Skip to main content



Čakám na rožky. Čerstvé pečivo bez AI je také zastaralé.


Potsdam: AfD-Politiker Dennis Hohloch sprach vor Grundschülern über Gruppenvergewaltigungen - eine Mutter beschwerte sich. Er veröffentlichte ihren Namen nebst Foto, nun wird sie bedroht. DER SPIEGEL
archive.li/2024.07.26-170151/s…


Before and after closing the #Matrix #Element chat client tab in Firefox: 2.8 gigabytes of RAM freed. For *one tab*. And this is not a memory-inefficient browser we're talking about.

I guess feature-rich public chatrooms have a cost.



Vozítko našlo na Marsu dosud nejzáhadnější horninu. Může být důkazem dávného života
zpravy.aktualne.cz/zahranici/v…


Jak jednoduše získat data bláhových uživatelů? Nabídněte jim neužitečnou blbost, která je cool. Na oplátku dostanete přístup k obsahu všech stránek které navštíví.

Možná žádná data tohle rozšíření nesbírá a neposílá, ale zdrojový kód není k dispozici a ujištění na webu nestačí.

#soukromí #bezpečnost

@365tipu mastodonczech.cz/@365tipu/1128…

in reply to Tomáš Odehnal

Práve si popísal modus operandi Alphabetu za posledných 15 rokov minimálne. 🎯

in reply to r0man 🇺🇦

Ja mám celkovo rád tie ich vstavané filtre, nielen WB. Totiž na fotení mňa nikdy nebavil ten postprocess. Nikdy som v tom nebol dobrý, ale automatické vyvažovanie farieb dáva sterilne uniformné výblitky, tak chtiac nechtiac musel som ísť do rawu.




#dobréRáno vespolek...

Dnes vám ukážu svoje letní kombo. Minerálka, smoothie a ledová káva.

in reply to Archos

@archos @PtrSrejber Jo no, lepší je právě nejdřív to studené mléko. Poněvadž horká káva ty kostky mnohem rychleji rozpouští a pak to není ledové, ale vlažné a bez ledu.
in reply to Zdenda Sazovský

@sazy @archos
Tak jsem to právě dělal , když jsem pracoval v kavárně. Led studené mléko, espresso, protřepat v šejkru a nakonec vanilkovou zmrzlinu. Ve skle do vypadalo moc dobře...


Non-Google search engines blocked from showing recent Reddit results arstechnica.com/?p=2039164
in reply to Patrick C Miller

I think I've got an opinion... But I better not present it publicly 🤯


The Physical Keyboard Challenge: Evaluating the Experience of Using an Android Tablet with an External Keyboard - Accessible Android accessibleandroid.com/the-phys…


This is a fresh 4K remaster of one of the most brilliant movies I have ever seen.
A+ sci-fi story with breathtaking hand-drawn visuals by a French-Hungarian team from '82.

If you get a chance, watch it.

youtube.com/watch?v=D4f4tprXSi…


in reply to Tuta

Honestly, I'm one more stupid decision by Proton away from switching to Tuta...
in reply to Tuta

tbh i kinda like mail, contacts, cal all in one app. All neatly placed in one app, so convenient. I hope you make this opt in and not remove cal from tuta mail app.


1Password is ISO 27001 certified — and more blog.1password.com/1password-i…


Laura Medcalf’s Legacy eastersealstech.com/2024/07/24…


TranscribeGlass™ Brings Captions to the Forefront eastersealstech.com/2024/07/23…


Abbott Warns Small Number of FreeStyle Libre 3 Sensors May Give Inaccurate Readings cnet.com/health/medical/abbott…


I Tried Improving My Sleep Quality With the Manta Sound Sleep Mask. Here's What Happened cnet.com/health/sleep/can-the-…


Interview with Jessica Hipp, the COO of WayAround evengrounds.com/interview-with…






New scraper just dropped (well, an old scraper was renamed):

Facebook/Meta updated its robots.txt entry for opting out of GenAI data scraping. If you blocked FacebookBot before, you should block meta-externalagent now:

User-Agent: meta-externalagent
Disallow: /

Official references:

#RobotsTxt #Scraper

This entry was edited (3 months ago)

Seirdy reshared this.

in reply to Seirdy

Obligatory “the W3C/EU effort for a standard TDM Reservation Protocol would solve this by letting sites globally opt out of datamining for purposes like this, without having to play robots.txt whack-a-mole”.

People really ought to support the NoAI and NoImageAI X-Robots tags in the meantime. They probably won’t, though. Unlike the TDM Reservation Protocol, there’s no legal incentive to do so; unlike noindex, there’s no self-interested reason to do so (noindex is often used on pages that don’t belong in search results, like duplicates and non-public pages).

This entry was edited (3 months ago)

Seirdy reshared this.



🇨🇭 Switzerland now requires all government software to be open source - @ZDNet

「 This new law requires all public bodies to disclose the source code of software developed by or for them unless third-party rights or security concerns prevent it. This "public money, public code" approach aims to enhance government operations' transparency, security, and efficiency 」

zdnet.com/article/switzerland-…

#switzerland #opensource

reshared this



Let me preface this by saying that I do absolutely believe the world would be better if Windows were Free Software, but fsf.org/news/lets-not-celebrat… is a very odd post. The failure was due to a bug in Crowdstrike's code when fed invalid data, resulting in uninitialised memory being interpreted as a valid pointer value, which exploded when dereferenced. Having more access to the Windows source code wouldn't have helped here. If the argument is that *Crowdstrike* should be Free Software, sure!


Fascinating read. The skydiver who survived a 14,000-foot fall espn.com/espnw/story/_/id/4042…


What should I do about this backpack food box that was dumped in my garden?

It's a little worn, but not excessively so. And it breaks my heart to see something like this go to waste.

Also, whoever you are, stop using my garden as a trash heap, k thx.

in reply to Martin Owens

dunno about where you live, but here those backpacks are highly prized by delivery people. Apparently delivery companies sell them to them for way too much, so used ones are gold.



Extremely symbolic seeing the Refugee Olympic Team enter the Games in a boat


Hey @chikim no pressure, but I do think this deserves your attention, especially in the context of VOLlama: ollama.com/blog/tool-support
Really cool development, I must say!
This entry was edited (3 months ago)
in reply to victor tsaran

I saw that and think it might have some potential. Looks very interesting.
in reply to victor tsaran

Yes, I posted about it earlier, but if you have any idea, let me know. The thing is you have to bake the functions into the code, so it's not like people can supply function in real time.
in reply to Chi Kim

Sorry, I was looking for your earlier post, but could not find it. Is what you said here the same what you said in the original post? Just wanted to make sure I did not miss the context.
in reply to victor tsaran

Nothing really important. I just posted Latest Ollama implemented Tool support! Agent time! lol




Is there anything that republicans do not mock? A party of mockers! :) mastodon.social/@chikim/112854…


@FreakyFwoof Also this is a good time to remind folks that bunch of republicans mocked #Kamala providing a brief visual self description at the roundtable discussion with #disability rights activists. Regardless whether self description is useful for visually impaired people or not, at least she acknowledged and tried to be inclusive. msnbc.com/opinion/msnbc-opinio…

in reply to victor tsaran

Just curious, do you use Moma? Did you quote the post or just link to it? When I quote posts they show up in the feed, I’m trying to see how universal this is.


I'm browsing an appliance site that has the following ADA statement as a terrible excuse to not make their site accessible.
"Welcome to our website! As we have the ability to list over one million items on our website (our selection changes all of the time), it is not feasible for a company our size to record and playback the descriptions on every item on our website. However, if you have a disability we are here to help you." Talk about a lawsuit waiting to happen.
mannystv.com


How a screen reader reads FFConf (or ffconf) and how the investigation I had done to "fix" it: remysharp.com/2024/07/23/scree…
in reply to rem

As a screen reader user, this seems mad. You shouldn't have the right to determine how I hear something anymore than you could stop someone reading it visually and deciding how to say it.
in reply to Sean Randall

@cachondo perhaps a strong opinion, but when I'm on stage introducing the event, and when I talk about the event, I say it as "F F Conf" (and we correct people who call it "Full Frontal")

It's definitely mad that I've tried to jump through hoops when in reality a screen reader is going to sound it out differently depending on the software you use.

As for the right - I've always argued for small and personal sites, you should have the right to experiment, so long as this doesn't hurt anyone.

in reply to rem

mine was perhaps a bit reactive, too. My apologies.
I think the lengths you've gone to for inclusion are remarkable and I admire the technical aspects greatly.
As you say, though, a person's choice of screen reader, speech synthesizer, voice characteristics ... it all impacts. Blind users will spell what they need to, if they don't already use Braille. It's the difference between calling me Shaun or Shawn or (If you're very british and blind, shorn), when the spelling is sean. It's part of our lives already.
in reply to Sean Randall

@cachondo on your closing point, I do agree. That's where I nearly (or should have) ended with "am I doing it wrong" in the blog post.

I did also consider changing all instances (across 14 years of micro-sites) from (lowercase) ffconf to (mixed case) FFConf. I didn't mostly because, well, I think screen reader users don't really think (or care) that the event is actually called ffconf (where the F letters aren't pronounced).

in reply to Sean Randall

@cachondo tangentially related question: are you, the user of the screen reader, able to fix pronunciations? For instance, I'm called Remy (pronounced "rem-me"), but I've had machines call me "ree-me". I'd want, as a user, to correct local pronunciations that the developers possibly had wrong. (Just curious)
in reply to rem

Oh absolutely yes, there are dictionaries. Some screen readers just let you do it globally, others on a per text-to-speech engine or even per-voice basis.
in reply to rem

there's also a huge disparity between speech engines. Some people want a Human-sounding version of the onscreen text, others see a screen reader as more of a way of accessing data. The predictability of a more monotone voice can allow for faster speeds, for example.
The models used to generate speech have varied wildly over the years as technology has changed.

And then there's the matter of how much the engine should do for you. Is "Dr." drive or doctor. Is MB always Megabyte? Sometimes it's millibars. One of the best examples of this is the Windows properties dialog box, the title of which reads "OS (C:) Properties".
One engine interprets the colon and closing parenthesis as a smily face emoteicon. as a kid I thought it did that when you had loads of space, and it'd change to c:( when space was low.

in reply to rem

The post, and eventual "solution" (in quotes because there was no problem) were interesting, and I think the attention to detail is excellent. I always enjoy seeing people learning more about what makes screen readers tick.

Unfortunately, the end result is that screen reader users everywhere will be copying the incorrect event name to the clipboard, creating an impression that they can't write a conference name correctly when they paste it on social media. It also uses more braille cells than needed (10/11 vs six, depending on translation table, repeated across the website).

Given that they end up hearing the "wrong" (in quotes because it isn't) version in the page title before they read any of the "fixed" (likewise) versions, I think the effort was disproportionate to the return on investment. @eric

@Eric
This entry was edited (3 months ago)
in reply to James Scholes

@jscholes @eric so much more eloquent than my pissed-off hey, who gets to decide how my screen reader works if not me?
Great points, as ever, and I also applaud the spirit to learn of course.
in reply to James Scholes

@jscholes @eric there's a lot to parse (and it's late for me, which is why I'm getting muddled), but! I'd say that *if* someone copies the text on the FFConf.org site, they're going to get the *correct* way to their clipboard (FFConf is correct, ffconf, looks "better" but not the letters I want copied).

Absolutely 100% that the effort was disproportionate to the actual problem - the small upshot is that I learnt more about navigating screen readers and a tiny bit about ligatures

in reply to rem

@eric some screen readers intercept the copy action I think, to provide formatted clipboard content. @jscholes will no more than I do I'm sure.



The Matrix Conference programme revealed, Circles going to the community, and bots updates.

That and more happened This Week in Matrix!

matrix.org/blog/2024/07/26/thi…

in reply to The Matrix.org Foundation

Looks like an amazing conference! I'm especially pumped to see the talks on Matrix 2.0, ElementX and decentralized identity!

Will they be uploaded to Peertube? (Another wonderful decentralized, FOSS platform!)

#Matrix #Peertube #ElementX #Element #Matrix2 #MatrixConf #FOSS

in reply to Blort™ 🐀Ⓥ🥋☣️

@Blort Hopefully yes! But we're not sure yet, we're still ironing out all of the details. We'll raise it with the team! It would certainly be fitting for us to publish on PeerTube.
in reply to The Matrix.org Foundation

I think the whole community is glad to see focus on hardening the base of protocol and targetting long standing problems. 🙌

Do you plan to publish recordings somewhere?



KDE signed an open letter for funding for Open Source. I'm not opposed, but I do want to say - given the billions that go in 'funding' to proprietary software - that changing the purchasing rules in the public sector is, in my opinion, far more important.

kde.org/announcements/2024_ngi…



re-upping this one again

so many forms of writing are rendered nearly useless if they have no info on *when* they were written

infosec.exchange/@0xabad1dea/1…