in reply to Al Sweigart

Going to disagree on this one. Having a project that's both useful and popular can achieve a greater good that outweighs the possibility of it being exploited by megacorps. I hope that my own main open-source project (accesskit.dev/) falls into that category; the more popular it is, the more applications are accessible to more people. _Maybe_ I could make some money selling copyleft exceptions, but I'd rather have more accessible apps, as long as I can make a living.
in reply to Ahmad Albahar

Yeah, they appear with what you wrote when you quoted, but not like on twitter, instead, it's a link to the post, so essentially, to a client that doesn't support quoting, which is a lot of them, even Enafore, although it says that we can quote, well, that's just not properly implemented, and whatever you can read over there to you, is because your app lets you, but I don't think the web will.

My @thunderbird regularly marks some S/MIME-signed mails with:

"Digital Signature is not valid!"

#TIL the reason: these mails are signed with sha1-algorithm which is considered insecure since TB 115.

There is a config to accept such signatures, see blog.thunderbird.net/2023/10/t…

However, I think the error message:

"There are unknown problems with this digital signature." is suboptimal.

Seems like an easy issue to fix, but on the other hand I was not even able to finde the appropriate repo. 😬

in reply to Tamas G

This is bullshit from someone who doesn't treat LLMs as tools. If you want them to interpret your asks and respond in a certain way, put that context in the system prompt and stop wasting tokens.

Saying "You act as though I always say please" is just as effective as actually saying it. Saying you want bullet points, or headed paragraphs, or responses without visual references, or whatever is even better. Not valuing politeness from a computer is even better than all of that.

This entry was edited (7 months ago)
in reply to James Scholes

I think the "valuing politeness" comes more from getting a greater effective output through it than otherwise, which isn't changing the core principle of LLMs: Since human data showed them that passages with polite dialog exert greater effort from the participant, it's likely that bias has now largely creeped into the average dataset of the models themselves. It's valid though to consider how this changes when applied to the prompt directly, as it least avoids repetition and can keep the positive sentiment the same rather than varying it across them.
This entry was edited (7 months ago)

I noticed a bug in our binutils port that had been generating semi-randomly broken branch relaxation trampolines for decades.

Why did it take so long to notice this code generation bug? The branch has to reach farther than +-32MB for the branch relaxation trampoline generation to kick in. And even then not all branches were affected (the type of relocation affected it, it had to be in a link library in a specific kind of segment and not in the beginning of it). Finally, for it to actually come into play, the branch had to actually be taken, too.

What did it do wrong then? It added the offset to the target function in the link library .text segment twice. So instead of jumping to the intended function it jumped somewhere random after that. Funnily, the jump might end up hitting some code that did not crash, but did something unintended. For you all non-developer peeps: That is very, very bad.

Fun features of this bug: Since whether the trampoline was generated or not depended on the order of object code and from where the affected call was being made, the bug would pop in and out of existence even on the smallest changes to the code or link libraries. If you know a thing about debugging, this is not very conducive to locating bugs.

Needless to say this one took a lot of head scratching to finally figure out. The fix was a change on a single line.

This bug was old enough to get into bars and drink alcohol legally.

EDIT: Oops, this was supposed to be a reply to this post: infosec.exchange/@harrysintone… - so added some context.

#bugstories #bugs #development #coding

This entry was edited (7 months ago)

With Andor s2 about to drop, it's good to keep in mind that in spite of popular assumptions, the Galactic Empire isn't based on the Nazis. George Lucas was very explicit that it's in fact supposed to represent the United States.

cbr.com/george-lucas-vietnam-w…

reshared this

LibreOffice had two major updates in 2024 with many improvements and fixes. Learn more from The Document Foundation's Annual Report: blog.documentfoundation.org/bl… #foss #OpenSource

Did you catch the recent ACB Vispero Presentation - Manage Windows 11 notifications with JAWS? In case you missed it, the archive is now available: acb-community.pinecast.co/epis…

#FreedomScientificTraining #JAWS

A cracked phone screen doesn't mean you need a new phone! Instead, choose #repair or refurbishment—it's better for your wallet, and the planet.

Repairing will refresh your device, but when you do replace it, remember to trade it in to somewhere like Back Market. Instead of ending up in a landfill, it'll find a new home in someone else's pocket (and put some money in yours!). This enables phones to stay in use for 10 years, saving cost and energy.

Learn more on our blog: ifixit.com/News/109306/fix-it-…

END OF 10

"Support for Windows 10 ends on October 14, 2025.

Microsoft wants you to buy a new computer.

But what if you could make your current one fast and secure again?

If you bought your computer after 2010, there's most likely no reason to throw it out. By just installing an up-to-date GNU/Linux operating system you can keep using it for years to come.

Installing an OS may sound difficult, but you don't have to go it alone! With any luck, there are people in your area ready to help!"

endof10.org/

#Microsoft #Windows #Windows10 #Windows11 #Linux #FOSS

This entry was edited (7 months ago)

A verified ID is required to access AI models in OpenAI's API to prevent IP theft. 😂 They went from stealing everything from everyone to now claiming it's their original work, despite hundreds of pending copyright cases against them in the courts. Why is this even allowed? That's hypocrisy at best.

#ai

#AI
This entry was edited (8 months ago)

【𝕭𝖆𝖘𝖘𝖕𝖎𝖘𝖙𝖔𝖑 🎬】
AxWax remixes at the Not Bonk What I Call Wave: Remixes Vol 2 launch party

░▒▓█▓▒░
░▒▓█▓▒░ ░▒▓█▓▒░

My two remixes on Not Bonk What I Call Wave: Remixes Vol 2, recorded at the release party Owncast stream. The tracks, Secret Deal - All Around (AxWax Remix) and Traiken - The Snag in Your Makeshift Cape (AxWax Remix) can be streamed / d...

░▒▓█▓▒░ ░▒▓█▓▒░
░▒▓█▓▒░

📺 v.basspistol.org/w/njehmVvLcBj…
⚠️ Your comments to this post are appreciated, but consider following this creator or interacting with the video directly by pasting the link above in your instance's search-box! 💌
🈁 #peerTube #Basspistol

🚨 EU citizens - Meta will start using your data for AI 🚨

Say NO and object!

Here's how 👉 tuta.com/blog/how-to-turn-off-…

#MetaAI #EU #AI

#AI #EU #MetaAI

I meant to try this out for a few days now, and I finally did it. I was able to view a Telemundo News report in Spanish and have Meta glasses translate it to me in real time using the Live Translate feature. There is about a 5-10-second delay, but over all it's quite impressive. I even tried watching trailers for some Spanish-speaking movies... Quite something!
This entry was edited (7 months ago)

reshared this

Love some quotes in this article, e.g. that democrats and liberals are more likely to buy Teslas than republicans. Guess Elon forgot whose pockets he should be thankful to... As Tesla profits plunge 71%, Elon Musk says he’ll spend less time on DOGE - NPR apple.news/AXomrUMyFRKGbQrC1aN…

"The old and new right-wing parties are amalgamating into a neoliberal and hyper-conservative ‘pleiad’ (social notables) that comprises neo-Christian fundamentalists, including Fatimists (a Catholic religious faction). The new desire to celebrate November 25 emerges from here. The latest extreme on the right is embodied in the anti-communism of IL (the Liberal Initiative) and the neo-fascism of Chega. The media normalizes it all as milquetoast ‘liberals’ rather than what they are – figures of the ‘radical / extreme right’. These groupings want to celebrate November 25 for what it was: a coup d’état against democracy at work, against the duality of popular power. In short, they want to celebrate the counter-revolution, the beginning of the end of the Revolution.

November 25 heralded Portugal’s reconstruction from a ‘national capitalism’ dependent on forced labour in its African colonies until 1974 to a subservient capitalism that is a de facto protectorate of the leading European power of Germany, France, and Britain. Portugal now helplessly depends on investments and capital goods from North America, Spain, China, and other countries.

The ostensible ‘democratic counter-revolution’ of 25 November 1975 is a central political event for understanding what happened in Portugal, and it was far less democratic and far more reactionary in its roots. As we rescue the authentic memory of the Carnation Revolution and lament its demise, no wonder Father Martins Júnior said in a piercing manner that what remains is “50 neo-fascists in the hemicycle 50 years after the April Revolution.”

The impossible became possible on that day 50 years ago. The dreams unleashed on April 25, 1974, neither have been vanquished nor fulfilled. The slogan could not be more urgent: April 25 forever, fascism never again!"

socialistproject.ca/2025/04/th…

#Portugal #CarnationRevolution #25deAbril #25deNovembro #Fascismo #Fascism #EstadoNovo

If you heard about that hacking of the voices of traffic light crosswalks in the US recently, the root cause is the devices all had the password '1234' and an app to reprogram the devices was on the Apple app store.
theregister.com/2025/04/19/us_…

reshared this