This water makes it's own sauce.
Dear People of Brazil: We, Germans, are very sorry for these arrogant, rascist and stupid things our (not loved) chancellor told you about #Brasil & your city #Belém - most of us didn't vote for him. He is a constant nuisance and says mean things even about German cities and is absolutely wrong about migration, climate crisis and its solutions. I often get woozy shaking my head just listening to news.
So, pls forgive us - our country Germany. We all do hope our head of gov will retire VERY soon.
So, pls forgive us - our country Germany. We all do hope our head of gov will retire VERY soon.
This entry was edited (1 day ago)
I really need to move instances; yiff.life softblocking (or whatever the hell it's called, I can only see stuff from there if it's replies or I followed the ) mastodon.social is a real fucking pain in the ass if you actually want to use the service.
Any suggestions?
One more reason why your social media platform should not have an attached, easily accessible API.
It's far too easy for LLMs (through coding agents) to stalk others for you.
"look through the last 1000 posts of this person and figure out if we can hire them, we have offices in <insert list of cities here>" is absolutely doable right now. That's an euphemism for "I want to know where my ex lives right now" that is palatable to LLMs.
2/2
Not that this is impossible on centralized socials, but you'd have to at least build some kind of LLM harness to do it, and that's almost as much work as just doing the stalking yourself. On Masto, you can go "here's curl, here's a Python interpreter, the username is so and so, go and do your thing."
Not that this is impossible on centralized socials, but you'd have to at least build some kind of LLM harness to do it, and that's almost as much work as just doing the stalking yourself. On Masto, you can go "here's curl, here's a Python interpreter, the username is so and so, go and do your thing."
On this day in 1998 the first module of the International Space Station, Zarya, was laucnhed. That's 26 years ago.
Make A New Seam
I've been avoiding new music lately. It's a painful thing, because I actually don't want to. But I also don't know what else to do. The truth is that everything could be AI, and I'm so skeptical of artists I don't know. Every new release, every song a service suggests, I'm asking. And the paranoia sucks.
Please don't tell me to just move past it and accept it, this is something I feel strongly about and I'm allowed my own convictions. I just hate the concept with a passion and cannot figure out how to simultaneously avoid it and be able to try/have access to new things.
Ironically, this is where TikTok comes in super handy. Most of the songs I hear on there are sung live, as some kind of demonstration or promo, and if the artist on the original sounds the same, I'm relatively convinced.
Your creativity is definitely not threatened by AI. If people choose to listen to it and prefer it over music that people put their soul into, that's their problem. If you enjoy what you do and you do it because you enjoy doing it, nothing else matters.
I have been rejected because of AI from a huge company that values quantity over quality these days. Was it disappointing? Sure, when you can't pay your bills because you don't have a job it's more than disappointing. But you are empowering AI more if you give up doing what you like because of it and I can't do that.
Today I'm reviewing soundness. I'm so used to thinking about proofs as establishing truth (compelling belief) it was hard for me to distinguish the |- (derivability) and |= (semantic entailment) worlds.
I think I got it eventually, but it's not a very natural way for me to think, at least yet.
A real Hackerone #curl report title!:
"Out-of-bounds read in *** potential crash. This is sharp, <reporter name>. We've got a real memory safety bug"
The AI is helpfully cheering the guy onwards to slopping. Of course, it is a false positive.
heartbreaking: your mutual that posts tons of great memes never uses alt text so you can't boost them
Wow! The iPhone 17 Pro Max can download at 1.8 gigabits per second, and upload at 1.3 gigabits per second. Pretty darn good for WiFi. Cellular performance is craptastic, no more than 400 megabits down by 80 megabits up, but I think I can pretty confidently say that is a result of shitty cell network performance, rather than device performance.
@jonathan859 Bass is increased pretty significantly over the Pro Original. I never owned the Pro 2, so can't comment. Also never owned Pixel Buds, so also can't comment.
Would you say it sounds more spatial? I know AirPods have the 3d Sound thing, but what I mean by that is rather, does the music sound more next to your ears and more bright/open/full? I know this description is probably horrible, but I don't know how to describe it in proper audio language, especially in English. The experience I've made with the AirPods Pro 2 was that music like techno sounded a bit, well idk, I can't say it better than flat. Like it has been tried to squeeze it into the little thing left and right to my ears. While I feel like with my Pixel Buds it sounds, more full. I know it's hard to say without having a comparison between the earbuds actually being talked about, so sorry if that's out the scope. Just interested in the improvements with the new model.
This entry was edited (5 hours ago)
@jonathan859 I have disabled all of the spatial and head tracking stuff. It sounds reasonable to me. In fact, a problem I was having with my Airpods Pro 1 was that Spatial and Head tracking shit would actually not really turn off when it said it was, so everything sounded like it was moving with my head, and also had a bit of a room echo effect on everything. I was most certainly not a fan. This is fortunately not the case with these new Airpods Pro 3, and the soundfield doesn't track with my head, and the audio sounds nearly as close to as intended as possible.
@jonathan859 The sound stage on the Pro 3 seems reasonable. I certainly own pairs of cans that cost three times as much that do a way better job, but for a pair of buds that you can take with you everywhere, and not having to have three pound cans balanced on your head is amazing. I cannot comment on any comparisons to Pixel Buds or Pro 2.
After two weeks of waiting for packages and help, we finally replaced our old an broken work desks with some new models yesterday. Finally we are also able to work while standing up. Also, having an actual desk plate and not this swedish crap of two thinny wooden planks with alot of nothing in between feels really good.
Strange Little Dance
Network Storage #NAS #HomeNetworking
Which of the following would you choose?
- Synology (40%, 2 votes)
- QNAP (20%, 1 vote)
- Ubiquiti - UNAS Pro 4 (0%, 0 votes)
- Ubiquiti - UNAS Pro 8 (40%, 2 votes)
I currently have an older Synology DS1815+, will probably replace with a UNAS-pro 8 when that needs to happen. I have a friend who works for an architectural firm, who also manages a multi-hundred site Ubiquiti controller and a bunch of junk on the side, and he seems to love them.
Sensitive content
reshared this
Sensitive content
@NVAccess @jcsteh I appreciate your insights here. I am planning to open an issue on GitHub, but I first want to make sure it actually needs a separate issue.
Currently, when NVDA encounters bold or italic text in HTML, it displays that word or phrase on its own line in the virtual buffer. Just like with links, NVDA then places the following non-styled text directly after it on the same line. This means that when navigating with Up and Down arrows, you may encounter several new lines in the middle of a paragraph, each starting with one or more bold or italic words, followed by normal text. Besides making navigation more difficult, this behavior is not very useful. For example, in online dictionaries where one or more words in an example sentence are bold or italic, the virtual buffer does not show how many words at the start of the line actually have that formatting.
To see this issue, which is very easy to reproduce, go to the Merriam-Webster dictionary website and check the definition of “workaholic.” Here is the direct link:
merriam-webster.com/dictionary…
In the Examples section, the first example displays “My brother is a real” on one line, and on the next line it shows “workaholic who almost never takes time off.” The problem is even worse in the Recent Examples section. There, in addition to an italicized word inside the example, the name of the publication is also bold or italic, and NVDA breaks the paragraph even further and places each styled element on its own line.
At the very least, there should be a setting that toggles this behavior, so users can choose between the current mode and an alternative mode where bold, italic, and underline formatting does not break the virtual buffer into separate lines.
Also, when pressing Insert+F on text that is bold or italic, NVDA reports formatting details only at the end. For example:
“Open Sans 14.5 pt very dark pale aqua-blue on white bold italic align left.”
It would be better if NVDA announced bold, italic, or underline at the beginning instead of the end, similar to how JAWS does it. For example:
“bold italic align left Open Sans 14.5 pt very dark pale aqua-blue on white.”
I want to open an issue requesting NVDA to stop separating bold and italic text from surrounding words, or at least provide a toggle for it, and to improve how formatting information is announced in general. There should be more efficient ways to convey this information without altering the structure of the virtual buffer.
Currently, when NVDA encounters bold or italic text in HTML, it displays that word or phrase on its own line in the virtual buffer. Just like with links, NVDA then places the following non-styled text directly after it on the same line. This means that when navigating with Up and Down arrows, you may encounter several new lines in the middle of a paragraph, each starting with one or more bold or italic words, followed by normal text. Besides making navigation more difficult, this behavior is not very useful. For example, in online dictionaries where one or more words in an example sentence are bold or italic, the virtual buffer does not show how many words at the start of the line actually have that formatting.
To see this issue, which is very easy to reproduce, go to the Merriam-Webster dictionary website and check the definition of “workaholic.” Here is the direct link:
merriam-webster.com/dictionary…
In the Examples section, the first example displays “My brother is a real” on one line, and on the next line it shows “workaholic who almost never takes time off.” The problem is even worse in the Recent Examples section. There, in addition to an italicized word inside the example, the name of the publication is also bold or italic, and NVDA breaks the paragraph even further and places each styled element on its own line.
At the very least, there should be a setting that toggles this behavior, so users can choose between the current mode and an alternative mode where bold, italic, and underline formatting does not break the virtual buffer into separate lines.
Also, when pressing Insert+F on text that is bold or italic, NVDA reports formatting details only at the end. For example:
“Open Sans 14.5 pt very dark pale aqua-blue on white bold italic align left.”
It would be better if NVDA announced bold, italic, or underline at the beginning instead of the end, similar to how JAWS does it. For example:
“bold italic align left Open Sans 14.5 pt very dark pale aqua-blue on white.”
I want to open an issue requesting NVDA to stop separating bold and italic text from surrounding words, or at least provide a toggle for it, and to improve how formatting information is announced in general. There should be more efficient ways to convey this information without altering the structure of the virtual buffer.
You may get some pushback though. By definition, turning screen layout off means you're asking NVDA not to respect visual layout and to instead break lines. Where lines are broken, then, is debatable. You personally just want this for links and I tend to agree that formatting and links are very different, but it could equally be argued that the same rule should be applied to other inline elements because they are still different from the surrounding text. VoiceOver breaks on every. single. element. for example. This is why I personally dislike non-screen layout in general, even for links, but I respect that there are different preferences and styles of navigation.
#Accessibility Statement
"
This statement was last updated on [enter relevant date].
We at [enter organization / business name] are working to make our site [enter site name and address] accessible to people with disabilities.
"
I really believe you.
Buenos días desde la Administración Pública.
Realizando el desempeño de las tareas correspondientes a mi puesto de trabajo de forma diligente y cumpliendo la jornada y el horario establecidos.
Y vosotros, ¿qué tal?
Put it in DFU mode, and upload a new firmware.
Linus Torvalds Says Vibe Coding is Fine For Getting Started, 'Horrible Idea' For Maintenance developers.slashdot.org/story/…
Linus Torvalds Says Vibe Coding is Fine For Getting Started, 'Horrible Idea' For Maintenance - Slashdot
Linus Torvalds is "fairly positive" about vibe coding as a way for people to get computers to do things they otherwise could not.developers.slashdot.org
And oh boy all terms seem so foreign to me as a long time #Linux user. The same disk is called ada0 with a partition like ada0p2 in #FreeBSD, will be called something like sd0 with sd0h in #OpenBSD, ld0 and dk2 in #NetBSD . Then to experiment, all the #dkctl, #disklabel and #fdisk commands are like blue and red wires on a dynamite you have to get rid of 😱. Linux distros nowadays seem to be going to the #gdisk way which feels very much like #gpart in FreeBSD.
#UseBSD #RUNBSD #BSD #FOSS #UNIX #Linux
#UseBSD #RUNBSD #BSD #FOSS #UNIX #Linux
This entry was edited (9 hours ago)
If all the people who work at airports told you, "don't fly that airline," I hope you would listen.
When all the musicians who have the freedom to be honest tell you, "don't use Spotify," you might consider other options for listening to music.
If I might weigh in here with some personal experience as an indie artist. I’ll refer to my streaming stats for March 2025 which totalled a bit over four dollars in streaming earnings.
Amazon Music: 6 streams, $0.00
Spotify: 84 streams, $0.00
Tidal: 2 streams, $0.01
Pandora Plus: 3 streams, <$0.01
Deezer: 1 stream, <$0.01
Facebook/Instagram: 21 streams, <$0.01
YouTube: 87 streams, ~$0.21
Apple Music: 27 streams, ~$0.15
Soundcloud: 23 streams, ~$3.78
If you are paying a Soundcloud subscription, a portion of that will go directly towards the specific artists you’re streaming.
I haven’t mentioned Bandcamp as I think of it as a music shop rather than a subscription streaming service, but straight up purchasing music there is a really great way to support your chosen artists.
> This security update will give you assurance that when you receive a message from a contact, you can effortlessly assume it’s really from them.
What does this mean? You can forge messages as other users on Matrix?
the world doesn't want you to know this but if you don't sleep tomorrow never starts. follow me for more pro sleep tips
I feel this needs to be repeated 🍪
The annoyance of cookie banners
doesn't come from the regulations, but from the malicious compliance of the corporations who want to exploit your personal data.
No data-harvesting cookies = No banner.
Simple.
My websites have no cookie banners,
because they don't use any non-essential cookies and don't track visitors.
Yours shouldn't either.
omg, so much this ☝️
"""
We value your privacy so much that it only takes four clicks on drab-looking buttons and tiny checkboxes to opt out of our invasive data slurping...*or* you can click the single big green "I consent to being slurped by you and your 9k partners" and we can just move on with our day like pals. So whaddya say, pal?
"""
reshared this
second day without nicotine.. #dobréRáno #goodMorning
edit: in fact without cigarettes still wearing nicotine plaster 😎
This entry was edited (9 hours ago)
keep on dancing
President Trump has signed the Epstein Files Transparency Act into law. thehill.com/homenews/5614156-t…
What are we going to do with this law once the files are released? I mean, it's a one-time action. Does it deserve to be a law? :)
Total talk time for 11/19/25: 00:45:49.737. Bandwidth used on Dallas node: 2.20GiB.
“Oracle is already underwater on its ‘astonishing’ $300bn OpenAI deal”
Of course it is. *Nobody* is making money on AI.
(Except Nvidia.)
(For now.)
ft.com/content/064bbca0-1cb2-4…
Oracle is already underwater on its ‘astonishing’ $300bn OpenAI deal
AI’s circular economy may have a reverse Midas at the centreBryce Elder (Financial Times)
Well, so much for the that one for the kids! Off the Christmas list! 🤪
CNN: Sales of AI-enabled teddy bear suspended after it gave advice on BDSM sex and where to find knives
cnn.com/2025/11/19/tech/foloto…
This entry was edited (11 hours ago)
fake garage
Don't Let It Hit You In The #$%@
November 19:
- International Men's day
- World Toilet day
Coincidence?
New Archive: Full Circle for November 19 2025 hkc.im/TPMTfc w/ @Liz
The easiest way to install chatmail server is a small project written by a single person, meanwhile the #deltachat official server installation process is cluster of unintelligible documentation and a docker image that need multiple commands to setup properly.
How does that one dev build a better hosting experience than a company?
This entry was edited (14 hours ago)
Are there any plans to make chatmail and friends configurations more distro-agnostic or at least documented in the future, because I'm in the middle of packaging it for RHEL (and possibly Fedora) and so far it has been an interesting experience to even get the build dependencies packaged and I have not even arrived at digging the configurations out of cmdeploy yet.
@phnt Please feel free to open an issue at github.com/chatmail/relay describing what you did/intend to do.
It's true that the repository for now targets Debian12 but the deployment client (where cmdeploy is installed) should not be distro-dependent. There recently landed a new documentation side btw chatmail.at/doc/relay/
GitHub - chatmail/relay: chatmail service deployment scripts and docs
chatmail service deployment scripts and docs . Contribute to chatmail/relay development by creating an account on GitHub.GitHub
Goa Goa Goa
Sylvia
in reply to Neko May • • •If you don't mind the local timeline being all Germans talking about Nazis bad and trains delayed chaos.social is honestly nice enough, I can hook you up with an invite just lmk.
They do block the worst scum and they ended up defederating Threads after seeing their moderation policy, but they don't seem to block instances lightly.
You are expected to at least somewhat CW some stuff I believe and mods can get annoyed if you are careless but I never ran into trouble myself.
Sylvia
in reply to Sylvia • • •I'm not complaining about the local timeline btw, I'm glad the general consensus here is "fascism is bad actually, let's not do that again"
Though I wouldn't mind more cut content