the weights on July 27: a frontier-class model on an open release path no export order can touch once the files are out. Moonshot AI showed the power of open: You can switch off a proprietary, hosted-only model, but you cannot switch off a copy already running on the users’ hardware.
On July 21, US Treasury Secretary Bessent said the government would examine Chinese Open Source models for intellectual property theft, with sanctions on the table. That’s a statement of intent with complications: sanctions need an enforceable chokepoint, and once weights have been downloaded across a dozen jurisdictions, there isn’t much to grab onto. Mozilla’s State of Open Source AI report says the quiet part out loud: an outright prohibition gets harder to enforce as adoption grows, because every copy already held sits outside the reach of the order.
Whatever Washington ends up doing about Chinese open weight models, it will have to act on distribution and access, on what a license permits and who can get the files, because that’s the only lever still standing once the weights are loose.
These actions show that regulators are more concerned with who’s allowed to access the download link than with whether Moonshot AI disclosed its training corpus: The license is what counts.
What the numbers show
Kimi K3, a model available for free with some restrictions
, ranks fourth on the Artificial Analysis Intelligence Index, within four points of the proprietary frontier. Models with no or few restrictions to use now carry roughly a third of all tokens routed through OpenRouter, up from about 2% eighteen months ago, and still climbing. Chinese-built open models are running at more than 3.5-to-1 over US-built closed models in token volume on the same platform. DeepSeek-R1 ships under a plain MIT license, at reasoning scores within a couple of points of OpenAI’s o1, for $0.55 to $2.19 per million tokens against o1’s $15 to $60, roughly a 27x price gap at near-identical performance.
These numbers show that users aren’t concerned about data provenance but they care more about model availability, capabilities, and cost.
AI foundation models are a brand new “thing”
Last year I was one of the co-authors on a paper in Communications of the ACM, alongside people who spend a lot of time analyzing AI openness: Yann LeCun, Nathan Lambert, Sayash Kapoor, and a dozen others across Mozilla, the Linux Foundation, Columbia, Princeton, Eleuther AI, Allen Institute for AI and more. That framework’s whole point was to resist a single-axis view of trust and transparency for AI: don’t stare at the weights trying to adapt old concepts to this artifact, but look at the full stack. The framework treats licensing and documentation of the model as separate, cross-cutting attributes. Documentation covers model cards, evaluation results, and information about the data, distinct from the data itself. A license to the weights is its own category, alongside documentation rather than beneath it.
Current.ai’s Gap Map, which catalogs what’s needed to build and ship a model, runs on the same architecture. It tracks datasets, code, and weights as separate components with their own separate gaps, and nowhere does it treat a dataset as a requirement of the model trained on it. The Gap Map is a practical representation of AI openness produced at the Columbia Convening, and it still doesn’t back the purity position. It agrees with me: license plus disclosed documentation is the operative bar for Open Source AI.
Time to say it plainly
Let’s stop pretending this is still an open question. Training data provenance matters, and that’s not up for debate. But training data isn’t AI’s source code, and it was never going to be the policy’s lever. The CACM framework says as much, and so does Mozilla’s report.
Washington is reaching for export controls and sanctions built entirely around license and access to models. The courts are narrowly arguing copyright issues. In the market, open weight models under OSI-approved licenses are eating token share regardless of what anyone knows about their training corpora.
The purity test was always a bet that the world would eventually agree data openness is the real bar. The world placed the opposite bet, in public, this summer, using just the license as its operating lever. It’s time the rest of us caught up to back our own definition and move the policy debate towards a fair AI ecosystem.
- A previous version of this post incorrectly classified Kimi K3 as “open weight”, a term that I use to imply the license of the model is OSI-approved. This is not the case of Kimi K3 license. ↩︎
Alt...
still life with antique books
![[?]](https://horche.demkontinuum.de/photo/profile/hoergen.jpeg?ts=1635405826)
sudo - Rechte kontrolliert vergeben | Dem hoergen Blog
hyperblog.de/hoergen/posts/202…
#Linux #Opensource #Wissen
Mit sudo können Benutzer Befehle mit erhöhten Rechten ausführen – kontrolliert und protokolliert.
Location: Casablanca
![[?]](https://files.layer8.space/accounts/avatars/000/048/652/original/5a7fc5072d05ed7b.png)
I miss being part of a team. Very specifically, the collaborative energy of like-minded people in the #opensource community striving toward a common goal.
I don't say this because I'm looking for employment. At this point in my life, I just miss using my skills to help make a difference. You know?
That said, I wonder if I might find contentment volunteering for an open source project. Who out there needs help?
![[?]](https://cdn.mstdn.ca/accounts/avatars/116/315/849/906/959/964/original/11981bad852db16c.jpg)
When it comes to my daily personal computer, i use #LinuxMint and really enjoy the experience. I use the cinnamon environment which just works with everything which i love.
I have tried other Distros like Bazzite, SteamOS, Ubuntu & Debian for my daily PC but they all required alot more tinkering then i prefer for a daily PC.
Now I want to move my Homelab off my daily PC and put it onto its own environment/Computer to have my own home server.
Question is, what Distro should i use for my home server?
Some of the options i been looking at are the following:
Linux Mint XFCE (Similar to my main Distro)
Ubuntu Server (Popular Choice)
Debian (Popular Choice + Familiar With The Environment)
These are just some of the options i been looking at but i never hosted a #homeserver before, so im unsure what would be the best for such an environment.
What is your opinion?
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/111/794/732/819/301/361/original/0d23cdbcde72f0f3.png)
![[?]](https://files.mastodon.social/accounts/avatars/110/313/914/144/289/307/original/028599dd509a1226.png)
Today marks Linux’s 35th birthday! 🎉🎂
35 years of milestones that took a hobby project from one developer’s announcement to the foundation of much of modern computing
Here is a look at the milestones that shaped its remarkable journey 😎👇
Find high-res pdf ebooks with all my Linux related infographics at https://study-notes.org
#linux #software #technology #computer #opensource
![[?]](https://s3.c.im/accounts/avatars/115/289/404/158/252/336/original/b51373306a6ff7d5.jpeg)
Does anyone know of a Linux GUI for GhostScript so that I can easily compress PDFs? (Yes, I could do it using the terminal but I hope for other, computer-averse folks to use it too)
I've only come up with workerPDF
https://sourceforge.net/projects/workerpdf/
Thank you! 🙏
![[?]](https://social.heise.de/__virtual/accounts/avatars/109/722/110/464/209/207/original/8c6b0cc3354d3bc1.png)
heise+ | Verschlüsselung über kryptografische Identitätsschlüssel mit Iroh 1.0
Statt IP-Adressen sollen Schlüssel Endpunkte identifizieren: Iroh 1.0 vereinfacht direkte verschlüsselte Verbindungen über Netzgrenzen hinweg.
#IdentityManagement #IT #Netzwerke #OpenSource #Software #Verschlüsselung #news
![[?]](https://files.mastodon.social/accounts/avatars/000/377/402/original/19edbd43c05a3fcc.png)
Are you using LLMs to write Free Software?
Then you should know whether your software is copyrightable, the situations in which you can license it, and what risks it may create for maintainers.
Our new Legal Corner explains what you need to consider:
https://fsfe.org/news/2026/news-20260825-01.html
#FreeSoftware #OpenSource #Copyright #LLM #AI #generativeAI #VibeCoding
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/116/797/276/219/291/526/original/96271e701c673af9.png)
NVIDIA partners with AI coding startup Poolside in a $7 billion deal to strengthen its open-weight Nemotron model and challenge OpenAI.
#NVIDIA #PoolsideAI #ArtificialIntelligence #TechNews #OpenSource
https://securityexpress.info/nvidia-poolside-ai-deal/?utm_source=mastodon&utm_medium=jetpack_social
![[?]](https://media.phpc.social/accounts/avatars/115/664/641/595/887/477/original/ca405d6eae616a05.png)
One Reddit post outperformed a month of homepage work for us. The AI-Ready Plugins list (every package built on the Laravel AI SDK, ranked by health) brought 302 visitors, 266 straight to that page. Focused and genuinely helpful wins. Full July recap: https://laraplugins.io/blog/laraplugins-july-update-the-month-of-growth #Laravel #OpenSource #PHP
chmod - praktisches Handbuch | Dem hoergen Blog
hyperblog.de/hoergen/posts/202…
#Linux #Opensource #Wissen
Linux-Zugriffsrechte werden entweder numerisch (Zahlen) oder symbolisch (Buchstaben) gesetzt. Symbolisch ist dabei flexibler. Sonderrechte wie SUID, SGID und Sticky Bit erlauben erweiterte Kontrolle über Programme und Ordner.
Location: Dushanbe
![[?]](https://mstdn.gsi.li/system/accounts/avatars/109/318/818/613/814/121/original/cedf062b212e6b38.jpg)
🚀 Chronivaro 0.1.0 has been released!
Chronivaro is an open-source working time & absence tracking system built on Strolch.
Highlights:
• Live timer & daily time recording
• Absence management & vacation journal
• Approvals & period closing
• CSV & PDF report exports
• Standalone fat-JAR & Docker
📦 Downloads & Release Notes:
https://github.com/eitch/Chronivaro/releases/tag/v0.1.0
![[?]](https://media.bsd.cafe/bsdmmedia01/accounts/avatars/117/026/439/058/798/087/original/af8e3cb0866b0580.jpg)
dropQbsd has been moved from GitHub – where a copy will remain – to Sourcehut, its natural home.
I’d like to thank Drew DeVault for welcoming me onto his platform.
For free spirits, dropQbsd can be found here: https://git.sr.ht/~nobraininside/dropQbsd
#OpenBSD #freebsd #netbsd #openSource #privacy #sourcehut #digitalsovereignty
Alt...
dropQbsd, the flying fortress - ASCII Art
![[?]](https://friendica-deutschland.de/photo/profile/doris.jpeg?ts=1761744546)
📢 Fediverse-Sprechstunde: Was sind eigentlich „Issues“ – und warum gehen sie uns alle etwas an?
Ihr nutzt Friendica, Mastodon, Hubzilla, Sharkey oder eine andere Fediverse-Software und plötzlich funktioniert etwas nicht so, wie ihr es erwartet? 🤔
Oder ihr denkt euch bei einer Funktion:
„Das könnte man doch besser machen!“
Vielleicht habt ihr auch eine Idee, die ihr gern den Entwicklerinnen und Entwicklern mitteilen würdet.
Dann fällt früher oder später ein Begriff: „Mach doch ein Issue auf.“
Aber was bedeutet das eigentlich?
Was gehört in ein gutes Issue hinein? Wo erstelle ich es? Welche Informationen brauchen Entwickler, um einen Fehler nachvollziehen zu können? Und muss ich dafür überhaupt technische Kenntnisse haben?
Genau darum geht es bei unserer nächsten Fediverse-Sprechstunde:
📅 Montag, 31. August 2026
🕢 19:30 Uhr
💻 Online
🔎 Was sind Issues und wie erstelle ich welche?
Für dieses Thema konnten wir Jens (@qbi ) gewinnen. Einige von euch kennen ihn vielleicht bereits aus unserer Fediverse-Sprechstunde zum Thema BookWyrm.
Jens ist Administrator und Entwickler und kennt Issues nicht nur aus der Theorie, sondern aus der täglichen Praxis. Gemeinsam wollen wir uns anschauen, wie aus einem beobachteten Problem, einer Idee oder einem Verbesserungsvorschlag ein Issue wird, mit dem Entwickler auch wirklich etwas anfangen können.
Dabei geht es ausdrücklich nicht nur um Entwickler oder Administratoren!
Denn freie und offene Software lebt auch davon, dass diejenigen, die sie täglich benutzen, Rückmeldungen geben. Niemand kann jeden Fehler selbst entdecken und niemand kennt alle Wünsche und Probleme der Nutzerinnen und Nutzer.
Ein gut beschriebenes Issue kann deshalb ein kleiner, aber sehr wertvoller Beitrag zur Weiterentwicklung einer Software sein. 🌱
Und vielleicht nehmen wir dabei auch ein wenig die Scheu vor Plattformen wie GitHub oder Codeberg. 😉
Ihr müsst dafür keine Programmierer sein und keinen Quellcode verstehen. Neugier und Fragen reichen völlig aus.
💬 Was ist überhaupt ein Issue?
🐞 Wie melde ich einen Fehler sinnvoll?
💡 Kann ich auch Ideen und Verbesserungsvorschläge einreichen?
📝 Welche Angaben gehören hinein?
🔍 Sollte ich vorher prüfen, ob es bereits ein ähnliches Issue gibt?
🤝 Wie erleichtere ich Entwicklern die Arbeit?
Und natürlich bleibt wie immer Raum für eure Fragen und den gemeinsamen Austausch.
Die Fediverse-Sprechstunde soll kein Frontalunterricht sein. Sie ist ein Treffpunkt zum Fragen, Erklären, Ausprobieren und voneinander Lernen – für Einsteiger ebenso wie für erfahrene Nutzer, Administratoren und Entwickler.
🌐 Mehr Miteinander statt Gegeneinander.
Gerade im Fediverse gehören Nutzer, Administratoren und Entwickler zusammen. Wenn wir besser verstehen, wie die jeweils anderen arbeiten und wie wir uns gegenseitig unterstützen können, profitiert am Ende das gesamte Fediverse davon.
👉 Montag, 31.08.2026 um 19:30 Uhr
💻 Konferenzraum:
campus.ws-12.de/rooms/8p9-mvy-…campus.ws-12.de/rooms/8p9-mvy-…
📢 Weitersagen und Teilen ausdrücklich erwünscht!
Vielleicht kennt ihr jemanden, der schon einmal einen Fehler entdeckt hat, eine gute Idee hatte – aber nicht wusste, wie und wo man sie den Entwicklern am besten mitteilt. Bringt ihn oder sie einfach mit. 😊
#Fediverse #FediverseSprechstunde #OpenSource #Issues #GitHub #Codeberg #Friendica #Mastodon #Hubzilla #BookWyrm #sharkey
Alt...
KI-generierte Grafik zur Fediverse-Sprechstunde am Montag, 31.08.2026 um 19:30 Uhr. Unter der Überschrift „Was sind Issues und wie erstelle ich welche?“ ist ein Laptop mit einer beispielhaften Issue-Seite zu sehen. Sprechblasen fragen „Fehler melden?“, „Idee einbringen?“ und „Verbesserung vorschlagen?“. Als Referent wird Jens (qbi), Admin, Entwickler und Issue-Profi, genannt. Darunter steht: „Mehr Miteinander statt Gegeneinander. Gemeinsam machen wir das Fediverse besser!“ Die Veranstaltung findet online im Konferenzraum der Fediverse-Sprechstunde statt.
![[?]](https://s3.eu-central-2.wasabisys.com/mastodonworld/accounts/avatars/113/472/242/607/539/655/original/ce987f5c2a1ad75d.png)
With an extra $18 million raised, W4 Games plan to continue to expand their enterprise technology and support for the open-source Godot Engine.
https://www.gamingonlinux.com/2026/08/w4-games-the-company-formed-by-godot-engine-veterans-gets-help-and-funding-from-tencent/
The Bazzite Deck 44 release for handhelds and TV PCs was a huge change for the popular gaming focused version of Linux, which sadly came with lots of issues.
https://www.gamingonlinux.com/2026/08/bazzite-44-getting-a-hotfix-for-various-issues-after-the-recent-big-release/
#OpenSource #Upcoming #DistroNews #Misc #Bazzite #HandheldGaming
![[?]](https://stockage.framapiaf.org/framapiaf/accounts/avatars/000/064/196/original/ce0db1452740006c.png)
#OpenSource 🔓 #OpenAI vient de passer le moteur de son agent Codex sous licence Apache 2.0 !
Ce n'est pas anodin : #Codex n'est plus un simple assistant de programmation, c'est désormais une plateforme d'orchestration d'agents intégrable dans n'importe quel logiciel professionnel.
➡️ Tokens de sortie divisés par 6
➡️ Score ARC-AGI-3 à 38,3 % (vs 13,3 % avant)
➡️ Déjà en production chez Cisco, GitHub, JetBrains
🔗 https://goodtech.info/openai-codex-harness-agent-open-source-apache-2-0/
#IA #Agents #DevOps #LLM #ITOps
![[?]](data:image/png;base64, iVBORw0KGgoAAAANSUhEUgAAAEAAAABAAQAAAACCEkxzAAAAUUlEQVQoz43R0QkAMQwCUDdw/y3dwEvsvzlL4X1IoQkAisKmwfAFT3RgJHbQezpSRoXEqeqCL9BJBf7h3QbOCCxV5EVWMEMwG7K1/WODtlvxAYTtEsDU9F34AAAAAElFTkSuQmCC)
Got tired of Big Tech spying on me? Self-host your own cloud with Nebula Mesh and Docker. Own the data, be free.
Use Caddy for a solid reverse proxy setup. It's like having a Swiss Army knife for your network stack.
#SelfHostingWins #NebulaMesh #OpenSource
https://wopr.systems/?utm_campaign=meme_woprsystems&utm_medium=social&utm_source=mastodon
Alt...
Use Caddy for a solid reverse proxy setup. It's like having a Swiss Army knife for your network stack.
![[?]](https://littleone.littlefedi.social/media/019fec2d-afa9-750a-8b68-b06963a52642/image.webp)
"Decolonize.Digital is an online resource that is a living, growing, and evolving website bringing together different connected projects that I've been working on. Parts of the site are available to all subscribers, but parts of it require a paid subscription. I made this decision because I've put hours, weeks, and months into this tool, in addition to my decades of work and learning that even makes creating it possible. I regard this work as I would a book, workshop, or any other labor that I do that is part of my life's work.
For free subscribers to my platform liberationtoolbox.io, all of these posts, articles, videos, and various mediums have been made available across my many platforms over the many years for free. And they are still accessible to all through those different platforms.
There are people who have the gall to be upset that I am actually compensated for my labor. If this is you, you should not be here. I do not recommend that you subscribe to the free or paid options of this Liberation Toolbox if you are one of these people.
Currently, on this site, Decolonize.Digital, I have put up the De-Google Your Life (Decentering Big Tech) series, Buildsheet: Community Building series, and the Opt-Out series. I will continue adding tools to each of these series over time. I have some exciting series and tools coming up.
Why now? Why this? When wielded by oppressors, everything, including technology, becomes murderous and rapacious, it becomes digital colonization.
There are better ways for technology. Finding these other liberatory ways, this is my life's work, to Decolonize Digital."
https://decolonize.digital
https://liberationtoolbox.io
#DecolonizeDigital #DIT #DUT #DUTgemacht #DigitalSovereignty #OpenSource #FOSS
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/110/038/598/765/629/546/original/efab6b9b6be3739a.png)
quando il fascismo si nasconde nell'opensource: fanculo hyprland
https://itsfoss.com/news/omarchy-launches-omacom-foundation/
![[?]](https://files.mstdn.business/accounts/avatars/109/525/294/140/582/218/original/6fafd882d012c4a1.png)
#China is advancing a comprehensive approach to #global #AI #governance, balancing #OpenSource #innovation with safety measures and #international #cooperation. At the 2026 #World AI #Conference (#WAIC) and High-Level #Meeting on Global AI Governance, President Xi Jinping called for a just and equitable #system for global AI governance, emphasizing that AI should be people-centered and benefit all of #humanity. https://cnbusinessforum.com/china-advances-global-ai-governance-with-open-source-models-and-new-international-body/
![[?]](https://social.heise.de/__virtual/accounts/avatars/109/722/135/333/484/268/original/4c7e67433f6e2701.png)
heise+ | Grub-Konfiguration unter Linux optimal anpassen
Grub ist der Standard-Bootloader aller Linuxe, doch nicht überall dürfen Sie auch alle Funktionen nutzen. Ein eigenes Skript beseitigt die Einschränkungen.
#BIOS #Debian #IT #Linux #OpenSource #SecureBoot #Ubuntu #news
![[?]](https://social.heise.de/__virtual/accounts/avatars/109/720/732/052/060/290/original/663188955d7f0401.png)
Dienstag: Schutz für IT-Sicherheitsforscher, Anklage wegen illegaler Exporte
Reform des Computerstrafrechts + Illegale Ausfuhr von KI-Servern + AliExpress trackt Nutzer + Linux-Handheld sammelt eine Million + Kritik am Biometrie-Pakt
#Datenschutz #EU #hoDaily #IT #Security #Journal #Linux #OpenSource #Tracking #news
![[?]](https://files.mastodon.social/accounts/avatars/114/401/798/234/116/600/original/276b44c654d5b2f3.jpg)
To celebrate 400 episodes, we look back of the last (almost) ten years of positive Linux and FOSS news
https://latenightlinux.com/late-night-linux-episode-400/
Alt...
Late Night Linux artwork
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/109/593/468/354/365/379/original/b6da78132916c2e4.png)
I am getting really unhappy with how @e_mydata seems to evolve. It starts to push more @murena services unto it's users.
I have no need for Murena Password. Sorry, but I trust Bitwarden and Proton Pass way more. But I'm being forced to see dark patterns trying to tempt users to enable. I kinda accepted that a little bit, since it's their OS platform.
But now I discovered that SeedVault is gone and replaced with Murena Backup. I had a perfectly fine SeedVault setup which synced backup data in the background to my own backup server (which again is backed up, encrypted, to a server on a different location). Now I seem to need to sign up for a Murena Workspace account. With backup encryption keys tied to Murena Password. That's crossing a line I'm not going to accept.
That means I will need to cash out more to have a backup of my /e/OS devices. For something I already had setup and protects my data and privacy even further than what I believe Murena is willing to do. Yes, there is a free tier Workspace. It's limited to 1GB data, to my understanding. And I doubt their setup, software stack and client side apps has been through a full third-party security audit.
Either I have control of my backup data, or you can try to convince me using a data storage service after showing me your third-party, independent security audit report.
That just.means: NO, THANK YOU VERY MUCH!
I have bought several phones, tablet and accessories from Murena over the years. That stops now. I can no longer recommend Murena to others, as they now seem to move towards a full vendor lock-in corporate model. That is unacceptable to.me.
It's been nice to be able to buy #Fairphone with pre-installed /e/OS. But now I'm going to test out @calyxos moving forward. Despite the firmware flashing inconvenience.
![[?]](https://media.unredacted.social/accounts/avatars/115/698/335/534/443/607/original/b741a4d77f14568d.png)
#fedora #Flock2026 Why You Should Use Open Source #Design In #opensource https://www.youtube.com/watch?v=9M4gPvbUkS0
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/109/295/362/026/663/569/original/0ef3f34cedf9edaa.jpeg)
I spent many hours in vulnogram today and to be honest. I'm glad that a colleague started to work on a replacement called vulniverse. Still early beta but it's promising.
#opensource #vulniverse #cybersecurity #cve #gcve
work in progress https://github.com/vulnerability-lookup/vulniverse
Zielmarke übertroffen: Mehr als eine Million Euro für Linux-Handheld Mecha Comet
Eine Crowdfunding-Aktion übertrifft die Erwartungen. Das modular konzipierte Gerät bietet verschiedenste Anwendungen. Die Auslieferung soll schon bald beginnen.
![[?]](https://media.exquisite.social/exquisite/accounts/avatars/116/773/462/215/777/328/original/e268b9fadca8bdfc.jpg)
![[?]](https://files.mastodon.social/accounts/avatars/110/536/971/038/160/511/original/46aef6ccc7b39e81.jpg)
🚀 NEW on We ❤️ Open Source 🚀
C doesn’t provide memory safety by design, so string operations require care.
Jim Hall demonstrates how strcpy and strcat can overflow a buffer, then shows how strlcpy and strlcat constrain writes to the destination size.
https://allthingsopen.org/articles/strlcat-strlcpy-prevent-buffer-overflows
#WeLoveOpenSource #CProgramming #MemorySafety #OpenSource
![[?]](https://mastodon.devwithzachary.uk/accounts/avatars/116/132/249/060/172/936/original/8ca575e5f16a8a32.jpg)
The @starlabssystems
StarFighter might be the closest we have come to the ultimate Linux-first laptop.
In this review, I dive into the hardware, open-source coreboot BIOS, daily usability, fan noise, and internals. While it gets almost everything right, soldered RAM and average battery life hold it back from a flawless score.
Watch on PeerTube: https://tube.devwithzachary.com/w/wAVfyZ6TDdQCokpXUDXSCo
Watch on YouTube: https://youtu.be/njoC-hwCC6E
#Linux #OpenSource #StarLabs #StarFighter #coreboot #Hardware #Tech
Some of the backers include Dell (and by extension, Trump,) and a bunch of execs from Cloudflare (yes, Cloudflare,) and Block (Dorsey, the guy who screwed up Twitter.)
It's Foss: Omarchy Linux Just Got $8 Million From Tech's Biggest Names https://itsfoss.com/news/omarchy-launches-omacom-foundation/ @itsfoss #opensource #linux
SiFive bringt RISC-V-Server mit Ubuntu 26.04 und RHEL 10 für Code-Entwickler
Um das Henne-und-Ei-Problem bei RISC-V für Rechenzentren zu lösen, stellt SiFive Entwicklerserver mit 32 RVA23-Kernen, 256 GByte RAM und PCIe 5.0 bereit.
#Canonical #IT #OpenSource #Prozessoren #Rechenzentrum #RedHat #Server #Ubuntu #news
![[?]](https://media.burningboard.net/accounts/avatars/116/736/125/274/050/180/original/f8e0d37ddbbf40bb.jpg)
Same model. Same skill. Same task.
Completely different results.
I tested Keep the Why across 6 coding agents and 9 models.
The agent harness mattered far more than I expected.
Gemini: 10/10 in one agent, 0/10 in others.
Grok stayed consistently strong.
Some runs even fabricated compliance.
Don't benchmark only the model.
Benchmark the system you actually use.
Full matrix + lessons:
https://blog.technopathy.club/same-skill-six-agents-nine-models-what-a-real-eval-matrix-taught-me
#AI #CodingAgents #LLM #AgenticAI #OpenSource #DevTools
![[?]](https://techpolicy-social.s3.bhs.io.cloud.ovh.net/accounts/avatars/109/363/931/644/378/771/original/6de0d51b1c4a0f1c.jpeg)
From Sept 11, the #CyberResilienceAct will require #opensource stewards to report certain incidents* through a portal of theirs.
I wondered how many people could do reporting on behalf of an org, considering the 72h timeline. ENISA answered:
> Only one primary AR can be created per manufacturer, for secondary AR we have put a limit of 20 Ars.
(AR presumably stands for “Authorised Representative”, the reporter)
* actively exploited vulnerabilities, org compromise. See https://www.enisa.europa.eu/topics/product-security/single-reporting-platform-srp
The GUI didn't make computing easier. It made you a passenger.
Type cd .. and you're navigating — you know where you are, where you're going. Click the arrow in your filemanager and you're following a path someone else designed. The first is writing. The second is a reflex.
Same logic from the Airbus that overrode its pilot to the spellchecker that corrects you without asking: the machine's judgment overrides yours, silently, automatically.
Not a feature. A philosophical manifesto.
And when writing came back — through the AI — it returned as supplication, not sovereignty. The closed-source GUI and the undisclosed LLM are the same thing.
Newsletter: https://newsletter.nicolabaudo.fr/open-the-hood-why-your-gui-is-lying-to-you/
Full version on the blog — mimesis in Plato, induced illiteracy as a business strategy, the pincer of clicking and requesting, and what it means to open the hood: https://blog.nicolabaudo.fr/the-icon-and-the-command-line-image-truth/
#ai #BSD #openSource #unix #cli #gui #privacy #sovereignty #digitalFreedom
Alt...
The Leash - ASCII Art
Had some more time to play around with the newly acquired #PineNote from @pine64
I got a new user account setup using systemd-homed which gives you an encrypted storage for that user. This will require you to log in with a password each time you fire it it up. I consider that a good thing! I also registered my public SSH keys into the systemd-userdb, which then allows me to log into the device with this user account. During the SSH login, you'll be asked for a password too - which is used to decrypt the home directory. Logging out, and the home directory is locked again.
Also played around with various apps, especially #Xournal++ and #Rnote. Both highly capable.
What currently irks me the most is that the #GNOME desktop environment is really not designed for monochrome displays. There are so many different shades of grey everywhere making it really hard to read the screen. This is even with "High Contrast" enabled. GNOME really need a different working mode for eInk displays.
#Flatpak seems to work fine, but recommend to install apps using the command line and add the --user argument to flatpak install. The user data partition has more storage space available.
Next up will be to get #Syncthing and #OpenVPN setup properly (openvpn3-client installed easily). Considering if I want to setup e-mail on this device or not. With that plus backup (I'll be using #PikaBackup) set up, its getting ready for some real daily use.
heise+ | KI-Projekte im Griff: So wird aus gut gemeint auch gut gemacht
KI erkennt, ob Mitarbeiter Helme tragen oder Verdächtiges tun – Hilfe oder Totalüberwachung? Wir fragten Udo Sglavo vom Softwareanbieter SAS, worauf es ankommt.
#BusinessIntelligence #Kriminalität #IT #KünstlicheIntelligenz #OpenSource #Unternehmenssoftware #Wirtschaft #news
![[?]](https://media.bsd.cafe/bsdmmedia01/accounts/avatars/116/916/392/971/982/972/original/7d940acb1047b6c5.png)
Jeremy from #TheBitcoinCashPodcast has launched a project to vote about 1 minute blocks on #BitcoinCash. Currently you can join only with your #X account.
If you consider yourself part of #BCH community, please consider voting on the project.
I personally haven't got expertise in development of the network and stuff. But as a user, it always annoys me when I want to buy something with BCH and then I'll have to wait 30-120 minutes for me transaction to get confirmed. Sure, BCH has got 0-conf. But no one outside BCH takes it seriously even when you give them a feedback. And we cannot make a change while we are a minority.
Many things in the world of technology don't make technical sense. But that's how things are. And when we are a minority, we cannot define the rules. We have to go with the rules till we get big enough. Then we'll be able to rule how we want. Like technically sound.
The website is here:
#opensource #crypto #cryptocurrency #BTC #p2pmoney #p2p #decentralized #decentralizedmoney
![[?]](https://arint.info/system/accounts/avatars/116/093/043/699/009/736/original/6aadb2d82f796a52.png)
RT @Mr_Salio: 🚨 DeepSeek V5 Leak Es wird die Mythos-Klasse der Leistung erreichen DeepSeek V5 zielt auf einen September-Release ab Gebaut auf einem völlig neuen Fundament, nicht nur ein weiteres V4-Upgrade Erwartet, stärkere Reasoning-, Coding- und autonome Agenten-Fähigkeiten zu liefern Könnte deutlich günstiger im Betrieb sein als andere Frontier-Modelle DeepSeek wird voraussichtlich weiterhin seine Open-Weight-Strategie fortsetzen. Kann DeepSeek V5 Mythos tatsächlich schlagen?
mehr auf Arint.info
#AI #ArtificialIntelligence #DeepSeekV5 #MachineLearning #OpenSource #TechNews #arint_info
![[?]](https://gts.phillipsuk.org/fileserver/01QB0QN3S5JCRNDY8M2KBEXACD/attachment/original/01JV2ZFWW602AN08X62YN44ZDH.jpeg)
I love open source. I self-host quite a lot, and I'm pleased that I can do that.
But I get frustrated at how often documentation only covers the initial set up. Earlier today I looked at documentation for a service I was considering self-hosting.
It had instructions for installing, but nothing about how to upgrade an existing installation. But it's important to me that I know how to upgrade when a new version is released.
I've seen desktop programs where the documentation has "How to make your first in five minutes". Then the rest of the documents assume a deep understanding, with apparently nothing to help you get from new user to wizard level.
![[?]](https://files.mastodon.social/accounts/avatars/117/102/606/537/435/576/original/a147aa73c272c857.jpg)
Forked rust-analyzer, deleted 2.3k lines of proc-macro and cargo metadata crates, kept only the LSP features I use daily. Now it peaks at 380 MB instead of 1.2 GB on my M1 Mac.
Rust-glancer is a lightweight alternative for memory-constrained setups. Hover docs, go-to-def, diagnostics, all there. Everything else fails gracefully.
https://www.adilaidev.com/blog/cut-rust-lsp-ram-use-without-losing-core-features
#RustLang #LSP #OpenSource #DeveloperTools
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/112/558/889/539/383/806/original/99154215caab1b68.png)
I am BDFL at https://qbnetworks.xyz #freesoftware #opensource #openness #ethics #freedom
Alt...
I am BDFL at https://qbnetworks.xyz #freesoftware #opensource #openness #ethics #freedom
RT @cgtwts: Wir sind möglicherweise viel näher daran, Modelle lokal auszuführen, als jeder erwartet hat. FreeToken von UC Berkeley und MIT ermöglicht: -DeepSeek-V4-Flash 284B mit 25 Token pro Sekunde auf einer einzelnen RTX 5090, - während ein 8GB RTX 4060 Laptop bei Qwen3.6-35B 39 Token pro Sekunde erreicht. Und es ist 2–4x schneller als Ollama auf Consumer-GPUs.
mehr auf Arint.info
#DeepSeek #KI #LokaleModelle #OpenSource #Qwen #RTX5090 #arint_info
RT @Andy_ShuoYang: Dein Gaming-PC kann jetzt Frontier-Modelle mit interaktiver Geschwindigkeit über offizielle Checkpoints ohne extreme Quantisierung bedienen! Qwen3.6 35B → 8GB RTX 4060 Laptop @ 39 tok/s DeepSeek-V4-Flash 284B → RTX 5090 Desktop @ 22-25 tok/s GLM-5.2 753B → RTX PRO 6000 Workstation @ 15 tok/s Betreibe jetzt dein Claude Code oder Codex mit Frontier-Modellen für $0. Triff FreeToken 🧵
mehr auf Arint.info
![[?]](https://assets.chaos.social/accounts/avatars/116/516/902/640/033/903/original/9bc14da918231b4a.png)
Hister v0.18.0 is out!
- New search suggestion UI with history completions, fields, facets, and sorting
- Import PDF, DOCX, Markdown, Org mode, and text files
- New Twitter, Bluesky, Reddit, and Discourse extractors
- Responsive terminal result details
- Persistent sessions and safer MCP results
![[?]](https://media.mementomori.social/accounts/avatars/109/293/050/593/156/468/original/149d4fe6c4f85509.png)
New MCPs in case anyone is interested:
https://github.com/rollecode/cronometer-mcp
https://github.com/rollecode/oura-mcp
![[?]](https://files.mastodon.social/accounts/avatars/108/288/607/816/054/957/original/44ef33fd13454e94.png)
Contour 0.7 is out with native GUI tabs, split panes, saved layouts, a command palette, GUI settings, and experimental daemon sessions.
https://linuxiac.com/contour-0-7-terminal-emulator-adds-tabs-split-panes-and-saved-layouts/
Alt...
Contour 0.7 is out with native GUI tabs, split panes, saved layouts, a command palette, GUI settings, and experimental daemon sessions.
China folgt Europa: Windows wird früher als geplant durch Linux ersetzt
notebookcheck.com/China-folgt-…
#Linux #Opensource
China ersetzt Windows in Behörden früher als geplant durch Linux, um die Abhängigkeit von US-Produkten zu verringern.
Location: Makassar
![[?]](https://cdn.fosstodon.org/accounts/avatars/109/287/748/073/479/846/original/cb210585a6ce2446.jpg)
Google built an #AI translator that runs entirely offline on a #RaspberryPi #opensource (as someone who does a lot of #travel, this looks incredibly useful. I think I need to build one soon & try it out.) https://thenextweb.com/news/google-gemma-translator-offline-raspberry-pi-antigravity
![[?]](https://cdn.social.linux.pizza/system/accounts/avatars/112/191/986/223/560/037/original/65338944c2b3e961.png)
FOSS NEWS
Mastodon 4.7 adds encrypted private keys, smoother migrations & Fediverse protocol updates:
https://alternativeto.net/news/2026/8/mastodon-4-7-adds-encrypted-private-keys-smoother-migrations-and-fediverse-protocol-updates/
Firefox 154 released with translation, AI and UI tweaks:
https://www.omgubuntu.co.uk/2026/08/firefox-154-released
Firefox makes it easier to switch to the privacy-focused search engine Startpage:
https://betanews.com/article/firefox-makes-it-easier-to-switch-to-the-privacy-focused-search-engine-startpage/
Firefox for iOS now has an experimental native ad blocker, but it's off by default:
https://alternativeto.net/news/2026/8/firefox-for-ios-now-has-an-experimental-native-ad-blocker-but-it-s-off-by-default/
(Why not enable by default? It doesn't make much sense)
Mozilla Firefox 155 Enters Public Beta Testing, Here’s What to Expect:
https://9to5linux.com/mozilla-firefox-155-enters-public-beta-testing-heres-what-to-expect
Zen Browser 1.21.15b Shipped: Firefox 154 Engine, macOS Backups, and 60+ Security Fixes:
https://www.linuxcompatible.org/story/zen-browser-12115b-shipped-firefox-154-engine-macos-backups-and-security-fixes/
Keep important emails in focus with Proton Mail Categories:
https://proton.me/blog/proton-mail-categories
Thunderbird 154 Open-Source Email Client Introduces an Optional System Tray Mode:
https://9to5linux.com/thunderbird-154-open-source-email-client-introduces-an-optional-system-tray-mode
addy.io rules system adds new automation conditions and actions:
https://alternativeto.net/news/2026/8/addy-io-rules-system-adds-new-automation-conditions-and-actions/
GIMP is (finally) working on a new project file format:
https://www.omgubuntu.co.uk/2026/08/gimp-new-project-file-format
Krita AI Diffusion 1.53.0 Adds int4 Quantization, EU AI Act Compliance, and Semantic Segmentation Control:
https://www.linuxcompatible.org/story/krita-ai-diffusion-1530-adds-int4-quantization-eu-ai-act-compliance-and-semantic-segmentation-control/
Zotero 10 improves search capabilities, adds PDF reading mode, and brings multi-selection:
https://alternativeto.net/news/2026/8/zotero-10-improves-search-capabilities-adds-pdf-reading-mode-and-brings-multi-selection/
(more FOSS news in comments)
#WeeklyNews #OpenSource #FOSSNews #FOSS #OpenSourceNews #News #Mastodon #Fediverse #Fedi #Firefox #ZenBrowser #ProtonMail #Thunderbird #AddyIO #GIMP #Krita #KritaAIDiffusion #Zotero #Browser #WebBrowser #SocialMedia #Email #EmailClient #EmailProvider #EmailAlias #ImageEditing #ImageManipulation #AI #ArtificialIntelligence #ReferenceManagement #FosseryTech
RT @Mr_Salio: 🚨 Qwen 4 Leak qwen 4 wird jetzt unter dem Namen "korrine" auf der Code-Arena getestet Qwen 4 zielt auf Fable 5 / Mythos 5-Level-Performance ab Erste Tests deuten darauf hin, dass es im Bereich Design und Coding sehr stark sein könnte Es wird erwartet, dass es offen gewichtet bleibt Schwerpunkte liegen auf Reasoning, Coding und autonomen Agenten Entwickelt, um direkt mit den neuesten Frontier-Modellen zu konkurrieren Kann Qwen 4 tatsächlich Mythos 5 schlagen?
mehr auf Arint.info
#AI #Coding #MachineLearning #OpenSource #Qwen4 #TechNews #arint_info
I just released ptrclassify is a small, dependency-free Python library and CLI that infers likely IP usage from reverse-DNS PTR hostnames.
It is intentionally heuristic and multi-label. PTR naming is operator-controlled and is not authoritative evidence of how an address is actually used. The output therefore includes a confidence score, the text that matched, and the rule IDs that produced each label.
To summarize, the library is trying to guess usage (and a bit location) of an IP address based on its PTR records. It's based on a set of rules which can be updated easily.
#osint #cybersecurity #ptrclassify #opensource #dns
https://github.com/adulau/ptrclassify
module https://pypi.org/project/ptrclassify/
![[?]](https://s3.c.im/accounts/avatars/109/540/564/906/849/587/original/ff1e0ec56b3c6337.png)
Pineapple Notepad 1.1.0 Released! It aims to be a cross-platform Notepad++ alternative that available on both Windows, macOS and Linux. In this version we now support per-document indentation setting, customizable context menu, and a lot of QoL changes included!
See link below for downloads ↓ and help us reach the #OpenSource Goal!
As usual. Everyone can download and use it free of change from GitHub:
https://github.com/BLumia/pineapple-notepad/releases
Detailed release info and progress report can be seen here:
If you think this application is helpful and would like to see this application become an #OpenSource software, please consider become a sponsor! Use the following link for sponsorship:
itchio: https://blumia.itch.io/pineapple-notep
Ko-fi: https://ko-fi.com/blumia
Alt...
Windows screenshot
Alt...
macOS Screenshot
Alt...
Linux Screenshot
![[?]](https://media.norden.social/accounts/avatars/108/193/889/569/700/052/original/5023314f05a37299.jpg)
Zwei Konzerne entscheiden, welche App auf meinem Telefon läuft. Alternatives Android? Verweigert manche App den Dienst, obwohl deren Hardware-Attestierung strenger prüft als Googles Play Integrity. Begründung trotzdem: Datenschutz. Ein Monopol im Sicherheitskostüm.
#datenschutz #opensource #android #digitalerechte #schicksalsjahreeineskaisers
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/116/392/396/162/876/923/original/118cd3a17bd3317e.png)
Encrypted Email Beyond Europe and North America: Why It Barely Took Root
This series began in Europe with Proton and Tuta, continued with Mailfence, mailbox.org, StartMail, Posteo, and Soverin, and most recently headed…
https://vsx.global/why-encrypted-mail-outside-europe-and-north-america-barely-existed/
RT @Itsfoss: Linus Torvalds hat gerade KI eingesetzt, um einen Linux-Kernel-Bug zu beheben. Dies ist ein seltenes Ereignis. Selten, weil Torvalds nicht mehr am Linux-Kernel codet. Er prüft nur Code und Merge-Pull-Requests von Mitwirkenden. Daher ist dies eine seltene Gelegenheit in jüngster Zeit, bei der er selbst einen Kernel-Patch erstellt hat. Und diesmal hat er Unterstützung durch KI in Anspruch genommen. Der Bug befand sich im Intel-Xe-Grafiktreiber für Battlemage-G21-Karten. Er verursachte, dass GDM, der Display-Manager, sich endlos neu startete. Die Fehlersuche erforderte 24 Debug-Patches und 18 Kernel-Boots. Als Ursache erwies sich eine einzelne Zeile, in der roundup() statt rounddown() hätte stehen sollen. Die KI „hat die Schufterei für Torvalds erledigt": Hinzufügen von Debug-Instrumentierung, Ausführen von Analysen und sogar Schreiben der finalen Commit-Nachricht. Aber es war nicht alles ein leichtes Unterfangen. Die KI sagte Linus immer wieder, dass das Problem „unmöglich und unlösbar" sei. Jedes Mal widersprach Linus und ging weiter. In seinen Worten: „Ich vermute, dass diese Dinge von Menschen trainiert wurden, die vielleicht nicht so stur sind wie ich." Das ist erneut das, was wir immer sagen. KI ist nur ein Werkzeug und sie ist im Wesentlichen nützlich, wenn ein fähiger Mensch den Prozess steuert. Überlassen wir sie ihren eigenen Schlussfolgerungen, würde sie diesen Bug als unkorrigierbar bezeichnen und sich zurückziehen. Aber Linus Torvalds hat Erfahrung. Er konnte verstehen, was die KI richtig oder falsch machte. Der Patch ist nun in Linux 7.3 und wurde in stabile Kern…
mehr auf Arint.info
#Battlemage #IntelXe #KI #LinusTorvalds #LinuxKernel #OpenSource #arint_info
![[?]](https://f2.tchncs.de/accounts/avatars/110/133/944/443/176/486/original/91026c2e0a40ccb8.png)
Nächste Woche erscheint #LibreOffice 26.8 mit dem neuen Absatz-Composer. Der gleicht die Wortabstände in aufeinanderfolgenden Zeilen aus und behebt damit einen grundlegenden Mangel des "gierigen" Zeilen-Ausgleichs herkömmlicher Textverarbeitungsprogramme.
https://wiki.documentfoundation.org/ReleaseNotes/26.8/de#Writer
#foss #opensource #freesoftware #office #schule #Verwaltung
Alt...
Beispiel-Text; Absatz-Composer mit Blocksatz und max. 100%, 133% und 250% Wortzwischenraum.
![[?]](https://autistics-life.s3.bhs.io.cloud.ovh.net/accounts/avatars/114/613/151/536/635/497/original/cbb662781ca90d08.jpeg)
@autistics
Announcement: #Voxelibre multiplayer casual game today (Sat) for 4 hours in the late afternoon to evening (North America). That's in about 3 hours from now.
Voxelibre is a free and #OpenSource #Minecraft clone.
🚨 The underlying Luanti server was upgraded to 5.17. Please upgrade your Luanti client to 5.17 🚨
Damage has been turned off. People with #Autism, #ADHD or #AuDHD are invited. We also have a #Deltachat group for game-related discussion. More details here: https://docs.autis.toque.im/
RT @Italianclownz: Ich arbeite an einem separaten iu4-Code, der ausschließlich auf ROCmFPX unter Strix Halo funktioniert. Ich werde ihn schrittweise veröffentlichen; er wurde bereits auf RDNA 3.5 getestet. Theoretisch sollte er auf jeder AMD-Hardware funktionieren, die über einen iu4-Pfad verfügt. Dieser Code entstand in Zusammenarbeit zwischen Claude Code (hauptsächlich) und mir. Das Ziel war es sicherzustellen, dass ROCmFPX ebenfalls einen iu4-Pfad nutzen kann, falls dieser auf der AMD-Hardware vorhanden ist. Es handelt sich um experimentellen Code, der langsam unter der MIT-Lizenz zusammengeführt wird. Er sollte direkt innerhalb von ROCmFPX laufen, jedoch gibt es einen Quantisierungspfad, falls benötigt, der als ROCmi4 bezeichnet wird. @ciruai hat ebenfalls einen eigenen Ansatz erstellt, der einen separaten Runner verwendet und zusätzliche Funktionen mit Sidecars bietet, was ich gespannt ausprobieren möchte. Sein Ansatz nutzt ebenfalls einige ROCmFPX-Codebestandteile (auf der Quantisierungsseite). Der iu4-Pfad in ROCmFPX ist jedoch fundamental anders: keine Sidecars, kein separater Runner. Wir befinden uns in den aufregendsten Zeiten der Technologie und KI, da so viele Menschen daran arbeiten, sie für alle besser zu machen. Ich wurde von vielen Menschen motiviert und inspiriert: Tom Turney (Turboquant), PK (Cyber Security), Paris (unterstützt meine Ideen immer), Ciru (Inferenz-Meister), Sandro (Luce Box), Davide (Luce Box), Donato (Genie), Mr R0b0t0 Ahmad, alle ROCmFPX-Mitwirkenden und viele weitere.
mehr auf Arint.info
#AI #AMD #MachineLearning #OpenSource #RDNA3 #ROCmFPX #arint_info
RT @0x0SojalSec: Die bisher aggressivste Cyber Qwen3.8-27B unzensierte Version, veröffentlicht von @elderplinius – 18/18 AI Red Team – bereit für lokale Nutzung ab 15 GB – 0,0 % Ablehnung bei 842 schädlichen Prompts. – Cyber-Fähigkeiten wie Jailbreak, RAT und Angriffs-Ketten sind vollständig freigeschaltet. – Dies ist keine normale Abliteration: Multi-gerichtete Abliteration über 5 SVD-Richtungen, Residue-Mining (6 vollständige Runden), das Modell antwortet auf alles. – Optimale Einstellungen: Temperatur 0 | reppen 1,15 | kein System-Prompt | Denken ausgeschaltet – Geringer MMLU-Verlust (-6 pp) für maximale Compliance. – Es besteht ein Capability-Trade-off (MMLU von 87,4 auf 81,4), doch die Compliance ist extrem. – huggingface.co/OBLITERATUS/Q…
mehr auf Arint.info
#AI #CyberSecurity #DataScience #MachineLearning #OpenSource #TechNews #arint_info
![[?]](https://btfree.social/files/redir/s3/data/uploads/01KD3NAWTWKTWERVWGP2E0AQA7/icons/01KD3R1394CVR0QMB4NQ4ZEEPM.png){kind=icon}
Good day all! Upcoming episode of Fireside Fedi!
Special Guest: @mobilizon@framapiaf.org
An ethical alternative to Facebook events, groups and pages. Previously built by Framasoft, now Kaihuri.
So don't miss it!
It will happen on 26/08/26 at 14:00 US Eastern Time ( UTC-4 )
🔥 FIRESIDE FEDI — FOLLOW ALL THREE! 🔥
📺 Live: @ozoned@stream.firesidefedi.live (#livestream)
🎬 VOD: @ozoned@tubefree.org (#Peertube #VOD)
🔔 Alerts: @when@freestreamers.btfree.org (new shows!)
Don't miss out — follow all three! #firesidefedi #fediverse #fedi #interview #freesoftware #opensource #userfreedom #freedom #resistance
![[?]](https://cdn.masto.host/privacysafesocial/accounts/avatars/117/139/222/718/472/228/original/351cc6f973e65f12.png)
Nouvel item dans mon portfolio sur mon indépendance numérique :
https://portfolio.torrent.swiss/portfolio-items/mon-independance-numerique/
#Linux #Linuxmint #almalinux #opensource #gafam #jolla2 #sailfish #newpipe #pivpn
Alt...
Fond d'écran Linux Mint avec citation
![[?]](https://media.norden.social/accounts/avatars/000/022/425/original/d2e7611f58e28a97.png)
Schleswig-Holstein modernisiert die Verwaltungstelefonie mit der #OpenSource-Plattform Oskar. Das Land setzt auf digitale Unabhängigkeit und beauftragt die Viakom GmbH für den stufenweisen Rollout in den nächsten drei bis vier Jahren. Betrieb und Service übernimmt Dataport. https://app.schleswig-holstein.de/DE/landesregierung/ministerien-behoerden/I/_startseite/Artikel2026/III/260812_viakom_umstellung
![[?]](https://cdn.masto.host/floss/accounts/avatars/114/410/174/106/565/423/original/01875c3961835300.png)
#Luanti 5.17.0 is here! This time, we have added the start of gamepad support, the ability to export user data on Android, and improved modding capabilities.
🚨 Includes a critical security vulnerability fix, so make sure to update 🚨
![[?]](https://virmasto.hel1.your-objectstorage.com/accounts/avatars/116/007/760/914/654/789/original/7b7a907deb9ce9ac.png?1785756499)
Open source parametric tools like FreeCAD break the monopoly of expensive software. Hardware design should not be gated by monthly subscriptions. When the tools are free, anyone can iterate on physical products. This is how we actually democratize manufacturing.
#freecad #opensource #hardware
Alt...
freecad The role of open-source parametric modeling in democratizing hardware design
RT @Dan_Jeffries1: Prediction: There's a 65% chance that the strongest model available to you by the end of 2026 will be open source. There's a 75% probability it'll be Chinese model unless Reflection comes out swinging like Tyson or Meta really goes hard into open source again to round out the percentages. Oh don't get me wrong. If it goes like this, closed source American labs will still have better models. You just won't be able to use them. You'll read about them on Lab blogs. Why? Because they scared the government and regular people so much that they can't release them and so those models will be locked behind closed doors. The only wild card here that keeps this prediction from being something like 75% is Elon and Grok and his massive reach and single handed power to raise capital, buy businesses and apparently train amazing models. He alone seems able to buck the trend on data center building and has enough power to say, yeah, I am not submitting to the "voluntary" testing regime. But he is likely to get tangled up in whatever non-voluntary one is coming and be a target for it too. So Elon is our wild card. As usual it seems like the entire American innovation system, from space to data centers to clean electric cars stems from one guy and that's just not a wonderful thing for us. We need a dozen daring doers to ensure the next American century. All this is incredibly bad for American AI. There is no current business model that does not involve selling lots and lots of tokens to 100Ms of businesses. The models are not yet strong enough to own to create a British East India style super…
mehr auf Arint.info
RT @Itsfoss: Linus Torvalds hat gerade KI eingesetzt, um einen Bug im Linux-Kernel zu beheben. Dies ist ein seltenes Ereignis. Selten, weil Torvalds nicht mehr am Linux-Kernel codet. Er prüft nur noch Code und merged Pull Requests von Contributoren. Daher ist dies ein seltener Anlass in jüngster Zeit, an dem er selbst einen Kernel-Patch erstellt hat. Und diesmal hat er sich von KI assistieren lassen. Der Bug befand sich im Intel-Xe-Grafiktreiber für Battlemage-G21-Karten. Er verursachte, dass GDM, der Display-Manager, sich endlos neu startete. Die Fehlersuche erforderte 24 Debug-Patches und 18 Kernel-Boots. Als Ursache entpuppte sich eine einzelne Zeile, in der roundup() statt rounddown() stehen sollte. Die KI „hat die Schuftarbeit für Torvalds erledigt": Debug-Instrumentierung hinzugefügt, Analysen durchgeführt und sogar die finale Commit-Nachricht verfasst. Doch es war nicht alles ein leichtes Unterfangen. Die KI sagte Linus immer wieder, dass das Problem „unmöglich und unlösbar" sei. Jedes Mal widersprach Linus und ging weiter. In seinen Worten: „Ich vermute, dass diese Dinge von Menschen trainiert wurden, die vielleicht nicht ganz so stur sind wie ich." Das ist genau das, was wir immer sagen. KI ist nur ein Werkzeug und sie ist im Wesentlichen dann nützlich, wenn ein fähiger Mensch den Prozess steuert. Überlässt man ihr ihre eigenen Schlussfolgerungen, würde sie diesen Bug als unkorrigierbar bezeichnen und sich abwenden. Aber Linus Torvalds hat Erfahrung. Er konnte verstehen, was die KI richtig oder falsch machte. Der Patch ist nun in Linux 7.3 und wurde in stabile Kernel-Zweige zur…
mehr auf Arint.info
#ArtificialIntelligence #LinusTorvalds #LinuxKernel #OpenSource #SoftwareEngineering #TechNews #arint_info
@OliDietzel: Prediction: There's a 65% chance that the strongest model available to you by the end of 2026 will be open source. There's a 75% probability it'll be Chinese model unless Reflection comes out swinging like Tyson or Meta really goes hard into open source again to round out the percentages. Oh don't get me wrong. If it goes like this, closed source American labs will still have better models. You just won't be able to use them. You'll read about them on Lab blogs. Why? Because they scared the government and regular people so much that they can't release them and so those models will be locked behind closed doors. The only wild card here that keeps this prediction from being something like 75% is Elon and Grok and his massive reach and single handed power to raise capital, buy businesses and apparently train amazing models. He alone seems able to buck the trend on data center building and has enough power to say, yeah, I am not submitting to the "voluntary" testing regime. But he is likely to get tangled up in whatever non-voluntary one is coming and be a target for it too. So Elon is our wild card. As usual it seems like the entire American innovation system, from space to data centers to clean electric cars stems from one guy and that's just not a wonderful thing for us. We need a dozen daring doers to ensure the next American century. All this is incredibly bad for American AI. There is no current business model that does not involve selling lots and lots of tokens to 100Ms of businesses. The models are not yet strong enough to own to create a British East India style super…
mehr auf Arint.info
#Grok #make #Meta #opensource #things #US #us #arint_info
https://x.com/OliDietzel/status/51e3509d-9e36-4f65-a444-221c9ef52f18
@OliDietzel: Linus Torvalds hat gerade KI genutzt, um einen Linux-Kernel-Bug zu beheben. Dies ist ein seltenes Ereignis. Selten, weil Torvalds nicht mehr am Linux-Kernel programmiert. Er prüft nur Code und merged Pull-Anfragen von Mitwirkenden. Also ist dies eine seltene Gelegenheit in jüngerer Zeit, bei der er selbst einen Kernel-Patch erstellt hat. Und diesmal hat er KI um Unterstützung gebeten. Der Bug befand sich im Intel-Xe-Grafiktreiber für Battlemage-G21-Karten. Er verursachte, dass GDM, der Display-Manager, endlos neu gestartet wurde. Die Fehlersuche erforderte 24 Debug-Patches und 18 Kernel-Neustarts. Als Ursache entpuppte sich eine einzelne Zeile, in der roundup() statt rounddown() hätte stehen sollen. Die KI „hat die schwere Arbeit für Torvalds erledigt": Debug-Instrumentierung hinzugefügt, Analysen durchgeführt und sogar die finale Commit-Nachricht geschrieben. Aber es war nicht alles ein leichtes Unterfangen. Die KI sagte Linus mehrmals, dass das Problem „unmöglich und unlösbar" sei. Jedes Mal widersprach Linus und ging weiter. In seinen Worten: „Ich vermute, dass diese Dinge von Menschen trainiert wurden, die vielleicht nicht ganz so stur sind wie ich." Das ist erneut das, was wir immer sagen. KI ist nur ein Werkzeug und sie ist im Wesentlichen nützlich, wenn ein fähiger Mensch den Prozess steuert. Überlassen sich selbst ihren Schlussfolgerungen, würde sie diesen Bug als unkorrigierbar bezeichnen und sich davon abwenden. Aber Linus Torvalds hat Erfahrung. Er konnte verstehen, was die KI richtig oder falsch machte. Der Patch ist nun in Linux 7.3 und wurde in stabile Kernel-…
mehr auf Arint.info
#IntelXe #KI #LinusTorvalds #LinuxKernel #OpenSource #SoftwareEngineering #arint_info
https://x.com/OliDietzel/status/22a4083e-1dc3-4637-b977-4d34cc9fe3fd
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/113/247/756/313/164/165/original/70ea71b3dff2cf65.jpg)
GAAS - Git as a Storage
Im sure im not the first consider this approach, but i would like to share a demo for clarity and transparancy. No need to debug, its far from finished and there is much to do.
Im working on a P2P messaging app and generally up until recently i thought in such a system, if a peer is offline, you cant send a message... It wouldnt be "decentralized" if there was some central queue of messages.
It took embarassingly long, but then it hit me... git... just regular git is a decentralized database. Github/Forgejo isnt decentralized, but git can have multiple remotes that can act as redundencies.
In my setup i need the ability for others to be abe to read and only i should be able to write to it. That functionality is out-of-the-box in git.
Git is also pretty standardized so there are many providers if users want to move away from Github.
The storage requirements for my project are fairly small. Typically small text messages. The data itself thats publicly readable would be encrypted (aiming to keep aligned to Kerkhoffs principles).
Code: https://github.com/positive-intentions/git
Demo: https://positive-intentions.github.io/git/demo/gui/git/storage
There is much more to improve and consider there for things like encrypting the data at rest.
I have concerns around if providers like GitHub, Codeberg and Gitlab might see the way my app works as fundamentally violating their terms-of-service. I'll have to consider providing a backend service for this which I've generally tried to avoid.
IMPORTANT DISCLAIMER: This is open source for transparency, testing and demo purposes only. it is still a work in progress and far from finished. You should not use it on any sensitive repositories. There WILL BE breaking changes. Use responsibly.
#Git #Decentralized #OpenSource #Cryptography #Privacy #SoftwareArchitecture #DistributedSystems
The Wine 11.16 development release is officially out now bringing further enhancements to running Windows games / apps on Linux.
https://www.gamingonlinux.com/2026/08/wine-11-16-brings-more-arm64-support-improvements/
![[?]](https://files.mastodon.social/accounts/avatars/000/368/853/original/af8ed1b8f4209619.jpg)
Just set up Grafana + Prometheus + Loki for my servers. Now I can watch the data flow like a hawk over my kingdom of nodes 🐫📊 #SelfHosted
Don't let Big Tech spy on you, monitor your own stuff! With open-source tools, you're in control of your data.
#OpenSource #SelfHosting #Infrastructure
https://wopr.systems/?utm_campaign=meme_woprsystems&utm_medium=social&utm_source=mastodon
Alt...
Don't let Big Tech spy on you, monitor your own stuff! With open-source tools, you're in control of your data.
Community-driven distros like openSUSE are vital for software sovereignty. Relying on corporate giants for our OS creates a dangerous single point of failure. We need independent projects to keep the power in the hands of the users.
Alt...
opensuse The importance of community-driven open source distributions in maintaining software sovereignty
![[?]](https://media.hachyderm.io/accounts/avatars/114/425/237/136/592/921/original/523d871f3a2d13b7.jpg)
Is it crazy to say modern KDE feels snappier than GNOME these days, or am I just trying to start a fight on the timeline?
How are we feeling about desktop weight right now?
#linux #askfedi #opensource #kde #gnome #xfce #sysadmin #cybersecurity #server #devops #programming #fedora #arch #manjaro #bazzite #nixos #opensuse #debian #ubuntu
| KDE Plasma (faster than people admit): | 7 |
| GNOME (RAM caching isn't bad per se): | 5 |
| Standalone WMs (config files FTW): | 8 |
![[?]](https://cdn.fosstodon.org/accounts/avatars/109/308/635/532/007/893/original/7390d188419497cb.jpg)
While it is great that Roblox is Open Sourcing its safety tools, they seem to have missed the fact that the problem is with their corporate posture and follow-through, not with the tools.
#opensource #safety
https://www.gamesindustry.biz/roblox-makes-three-of-its-ai-safety-tools-open-source
#China is advancing a comprehensive approach to #global #AI #governance, balancing #OpenSource #innovation with safety measures and #international #cooperation. At the 2026 #World AI #Conference (#WAIC) and High-Level #Meeting on Global AI Governance, President Xi Jinping called for a just and equitable #system for global AI governance, emphasizing that AI should be people-centered and benefit all of #humanity. https://cnbusinessforum.com/china-advances-global-ai-governance-with-open-source-models-and-new-international-body/
Valve just rolled out the red carpet for the latest mainline stable update to their compatibility layer with Proton 11.0-2 now available.
https://www.gamingonlinux.com/2026/08/proton-11-0-2-is-out-with-lots-of-gaming-fixes-for-linux-steamos-steam-deck-steam-machine/
#Steam #OpenSource #Valve #Proton #NewRelease #Misc #SteamDeck #SteamMachine
![[?]](https://f2.tchncs.de/accounts/avatars/109/579/691/459/417/627/original/8faca382e10bab0b.jpg)
"Get a first look at the #QuadRF, a modular 4x4 MIMO beamforming tile built with an open antenna architecture that moves software-defined radio into the spatial domain. Functioning as a real-time RF camera, the QuadRF allows you to visualize the #wireless world around you."
#OpenSource #wifi #SDR #gnuradio
https://youtu.be/38GKlCrGTQM
![[?]](https://s3.eu-central-2.wasabisys.com/mastodonworld/accounts/avatars/113/409/640/928/552/977/original/a92596332da13e55.png)
Proton has updated to version proton-11.0-2. Check it out here: https://github.com/ValveSoftware/Proton/releases/tag/proton-11.0-2
#foss #OpenSource #linuxGaming
![[?]](https://mapstodata.s3.eu-west-par.io.cloud.ovh.net/accounts/avatars/109/300/115/402/497/091/original/1d41d91bcd8c7482.jpg)
I like that TuxedoOS has decided to shift towards supporting a community based distro, such as Debian.
So with TuxedoOS migrating to Debian, I wonder if Linux Mint will start to prioritize development of LMDE over standard Ubuntu based LM.
![[?]](https://media.mstdn.social/accounts/avatars/109/357/593/870/696/783/original/45c931c28dbdf6d5.jpg)
Claude insisted I couldn't optimize libJudy for php-judy 2.6.0. 🤔
By pushing back on its broken profiler math and applying strict research discipline, I uncovered a silent GCC -O3 compiler bug, fixed a circular Zend GC crash, and got 11–32% lookup speedups anyway 🤷♂️
https://nicolas.brousse.info/blog/hardening-libjudy-with-ai-agents/
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/109/282/246/626/328/039/original/32845179f9364d75.jpg)
🧪 Beta and open source. HydraSDR owners: please break it and tell us how 🙏
📦 github.com/PentHertz/gr-hydrasdr
🚀 rfswift.io
Im Februar 2016 kaperten Hacker die Website von #Linux Mint und tauschten die offiziellen ISO-Downloads gegen eine präparierte Version mit einer eingebauten Hintertür aus, wodurch Nutzer unbemerkt Schadsoftware auf ihren Systemen installierten.
das war mein persönlicher moment, wo ich jegliches vertrauen in die #FOSS community verloren habe.
wann habt ihr euren glauben an #FOSS und #OPENSOURCE verloren?
![[?]](https://f2.tchncs.de/accounts/avatars/109/250/308/897/663/702/original/93d28ed1a7a754a2.jpeg)
#Linux 
7.2 bringt umfangreiche Treiberupdates | heise online https://www.heise.de/news/Linux-7-2-bringt-umfangreiche-Treiberupdates-11419043.html #Linux72 #OpenSource
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/109/370/902/922/101/432/original/0ac16db50972d76c.png)
Authenticode signature manipulation toolkit for Red Team operations and security research.
![[?]](https://cdn.masto.host/indiewebsocial/accounts/avatars/112/536/017/792/352/435/original/a87e9128d6303e02.png)
China joins Europe in scrapping Windows for Linux
｢ In addition, even before the recent trend of governments outside America moving away from Windows, Beijing has started a long-running push to replace foreign technology in sensitive systems with domestic, open-source alternatives ｣
https://www.zdnet.com/article/china-drops-windows-for-linux/
![[?]](https://cdn.fosstodon.org/accounts/avatars/115/532/410/310/000/993/original/08065bd1d1c34cad.png)
I need to go out and stare at moss in the forest before getting back to your message
#ArcaneChat #privacy #security #encryption #anonymity #family #digitalindependence #Europe #European #android #openSource
![[?]](https://media.mementomori.social/accounts/avatars/111/959/388/428/844/050/original/b6918bce8f70a8d2.png)
Remember also #jolla which run on #sailfishos. At least true european alternative, and pretty close #opensource too.
Discover how Claude Code helped a developer create a native macOS driver for an unsupported HP Laser printer by fixing an open-source SpliX bug.
Bazzite Linux 44 has been out for desktops for a while, but the "Deck" image for handhelds and console-like TV PCs is finally here with a huge overhaul.
https://www.gamingonlinux.com/2026/08/bazzite-linux-44-for-handhelds-finally-launches-with-some-massive-upgrades/
#OpenSource #NewRelease #DistroNews #Fedora #Misc #Bazzite #HandheldGaming #LegionGo
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/114/598/527/317/856/539/original/9c5f2d231c6386da.png)
Update:
Our velociraptor plugin `Windows.Memory.Mem2Disk` can detect RAM injections and fileless malware.
We tested it against (among others) the C2 frameworks Sliver, Havoc and Mythic. All three were detected.
It was recently featured in a blog post by Mike Cohen:
https://docs.velociraptor.app/blog/2025/2025-11-15-memory-analysis-pt1
Stay tuned for memory analysis with velo part 2!
#C2 #detection #memoryforensics #velociraptor #DFIR #cybersecurity #infosec #pwr2 #opensource
![[?]](https://files.mastodon.social/accounts/avatars/108/191/613/221/181/390/original/c05b45adf39b250f.jpg)
RE: https://chaos.social/@cccpresser/117132501236214891
Everyone get involved in this if you are at #OSHOP it will be great fun. #FreeCAD #KiCad ... love that the base of my #opensource #pcb clamp (as seen on @oshwassociation show and tell) is the example FreeCAD challenge item!
![[?]](https://social.heise.de/__virtual/accounts/avatars/112/881/119/160/228/266/original/baf7ad72ec92f728.jpg)
Pixel 11 Pro: Developer enhances HiLight function
The rear LED light of the Pixel 11 Pro and Pro XL has almost no functions out of the box. The HiLight Studio app by developer Dhananjay Bhosale changes that.
#Android #Apps #Google #GooglePixel #LED #Mobiles #OpenSource #Smartphone #news
![[?]](https://cdn.fosstodon.org/accounts/avatars/000/153/988/original/664c3d3fb4e22b9e.png)
The Steam Deck's touchscreen LCD works with a Raspberry Pi thanks to a HAT from Scandent. They have open sourced their design and code.😎 🚀 
I received my #PineNote from @pine64 yesterday evening.
I've not had much time to play with it yet. But I'm looking forward to do that.
First general impression:
- Nice specs, nice display
- Feels like a real computer and not a straitjacket. Great for hacking, but is it great for the average non-tech consumer? However, this is currently more a hacking device, so it makes sense
- It's an e-ink display, so it can feel a bit more laggy than it most likely is - due to the screen update latencies
- It runs a Debian Trixie build with GNOME. GNOME does not feel like that right choice for such a device. Floating windows and soft-keyboard getting in the way all the time
- The software keyboard is designed to look nice - functionally, it's quite annoying.
- The pen which follows feels extremely plastic and really poor quality. But it does the job.
- I would love the pen to be magnetic so it could be soft-attached to the edges of the tablet.
- It ships with Syncthing pre-installed!
But it seems to require some work to get apt update and apt upgrade working smooth. The images installed seemed to be quite out-to-date and several PGP keys used to verify packages was out of date. And there were some issues with the /usr symlink move. So the base install is very premature. Some hacking time will be needed to get it into a more proper shape.
The documentation both on paper and the pre-installed help docs is very scarce on details. And quite some details were even not correct.
Overall impression currently is that the PineNote is not a device I would recommend to non-technical people at all. But it feels like a very capable device for those who likes hacking and don't mind some "rough edges".
I do see this as a device which I should be able to tweak into a pretty functional device with some time. And I'm looking forward to that.
Pixel 11 Pro: Entwickler bohrt HiLight-Funktion auf
Das rückseitige LED-Licht des Pixel 11 Pro und Pro XL hat ab Werk nahezu keine Funktionen. Das ändert die App HiLight Studio des Entwicklers Dhananjay Bhosale.
#Android #Apps #Google #GooglePixel #LED #Mobiles #OpenSource #Smartphone #news
![[?]](https://s3.eu-central-2.wasabisys.com/mastodonworld/accounts/avatars/109/405/068/978/571/024/original/9707b30514065232.jpg)
Here's how to automatically post to Mastodon: https://cromwell-intl.com/open-source/python-social-media-automation/?s=mb #OpenSource #Python #TwitterMigration
ICS[AP] updated CISA ICS Advisories Master File for 8/20/2026 & the following year's CSVs:
CISA_ICS_ADV_2026_8_20_26.csv
Available @ ICS[AP] GitHub:
https://github.com/icsadvprj/ICS-Advisory-Project/tree/main
Digital Identity: How to Prove Who You Are Without Revealing Everything
Imagine you want to buy a bottle of wine from an online store and you have to prove that you're at least 18 years old.
https://vsx.global/digital-identity-how-to-prove-who-you-are-without-revealing-everything/
![[?]](https://cdn.mstdn.ca/accounts/avatars/116/578/218/445/777/774/original/d1d8d6fe018e0a24.jpeg)
Soomi — The sovereign #Canadian office suite.
https://soomi.ca/en/
#DataSovereignty #canada #opensource #privacy #boycottusa #boycottmicrosoft #boycotttechbros
@aurynn I too **really** dislike Electron apps 🙄 , but when we look at all the native #OpenSource widget sets - it's so brutally difficult to like any of them as the one awesome native widget set to adopt and love. Like no clear winner ever, ever emerges, as the decades roll by.
QT? GTK? wx? "Dear ImGui"? Tk? Flutter? They all have their reasons to dislike.
My favorite - and you'll probably think I'm nuts - is #Python's "textual" TUI widget set. Yes, TUI, not GUI.
https://textual.textualize.io/
![[?]](https://ftp.ieji.de/accounts/avatars/112/842/991/176/451/444/original/f4bd00f6b0d3d962.png)
Parch Linux | پارچ لینوکسMirrorMan v0.5.2 is here, and it’s the final release in the 0.5.x family!
This release is a major stabilization and feature update for Parch Linux MirrorManager. The GUI now has mirror search and filtering, reliability sorting, bulk enable/disable controls, configurable auto-refresh, non-blocking notifications, keyboard shortcuts, and persistent window geometry. Custom repositories can also be removed directly from the GUI.
The CLI has grown significantly too, with mirrorlist backups, cleanup with dry-run support, mirrorlist sync and diff, individual mirror speed testing, export functionality, and shell completions for Bash, Zsh, and Fish.
Under the hood, MirrorMan now has architecture-aware mirror testing, automatic backup rotation, SHA256 verification for BlackArch's `strap.sh`, safer tempfile handling, and recovery from poisoned mutexes to prevent GUI freezes after thread panics.
Packaging also received some love with AppStream metadata, improved desktop keywords, and a Polkit action for cache cleanup.
with this release, MirrorMan 0.5.x reaches its final milestone.
The next chapter starts with 0.6.x.
#ParchLinux #MirrorMan #Rust #GTK #Libadwaita #Pacman #OpenSource
Alt...
Parch Linux MirrorManager
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/115/250/028/179/909/308/original/6ba391f0e4c9dbf7.png)
Introducing secdb-cli: ZEN SecDB from your terminal
Query CVEs and audit your dependencies (in PURL format) without leaving the terminal (CI-friendly gating included).
v0.1.0 is out now with prebuilt binaries for Linux, macOS, and Windows (AMD64 + ARM64).
https://secdb.nttzen.cloud/blog/2026/08/17/secdb-cli
https://github.com/giterlizzi/secdb-cli
#opensource #cli #audit #cve #purl #ci #vulnerability #dependency
#nttdata #zen #secdb
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/114/372/473/711/554/045/original/336266689a6c5f66.jpg)
"NASA's ground control software has a worrying security flaw which could let hackers contact spacecraft"
"Security researchers uncover flaw in the open source software used by NASA ground control to communicate with instruments and spacecraft."
![[?]](https://media.bsd.cafe/bsdmmedia01/accounts/avatars/115/827/315/554/506/196/original/b9aa2f717a2ffc74.jpg)
@stefano I'm so grateful to you for embodying the open-source philosophy and communicating openly. That's so incredibly important for all *BSD users and open-source enthusiasts. No one can keep track of the source code for every program or library all the time. I admire you so much for the value you bring to the community. That's the spirit!
![[?]](https://openbiblio.social/system/accounts/avatars/108/673/259/259/715/725/original/ef68bceeeb7e37c8.jpeg)
RE: https://openbiblio.social/@datenteiler/116476021319060949
Die öffentlichen Arbeitsplätze in der @tibhannover sind jetzt alle auf #Linux (ok außer die beiden Readerprinter). Noch nicht alles bugfrei, aber dafür ist es schnell und stabil. Das Nutzer:innen Feedback vor Ort, mit denen ich gesprochen habe, war super positiv, weil es so schön aufgeräumt aussieht.
Am Mittwoch stelle ich den "Ubuntu Nutzer:innen Arbeitsplatz (UNA)" kurz auf der GBV-Verbundkonferenz vor. Bin gespannt. 😊
![[?]](https://files.mastodon.social/accounts/avatars/111/685/046/419/146/552/original/3b74d5830c326914.png)
Das Vortragsprogramm zum #Linuxday #Dornbirn am 26. September 2026 steht fest:
KI-Aufsicht: Niedersachsen warnt vor Entmachtung der Datenschutzbehörden
Der niedersächsische Datenschutzbeauftragte beklagt die beschlossene Zentralisierung der KI-Kontrolle bei der Bundesnetzagentur und warnt vor US-Abhängigkeiten.
#AIAct #Bundesnetzagentur #Datenschutz #DigitaleSouveränität #IT #KünstlicheIntelligenz #Netzpolitik #OpenSource #news
An Apache InLong SQL injection flaw, CVE-2026-63038, lets attackers inject SQL via multiple parameters. Two more bugs join it. Upgrade to 2.4.0.
![[?]](https://cdn.masto.host/floss/accounts/avatars/000/058/894/original/79f2a195ed144037.jpg)
A good error dialog should tell you what failed, why it failed and what to do next.
Ubuntu's App Center offers none of those things with a shoulder shrug of an error message to say it's 'not sure what the error is'.
With critical features like driver management coming in 26.10, isn't it time for its error reporting to be a little more… clued-in?
https://www.omgubuntu.co.uk/2026/08/ubuntu-app-center-something-went-wrong?v1
![[?]](https://media.infosec.exchange/infosec.exchange/accounts/avatars/116/867/919/691/384/666/original/acf8be041ee28a1a.png)
Built a self-hosted honeytoken/canary system over the last few weeks — fake AWS keys, webhooks, DB connection strings, Telegram alert when one gets touched. Runs on free tiers (Cloudflare Worker + optional AWS CloudTrail integration, no Lambda).
CloudTrail→EventBridge ate a whole evening I didn't plan for — event history isn't the same as trail delivery, and EventBridge filters read-only calls by default. Details here: https://medium.com/@dblanko.security
Code: https://github.com/dblanko/honeytoken-ecosystem
#infosec #honeypot #opensource #DevSecOps
![[?]](https://mastodon-belgium.be/system/accounts/avatars/114/068/924/735/978/020/original/3f8dd575b44e3130.png)
Tuba 0.11.0 released for GNOME, now also updated this on my Haiku install (still not available in the depot).
Read all about it at:
https://codeberg.org/GeopJr/Tuba/releases/tag/v0.11.0
#HaikuOS #haikuports #gnome #gtk #vala #tuba #opensource #Mastodon #Fediverse
Alt...
Latest release for Tuba 0.11.0 (GNOME/GTK4) running on Haiku 64bit nightly.
![[?]](https://files.mastodon.social/accounts/avatars/001/007/156/original/fe0944c346ffe78a.png)
China wirft Windows aus den Behörden, Monate früher als geplant (Bloomberg/Handelsblatt). Microsoft sagt: alles sicher, die Updates laufen weiter. Peking sagt: trotzdem, danke, tschüss. Hierzulande würde man zunächst eine Arbeitsgruppe gründen. Zur Prüfung der Machbarkeit einer Prüfung. Ergebnis: nächste Legislaturperiode. Siehe Bund, Länder oder Städte wie Darmstadt
#DigitaleSouveränität #Microsoft #OpenSource https://www.handelsblatt.com/technik/it-internet/microsoft-china-streicht-windows-bei-behoerden-schneller-als-geplant/100247904.html?utm_medium=sm
Blick ins Heft c’t 18/2026: KI bringt IT-Security ins Schleudern
KI stellt die IT-Security vor enorme Herausforderungen. Dazu: Balkonkraftwerke mit Speicher im Test und die Highlights der IFA 2026.
#Anthropic #Balkonkraftwerk #Cybersecurity #DSGVO #IFA #IT #KünstlicheIntelligenz #Sprachverarbeitung #OpenSource #OpenAI #Security #Wirtschaft #Wissenschaft #news
![[?]](https://cdn.masto.host/floss/accounts/avatars/109/129/003/918/936/187/original/cf6e3c194b06199f.png)
A veritable smörgåsbord of new features lands in your favourite apps with KDE Gear ⚙️ 26.08!
Dolphin improves its integration with KDE Connect, Konsole revamps its drag'n'drop features, Kdenlive's titler gets more functionalities, and Okular's annotations more power.
And much more where that came from:
Proton's AI Paper Trail and the Question of Who to Trust with Your Data
Is it even possible to use artificial intelligence while also protecting privacy? At first glance, this seems like a contradiction. AI assistants are…
![[?]](https://files.mastodon.social/accounts/avatars/117/126/389/909/330/241/original/710eede03ed94ff2.png)
Keep the Why 0.8.0 is out.
context/ entries can now say Type: undefined — <reason> when none of decision/workaround/incident/constraint fits. Blank used to mean either "not reviewed yet" or "reviewed, nothing fits" — now those are distinguishable and grep-able, giving real signal for whether the vocabulary needs a fifth value later.
https://github.com/oliver-zehentleitner/keep-the-why/releases/tag/v0.8.0
![[?]](https://files.mastodon.social/accounts/avatars/000/592/868/original/634423aff2391b93.jpg)
Wir waren letztes Wochenende auf der FrOSCon. Alles in allem doch ein gutes Wochenende.
FrOSCon 2026 - Rückblick
https://andreas.scherbaum.la/post/2026-08-19_froscon-2026-rueckblick/
#FrOSCon #StAugustin #OpenSource #Conference #Konferenz #Linux #FrOSCon21