🚨 noyb has filed two complaints against #Microsoft (365 Education) for violating children's data protection rights!

👉 The company shifts responsibility for #GDPR compliance to local schools...

👉 ...and places tracking cookies without consent

Read all about it here:

noyb.eu/en/microsoft-violates-…

Microsoft violates children’s privacy – but blames your local school

noyb has filed two complaints against Microsoft for unlawful data processing in 365 Education

^noyb.eu

Things the guys who stole my phone have texted me to try to get me to unlock it

Link: gothamist.com/news/things-the-…
Discussion: news.ycombinator.com/item?id=4…

Things the guys who stole my phone have texted me to try to get me to unlock it

^{Veronica de Souza (Gothamist)}

Total Recall... táto téma zreje ako vino. Teda nie - zrie oveľa rýchlejšie.
Zhruba pred mesiacom sa Microsoft bil do pŕs, ako je pre nich bezpečnosť nadovšetko. Pred dvoma týždňami ohlásili obludný spyware, ktorý ale je "super zabezpečený", veď všetko ukladá a spracúva lokálne, navyše "ukladá šifrovane" - viac v mojom ostatnom blogu: herrman.sk/home/a-tento-poznat…

No a dnes, prichystajte si prekvapený výraz v tvári, ono je to naozaj tak zlé ako sme predpokladali: wired.com/story/total-recall-w…

#recall #blog #privacy #fail

@CREW is tracking companies that pledged to make no donations to politicians who "who voted against certifying the 2020 election [&] later, the 20 new members who won in the 2022 midterms promoting the #BigLie that the 2020 election was stolen."
citizensforethics.org/reports-…

It just published its June 2024 update:
* 37 companies kept their promise
* 1,794 companies did not, donating $145.5 million to the "Sedition Caucus" since January 6, 2021.

#Boycotts #Insurrection #USPolitics

Bankrolling the Big Lie - CREW | Citizens for Responsibility and Ethics in Washington

We're scoring companies’ support for democracy based on how much they are giving to members who have done the most to undermine it.

^{CREW | Citizens for Responsibility and Ethics in Washington}

Some of us Internet Weirdos are doing a survey on Neurodivergence and Education.

If you are in school, college, university or whatnot, or have been in the last five years, could you spend a few minutes filling in this out for us?

We're writing guides for educators and learners and the more input we have, the better we'll be.

Appreciated folx!

forms.gle/zoww1VXHG4AHu9kV6

#ND #neurodiverse #education #school #college #university #neurodivergent #autism #actuallyAutistic #ptsd #survey #surveys

v.2 NdSA General Survey

Share for bigger sample size

^{Google Docs}

"Lifesaving “Golden Rice” is once again under attack from vested interests like old-fashioned environmental groups, powerful politicians and some farmers."

Sign the petition, save lives 👇

act.weplanet.org/defend-lifesa…

Defend lifesaving golden rice

^WePlanet

#AndroidAppRain at apt.izzysoft.de/fdroid today brings you 13 updated (as yesterday) and 2 added apps:

* Ollama: chat with an Ollama server
* Caffeinate: keep your phone's screen On

Enjoy your #free #Android #apps with the #IzzyOnDroid repo

IzzyOnDroid F-Droid Repository

This is a repository of apps to be used with F-Droid. Applications in this repository are official binaries built by the original application developers, taken from their resp. repositories (mostly Github, GitLab, Codeberg).

^{IzzyOnDroid App Repo}

new test release 1.45.1 for #deltachat_desktop is ready for testing.

We fixed the map and added a small screen mode making it more usable on small screens such as #linuxmobile phones (even though that is still not really a supported use case)

support.delta.chat/t/help-test…

🧪 Help testing the upcoming v1.45.X release

The time has come for a new Desktop Release! 🎉 Help us find the most annoying bugs so we can make sure this update is ready to be released on the website and in stores.

^Delta.Chat

«Telegram Combolists and 361M Email Addresses»
– by @troyhunt

Telegram was never secure but now is it very clear. You us it, why?

🤷 troyhunt.com/telegram-combolis…

—
#telegram #itsecurity #email #combolist

Telegram Combolists and 361M Email Addresses

Last week, a security researcher sent me 122GB of data scraped out of thousands of Telegram channels. It contained 1.7k files with 2B lines and 361M unique email addresses of which 151M had never been seen in HIBP before.

^{Troy Hunt}

I am not the only one to say this:

Please remember that WCAG itself is the bare minimum of #accessibility. Conforming to WCAG does not guarantee something is accessible. It does not even guarantee something is usable. All WCAG does is provide you with a starting point. Lots of WCAG failures suggest the page has not even made it to the starting line.

Which I say at the end of this post comparing free automated WCAG testing with manual testing:
adrianroselli.com/2023/01/comp…

#a11y

Comparing Manual and Free Automated WCAG Reviews

Automated accessibility testing tools cannot test 100% of WCAG. This position is not controversial. Other than overlay vendors, no automated tool maker makes that claim.

^{Adrian Roselli}

We've got a Thunderbird for Android/K-9 Progress Report for May, coming in hot! 🔥

✅ Material 3 Updates
✅ Android 14 Targeting
✅ Good News Re: F-Droid Metadata

...and more, including Community Contributions! Check it out on the Thunderbird Blog: blog.thunderbird.net/2024/06/t…

#Thunderbird #Android #Development

Thunderbird for Android / K-9 Mail: May 2024 Progress Report

Material 3 improvements, better developer documentation, and more in our May 2024 progress report for K-9 Mail / Thunderbird for Android.

^{cketti (The Thunderbird Blog)}

Ost-West -Achse in @koeln :

„Da es kein vernünftiges Argument für den Ost-West-Tunnel gibt – welche Interessen stecken eigentlich dahinter?

Wer will solche Tunnel und warum?“

@verkehrswendek führt Euch in die Tiefen des Kölner Klüngels.

Der in Köln ansässige Bundesverband der Tunnel-Lobby, die „Studiengesellschaft für unterirdische Verkehrsanlagen“, kurz STUVA, hat beste Verbindungen ins Rathaus und zur KVB.

Mehr im Blog:

koelle4future.de/blog/2024/06/…

#ObenBleiben #Mobilitaetswende

Offenes Plenum Juni

Herzliche Einladung zum offenen Plenum an alle kölschen Klimagruppen und klimainteressierte Menschen. In einem gemütlichen Rahmen wollen wir uns besser kennen lernen, gemeinsam austauschen und die Aktionen der nächsten Monate...\n

^{Verkehrswende Köln (Kölle For Future)}

I was sad to find a surprising lack of visual scripting apps for Linux. Apple's Shortcuts app is very powerful and can be incredibly useful. It's not like we don't have the technical ability to make something similar.

So I started working on one myself

#Linux #GNOME

Sensitive content Reveal/hide

"This last encounter was so much more than just a fling. Not because we’re falling in love… but because it proved to me that I have accepted myself, and that others can and will accept me too." #ICYMI last week's gorgeous guest blog by Ana ❤

girlonthenet.com/blog/casual-s…

Guest blog: Casual sex helped me learn to accept myself

Ana talks about the casual sex experience that made her realise she'd managed to shake off shame and insecurity and accept herself.

^{Girl on the net (Girl on the Net)}

Researchers have developed SaySelf, a novel training framework addressing a major issue I've observed with LLMs: their inability to express uncertainty or accurately convey confidence in their responses.

SaySelf tackles this problem through a two-stage process:

1. Fine-tuning LLMs on model-specific datasets containing self-reflective rationales and numeric confidence estimates.
2. Reinforcement learning with a reward function that encourages accurate, high-confidence predictions and penalizes overconfidence in incorrect responses.

The result is LLMs that can generate responses with human-like confidence indicators, potentially making them more trustworthy and reliable.

As someone who has interacted with AI, I've often been frustrated by their lack of uncertainty expression. They tend to present all responses with equal confidence, even when proven wrong. In contrast, humans often preface their answers with phrases like "I'm not an expert, but..." or "I could be wrong, but...". This absence of uncertainty expression in AI can lead to over-reliance on
potentially inaccurate information.

SaySelf's approach shows promise in addressing this issue, significantly reducing calibration error while maintaining task performance. The generated rationales provide valuable insights into the model's uncertainty.

Could this be a step towards more transparent and human-like AI assistants?

github.com/xu1868/SaySelf
arxiv.org/pdf/2405.20974

We're getting ready for the B2Run here in Hanover. 🤩

Give us a virtual clap to get us through the finish line! 👏

#gemeinsamaktiv #TutaTeam #Privacy4All #Hannover

Microsoft Copilot helpfully summarizing an obvious phishing email, without alerting that it's a phishing email!

#ai #bullshit