The first day of the Hackathon at #IETF124 in Montreal is coming to an end. We implemented the relatively niche feature of XEP-0444: Message Reactions that allows channels or group chats to restrict the type and number of emoji reactions users can do.¹
This brings us a step closer to advancing and stabilizing the XEP. We didn’t want to do that before, since we aim to have complete implementations before requesting a Last Call.
Oh cool, did you implement a "restricted emoji picker" in Conversations? I only implemented "1 emoji/message" in @gajim
For groupchats I followed the spec, but for 1:1 chats lovetox did not want to rely on anything resource-related (client disco features/caps). I added the restriction form at the "server" (=gateway component) level, by adding a `scope=domain` field. It may be worth an extra sentence in 2.2 of the spec?
@nicoco @gajim Yes, I've implemented both the allow list and the max reactions per user.
Thus far I only support this in MUCs (We mainly wanted to proof that the XEP is feasible rather than create functionality that many users will use immediately.)
Announcing it on the domain (with 'scope') would probably solve a problem that @larma and I have been talking about too.
@nicoco @gajim I'm not sure we need to announce a scope, I would usually expect restrictions to apply to all "sub JIDs" (that is, it applies to full jid if disco'd on bare JIDs and on bare JIDs when disco'd on domains).
That said, the main issue with announcing it on the domain is that many clients don't disco their peer's domain (also typically not the bare JID) and it would be weird to disco all domains of all contacts just for the purpose of reaction restrictions.
@larma > it would be weird to disco all domains of all contacts I think the reasonable use-cases are MUCs and gateways. If you use a gateway, you probably have its disco cached somewhere. It makes sense to check for restrictions there specifically if one wants a client that integrates nicely with gateways. It should be "almost" free if it is already implemented for MUCs, I think?
@nicoco Can you elaborate why a client would typically disco the domain of a gateway service? Sure, if a service was disco'd before and I thus have the data available locally, I'd use it. For what I know, most client implementations only ever disco their local server, bare JIDs of MUCs and full JIDs for which they got a caps string in their presence.
@larma If you want to use a gateway, you likely need to disco it to look for registration mechanisms and adhoc commands you can use on it. I can't speak for all clients, but gajim disco JIDs in your roster, and if you have a slidge-based gateway that properly advertise its gateway type, it also uses that to add an icon to the avatar.
No, sorry, I can't possibly start on the thing that's due Monday or the one that's three weeks overdue, because I'm waiting for an entirely unrelated thing that won't be here until the end of next week.
"How to make emails sound human with the growing use of AI" Um, how about just write an email with actual thoughts from your own brain? I am also annoyed and will mute people who social media post with AI. If you can't come up with a post that isn't riddled with all the AI tells like over emoji, bullet points and ***'s everywhere? I'm not here to engage with you, and I am not sorry about that. ffs
@miki With all due respect, I was not referring to non-english speakers and I didn't think it needed to be said. But also, people who didn't have good english were able to social media just fine before ** and bullets and an emoji every other word became a thing. I just see people letting AI take over so many needless areas. Do we really need to have AI taking over how we communicate with our contemporaries too? It just strips the heart from a post for me and it feels especially annoying coming from a human. We shouldn't be ok with casually letting an AI essentially think for us because I just think that will lead to a generation of people who don't trust in their own thoughts and ideas and think AI can just automatically tweet/toot better.
@miki I personally translate all my English texts using GPT. I haven’t noticed it forcing its own style, like adding bullet points, asterisks, or emojis where I didn’t explicitly ask for them. Maybe the key is that I always write a prompt like “write in English,” followed by the exact text I want to be translated. If someone asks it to write something but doesn’t provide the precise wording, that’s when the AI lets its imagination run wild. :)
Gave into Windows 11, Microsoft's tactics worked on me. I really wanted to have no Ethernet connections to my desktop, just Wi-Fi - and 2.4 GBPS is just shy of the 2.5 over wired. So, Windows 11 it is. In order to get ram leaks and usage down to a lower amount, a lot of compromises had to be made. - no Search box on start menu. SearchHost.exe uses anywhere from 200 to 400 MB Ram, with it growing and never shrinking, so it was nuked per group policy in the registry. - Quiet CrossDevice (CDP): HKLM + HKCU policies; stop/disable CDP services; disable CDP/SharedExperience tasks. CrossDeviceResume.exe was always using anywhere from 62-100 MB ram. Thus its presence had to be removed. I don't care about doing things across devices, I just want my OS to be an OS, stay out of the way, please. - Memory knobs: disable MemoryCompression & PageCombining; App-launch prefetch & Superfetch OFF; SysMain disabled. I don't need memory compression when I have 64 gigabytes of it. - WpnUserService → Manual; TabletInputService → Disabled. Don't care for push notifications in Windows 11, they can die a blissful death to me. With this, Windows 11 at idle, after boot and using MemReduct to clean standby lists, can go as low as... Drumroll... 1.85 GB. Of course, I also have Windows Defender and security center killed because I don't want them. I'd rather be mindful and use a tool to manually scan a file against a virus after download, and run weekly scans myself, than deal with the overhead.
I'm curious how your ram usage will look after about 12 days of uptime. On my system with 10 IoT LTSC, with only TeamTalk and MUSHclient open, Windows is using an insane 9GB. I don't want to restart just to fix it, then have it spike again. Any helpful tips? or should I bite the bullet myself and put 11 back on.
@Blobsta I think you have to use MemReduct to clean out standby and cached pages. You can check additional ones in tis settings, like for the standby ones - easiest is from their release page at github.com/henrypp/memreduct/r… - with 9 GB, I bet something like it is holding most of the ram as hostage either in perged pages or standby cache.
@x0 oddly, Windows on 16 GB can run in as low as... 1.3, to 1.5 GB ram. It's crazy. It's somehow between 20-30 GB that it begins to push other caches higher and then it also has more. I had experimented with the removememory command in BCD to take off a chunk of ram that way, and it always amazes me how at least 5-600 MB of it comes around as you scale up from 12 GB to around 25. After 25 it gets stable at that 1.8 to 1.9 GB mark.
I figured it was doing something like that, because it manages to idle at about half usage on both 8 and 16 GB systems, or rather 10 did, IDK if anyone has the guts to run Windows 11 on 8 GB RAM.
I created a Github Gist with my script here, it has both a do and undo command to revert back some of the changes (but not app uninstalls.) If you're willing to deal with no search box in your start menu and no cross device experiences (phone link, content syncing across devices.) win11LTSC-debloater.ps1: gist.github.com/tgeczy/2d847e2… Also on Eurpod: eurpod.com/win11LTSC-debloater… (right click and save)
Some of y'all are saying, "I nee notification center!" For some, this just means you want notifications to appear still. In Windows 10 and even 11, you can re-enable the old balloon notifications. So that is what I also did. The script is now updated on the Gist to also do this and set the proper registry tweak. It's at: 'HKCU:\SOFTWARE\Policies\Microsoft\Windows\Explorer and the DWord item: "EnableLegacyBalloonNotifications". They still do appear like they did with Windows Xp and above, too, which is nice.
How do you actually activate balloon notifications with a screen reader though? Or keep a history of them? I was under the impression you could do neither. Unless JAWS did something for that that NVDA never did.
@x0 they were never actionable, that is, I think balloons were meant to just show a tooltip of info. Maybe though? Would be interesting to find that part out.
@x0 They were most definitely actionable. I suspect a significant factor in retiring the balloons was that the default visibility time was too brief for most people-- 5 seconds I think. God knows how many calls I fielded about them at my last job.
@remixman @sclower @x0 yeah precisely. I think you could sometimes press your mouse button to activate it, but all it did was launch the app that fires the baloon, it didn't have any fancy intents and stuff like that. Windows 10's notification center is quite nice, and was less bloated, because it didn't have to use these RuntimeBroker sandbox processes, and lived on as part of ShellHost still, rather than as a direct push notifications service.
@remixman @sclower Later WIndows 10 notifications are shell experience host actually, and runtime broker is certainly there, though unsure if it's responsible for them.
@x0 @remixman @sclower oh interesting - now I'm tempted to go back to just check out how it's changed. Ha, just another reason to. I recently started to also notice Windows 11 chew threw about 1 GB of SSD writes (checking precipitously with crystal disk info like a hawk) per 2-3 hours, and for the life of me I cannot figure out what would cause it to write that much over time when I'm just doing quick google searches, checking AI overviews, closing browser. Argh. Then I recently read an article about Win11 potentially chewing through SSDs at idle and I feel done with it again. xD
Apple's recently announced memory integrity enforcement (a combination of hardware and software changes) promises substantially to reduce the risk of memory safety-related security vulnerabilities. This podcast offers a detailed discussion.
Apple announced its new suite of memory security improvements from the top of the stack all the way to the bottom, so we dug through what they did and how th...
If you're the kind of person who is willing to create troll accounts of people, use their info and twist it just enough so you're enflamatory, I feel like you're also the kind of person who needs a little more happiness in your life. Please go outside. Breathe some fresh air. Touch some grass. live a little. The way to make yourself happy is not by hurting other people.
And more importantly, and what I actually came to say here, if you're that kind of person, please leave me and mine alone. Please unfollow quietly if you're here. Please walk away. I can't stand this petty, childish, and vaguely insidious bullshit. I have neither the time nor the spoons.
@miki Yes, I care. I know you're right, and I know that I probably shouldn't have said anything, but nobody touches my people. Nobody gets to have so much power over me that I'm scared into silence. Nobody is worth that.
They call us the "In-Between Generation." We are the last kids to play outside until the streetlights flickered on, our names carried on the breeze as our mothers called us home for dinner. We are the children of scraped knees, dial tones, and quiet Sundays. We are the bridge. We are the ones born roughly between 1955 and 1985. We grew up in a world that moved at the speed of a bicycle, not a broadband connection. We knew how to be bored. And in that boredom, we found magic. When it rained, we didn't have a thousand streaming shows at our fingertips. We sat on the carpeted floor, listening to the static hiss of a record before the music started. We watched raindrops race down the windowpane. That was our high-definition entertainment. Our afternoons were saved by a ball, a few friends, and a stretch of sidewalk. We built forts in the backyard from old blankets and cardboard boxes. We rode bikes with banana seats and high handlebars, our knees patched with bandages, our hearts full of stories. We are the last generation to remember the smell of mimeograph ink from school worksheets—that distinct, sweet chemical scent of a test or a permission slip. We knew the art of waiting. We wrote letters—on real paper, with real ink. We’d pour our hearts out, fold the page, seal the envelope, and wait. We’d wait days, sometimes weeks, for a reply. And when it came, we’d read it a dozen times, memorizing every word. We communicated on a phone that was tethered to the wall. The cord would stretch across the kitchen, tangling around our legs. If we wanted to talk to a friend, we had to be polite and talk to their parents first. We memorized phone numbers, a secret catalog of our social lives etched into our brains. We are the generation that made music. We sat by the radio for hours, finger hovering over the ‘Record’ and ‘Play’ buttons simultaneously, trying to catch our favorite song without the DJ talking over the intro. That mixtape was a work of art. It was a declaration of love. We grew up with trust, not passwords. We knew our neighbors. Friends didn’t text "here." They rang the doorbell. They stood on the porch and yelled your name. We left our doors unlocked. We were raised in a world where failure was normal and boredom was just the pause before a new idea. It was a world where things were allowed to break—and then they were repaired. We didn't just throw things away. We fixed toasters. We mended jeans. We worked on relationships. We witnessed the great shift. We saw the world change from black-and-white to color. We were the first to play Pong and Atari. We were the teenagers who first heard the scream and crackle of a dial-up modem connecting us to a new, invisible world. We learned to type on a keyboard, but we still have the cursive signature our grandmothers taught us. We understand what "the Cloud" is, but we still have shoeboxes full of faded photographs with handwritten dates on the back. We see the children of today, their faces illuminated by screens, and we wonder, quietly, what they have lost in exchange for all they have gained. We are the generation of real conversations, of kitchen tables smelling of coffee and cigarette smoke, of long-distance calls from a payphone with a pocketful of quarters. We are the generation that learned that happiness isn't a filter, it's a feeling. That connection isn't a Wi-Fi signal, it's looking someone in the eye. That memories are not stored on a server, they are stored in us. The world has gotten faster. Brighter. Louder. And sometimes, it feels colder. But we carry the quiet inside us. We are the bridge between yesterday and today. We are the translators. They may look at us and see an older time. But we are the anchors. We are the ones who remember the why. The past wasn't better. It wasn't perfect. But it was real. It was tangible. It was human. And that is a memory worth fighting for.
We apologize for the long performance degradation today. Finally, we identified all of the 'tricks' that AI crawlers found today. They no longer bypass the anubis proof of work challenges.
A novelty for us was that AI crawlers seem to not only crawl URLs that are actually presented to them by our frontend, but they converted the URLs into a format that bypassed our filter rules.
By the way, you can track the changes we have been doing via
💯 Agree. But pushing Carney to repeal anticircumvention laws will be a big challenge given the laws were essentially drafted by big tech, and exactly no one in government actually understands them.
"The world's governments have all signed up to "anticircumvention" laws that criminalize reverse-engineering & modifying US #tech products. This was done at the insistence of the #US Trade Rep... using the threat of #tariffs .... getting rid of anticircumvention laws only requires that governments control their own behavior – unlike taxing or fining companies, which only works if governments can control the behavior of companies" - @pluralistic
My work #EMR at now has integrated #AI that summarizes a patient's chart whether I want it to or not. This week it told me the wrong reason for admission, the wrong hospital course, and the wrong medications as compared against the human-written discharge summary. To review it and find the error took 3 minutes; to document the error and report it took another 10.
Anchoring bias exists. What we read stays with us, truth or lie, influencing decisions.
There is a lack of COVID-19 vaccine for each and every Canadian. Provinces and territories FAILED to plan when the federal government left it up to them to acquire the COVID-19 vaccine. THAT is why certain governments are charging for COVID-19 vaccine in 2025-2026.
There are people in Québec with appointments for FREE COVID-19 vaccines whose appointments are being pushed by a number of days by their local CIUSSS b/c of a lack of COVID-19 vaccine.
First impressions of #Gemini: 1. Interesting custom instructions management (bit by bit rather than one big text area which is #ChatGPT way). 2. A blazing fast and free gems creation (those are like custom GPTs in chatGPT, I believe — I can't create GPTs because it doesn't seem to be available for free users). 3. lots of weird accessibility quirks absolutely not expected from #Google: many unlabeled buttons and buttons like "Accessibility label for the button that closes the confirmation dialog" (instead of just "Close" or "Close dialog", at most).
from this reddit thread :) Microsoft: you're going to have to recycle your 2018 PC because we aren't providing updates anymore and we won't let you install the latest version of Windows (even though it works fine) Debian: just a heads up, the next version of Debian might not work on the Amiga 1000 or Atari ST. It is up to you to make the necessary changes. old.reddit.com/r/linux/comment…
Updated to #fedora43 and #kde plasma 6.5. As always, the process is smooth, fast, well-documented. It baffles me how macOS and windows became more and more clunky and broken, while overall Linux is stable requiring minimum maintenance.
Is there any escape from the "rich get richer" phenomenon in social networks where the people who have the most connections will garner the lion's share of the audience, drowning out those who are quiet and thoughtful, and the majority of users get basically no reach and no interaction, or is that inherent in the nature of all such networks? Seems like a lot of modern society to hang on essentially random founder effects.
My apologies if you're already familiar with the model, but this problem has been studied under the name of scale-free networks (networks having a power law distribution). There are a few simple generative mechanisms proposed for them, and it sounds difficult to get rid of them. en.wikipedia.org/wiki/Scale-fr…
#AI using tip: consider going to memory settings (#ChatGPT memory, #Claude now has it also) and removing garbage. I use free tier of ChatGPT, and it complains about full memory. I went there and now I'm removing lots of insignificant facts taken from random one-time chats.
Yeah, I have to constantly go through Gemini history and delete some of those one-off chats. I especially love those where the title will say something like, "a greeting" with the "hello" as the only message in the chat. :)
@vick21 Oh, I have to revisit Gemini actually, I'm kind of sticking with ChatGPT and Claude. What would you suggest, browser version or something else? I'm on Windows 99% of the time. Thanks!
To ,e honest, I mostly use Ai Mode inside Google search these days. Though if you search in languages other than English, then this could be problematic for the LLM. As for Gemini, a simple gemini.google.com will do! :)
@vick21 Oops, there is a bug. If I go to gemini.google.com/gems/view, I hear "Start a new conversation with this gem." for every gem. JAWS 2025, Chrome 141. Where can I report this please?
I am happy to say that I have only seen that horrible, wrong and incorrect phrase "daylight savings time" once in the last few days. Well, crap. Now I've doubled it. Anyway, there is no savings. Only daylight. And ya know what? Screw the daylight. Don't save it. Yep, that's right, no more Daylight Saving Time. The save is being inged. It is not the saves that are inged.
Now, if all that sounded entirely ridiculous, you're correct. So is Daylight Savings, because it's wrong, and the concept of Daylight Saving Time itself, because it's (mostly) dumb.
I mean, seriously? With 5g on my status bar, I get these errors: An error occurred while sending your post, A TLS error caused the secure connection to fail.. Is Apple lying or t-mobile does? ?
Otevřu aplikaci mujRozhlas a tohle na mne skočí. Nevěřícně na to koukám... Po celém dni online na #OpenAlt přemýšlím jak to pánové zařídili, že mají reklamu i tu. Chvíli to zkoumám a konečně pochopím, že to nejsou "naše" VHSky. cc @vhsky
Kelly Sapergia
in reply to David Goldfield • • •