Apple @ Work: Red Canary Mac Monitor is a newly available tool for collection and dynamic system analysis on macOS endpoints 9to5mac.com/2023/06/17/red-can…
Discover the extensive capabilities of Mac Monitor, a dynamic analysis tool for macOS security. Explore how Red Canary leverages this powerful tool to perform rigorous threat research and uncover forensic artifacts.
could this blog post actually be right? did pgp really age so poorly? for those of you who know more about cryptography and how all this stuff works, can you shed some light on this? is it true, or mear exageration and it was written for the explicit purpose of writing a rant, like someone would write about absolutely anything at some point? latacora.micro.blog/2019/07/16…
Cryptography engineers have been tearing their hair out over PGP’s deficiencies for (literally) decades. When other kinds of engineers get wind of this, they’re shocked.
The UX is bad, because it could be a little better. But I don't get why the encryption is bad. It's not breakable, so it sounds good to me.
I only know gnupg, and this gives the option of separate signing, and encryption keys, but I don't know why I'd want that.
I don't know why the article says everyone should rotate keys. I'm not aware of rampant hacks which pull keys from the ram.
The article keeps mentioning normal users, but I don't see why. It's a tool for engineers, like SSH. And the idea that it's complicated is overblown.
The only 'key-signing' party I attended was people swapping keys at a standard Linux meetup. It takes like 5 minutes.
The statement that keyservers are centralised authorities is false. They're neither. You can find mickey mouse's keys and that's fine, because keyservers shouldn't verify anything for you.
I'm not sure how the alternative signing works if there's no ring of trust. Atm I have some reason to trust keys three steps from friends
@malin hmm, interesting, someone from the other side of this answers. I think the issues revolve around the fact that pgp mostly invented interesting ways to use existing algorythms, if not make some of its own in the process. Also the transport protocol, also key management, forward secrecy apparently is a big one, etc. About pgp being for engineers, no, it isn't, or at least it shouldn't be. All the encryption dogmas, mantras and sayings thrown around by privacy concious people were made by the creator of pgp, and they all imply that everyone should be able to have privacy, aka it's more of a tool to keep communications encrypted, while ssh is indeed not for the average user, since few people have an actual use for it
I'd agree that standard users should have privacy. If the PGP creators thing that anything involving a terminal will help, then they're crazy, but I'd assume they just mean normal solutions, like Matrix.
@malin Is the assumption made though not that PGP is used *by* other, more user facing apps and services, thus opening them up to some or all of PGP's design deficiencies? I think the author is specifically targetting email for a lot of this, although I know absolutely no-one who isn't a nerd who knows what PGP emails are to be fair, and everyone still assumes that emails are secure unless they're told otherwise. It's crazy to me that email is perhaps the world's largest communications backbone for sensative data, and it's probably also the least secure.
@malin In recent versions of git you can actually sign commits with SSH keys. I've set that up, Github and other forges support it, or are (hopefully) adding support, and it works great.
Nechci sem sdílet z Facebooku, tak příspěvek sdílím přes náš PrivateBin privatebin.arch-linux.cz/?fc81… Je tam skupina co bojuje proti dezinformacím. I v komentářích se snaží vše vysvětlit. Zatím myslím super počin.
no, ok, ale ten "marxismus" je tam fakt hodně fiktivně, na tom seznamu. Většina jakýchkoliv rebelů proti oficiální vládě kdekoliv, když chtěla pomoc z bývalého východního bloku, tak se z pí-ár důvodů prohlásila za něco komunistického nebo marxistického. Ale často to byla poměrně fikce, leda se to někdo dodatečně snažil fakt zinscenovat.
Dneska když nějaké režimy chtějí podporu Ruska, tak si už na žádný marxismus ani socialismus nehrajou. (myslím se naopak na konci SSSR dnešní Rusko inspirovalo u režimů, jako je ten Íránský a uvědomilo si, že jakkoliv progresivní ideologie jsou většinou v rozporu s agendu, kterou potřebuje prosadit a hodně na tuhle strunu hraje i v rámci svých současných vlivových operací třeba v Evropě...)
trurl 0.8
Changes since previous release
o add --urlencode
Bugfixes since previous release
o --json/--get: .[].parts and {component} should be urldecoded not encoded
o tests: use empty file instead...
Trump Says His Hand-Picked Chief of Staff Was ‘Born With a Very Small Brain’ The former president is appealing to voters by explaining how he appointed a bunch of incompetent losers the last time he was in the White House #News #GOP #Politics #USA #Trump #MAGA #Republicans
We've been told Google Play transparently shows what data is collected etc, right? And whether an app has Ads? Well, dream on. I've just checked an app to see if it would fit my repo. Below is what Play (play.google.com/store/apps/det…) says about it (no "Ads" marker), then the results from my library scanner.
@MartinBe 😮 Yeah, but why bother? If there's "no data collected", then… wait, is that a double-meaning? "This not collected data is not shared" (double-negation) 🤔
@MNandor Sure I can. And yes. And yes, I should hope so. F-Droid thinks so, too, as it's one of the scanners used there (with Issuebot).
OK, let me check: android.izzysoft.de/articles/n… gives you some background and also tells you were to find it. It's FOSS of course. Library definitions are regularly updated (about once every 3..4 weeks – last update just 2 days ago). Help welcome, especially with filling gaps like these: gitlab.com/IzzyOnDroid/repo/-/…
Die meisten Android-Apps enhalten zusätzliche Module anderer Anbieter/Entwickler. Manche davon sind für die eigentliche Funktionalität relevant, andere nicht.
So yeah the quality is bad, and once you play with it enough, it seems to be very predictable and very cautious. So uh... just... keep making music manually please. This ain't worth it. In the slightest.
Don't get me wrong. This is a pretty good imitation. Probably some of the better ones I've heard. Once it chooses a theme it seems to be relatively consistent and stick with it. But just like most AI you have to ask yourself is it really worth it? This thing requires significant resources to run. And so far, none of the prompts I gave it were exactly what I was looking for. Just like how you have a broad idea of what you want, type your image prompt into stable diffusion, it comes out with something you didn't want and suddenly you realize you did know what it is you wanted and make it yourself? It's kinda like that.
Yeah it's all a load of rubbish. even ElevenLabs as fun as it is to play with, you don't know what the output will sound like until you hit generate. If I needed something like that recorded I'd pay a human to do it.
> Extrapolating the gains of recent years might suggest that by 2025 the error level in the best deep-learning systems designed for recognizing objects in the ImageNet data set should be reduced to just 5 percent. But the computing resources and energy required to train such a future system would be enormous, leading to the emission of as much carbon dioxide as New York City generates in one month.
Yeah that's... great isn't it? Feels like crypto all over again.
you are right. The same people that were peddling crypto are peddling this. With the same rhetorical tools. The same critics using the same tools they used for crypto, critical thinking and rationality and system thinking, to debunk it.
And we are in the same cycle. See also VR, 3D video, metaverse, Blitzscale, etc etc
i mean it churn for a reason. We need something to pour all the boomer money into. Otherwise how could they get their super high pensions...
I think we are at 85% of the public market owned by institutional investors aka charities and pension funds. So they moved to the private market 15 years ago.
@Di4na I almost can't believe I'm saying this and I don't want to turn into the "actuaaaalllyyy crypto has genuine usecases" person but I've played with llava, which is an open source multimodal network that can answer questions about images, and as a blind person this is exciting. But that's a very specific and niche usecase. Just like how people losing their voice due to disability could make use of voice cloning.
That is the fundamental question right? We can imagine tons of tools that you could graft crypto to, but does it genuinely make it better and is it genuinely making it better than less problematic alternatives?
aaaah oh it definitely has use cases. The question become, can you actually trust the result. Knowing that they are specifically trained to deceive humans into thinking the result was good.
And then, once you built the stuff around to reach a good enough level, was it cheaper and better than actually funding humans to do it without the AI?
I mean i get it as a hack around our societal limitations into providing it buuuut...
@Di4na I know the results are unreliable and keeping that in mind it really does help. It's also why I'm advocating being able to run this stuff on device, because I might have some pictures that I feel awkward about having someone describe, or privacy reasons, or whatever. Obviously it is no substitute for good alt text. That is always more important. But if I take a picture of something then obviously there won't be any alt text.
right but my point is not that it is unreliable. It is that by definition of how it works, it is really hard to know if it fits or not. That is kind of worse than unreliable.
@Di4na so I shouldn't be able to get automated help with getting images described in a privacy friendly manner? I mean if there's a better way I'm all for it, but right now it's the best we have. I don't always have people I trust to get images described for example. I dunno if you don't need a tool and have no use for it, saying it's worse than useless is perhaps a bit heavy handed. By no means am I saying alt text isn't preferable. It always is. But there are situations in which manually written alt text is unavailable, and I'm very grateful that I can have technology help me with that.
yes. As i said the question become if the "bad" one impact is more than compensated by the help. So far in a lot of situation research shows no, but as pointed above i am not saying it has no use. It is also really far from what is being hyped :)
That may change... There was a paper that was published recently about a new method to basically replace the existing transformer model with one which is much more efficient. Something about context I think, can't remember now. It aims to replace the transformer architecture entirely which means models of course will have to be retrained, but they did resource comparisons and it was staggering. Something like 20x to 100x improvements in speed and resource usage I think. I wrote a post about it a while back about that allowing us to run it on embedded devices.
There have been some important improvements since that article, but it is still very much true that a neural network is (at best) the second best solution to any problem.
@pettter Yeah I should probably have made my relative ignorance about the subject clearer. I'm personally not really interested to actually build any neural networks so I haven't been keeping super up to date with advancements in architecture.
I agree that's a huge problem. At the same time some smaller free software models have been getting very effective for more specific tasks, with better training data. This is an interesting article on it: semianalysis.com/p/google-we-h…
And, as the energy transition happens, one expects the issue of carbon will also improve.
@modulux As we run into limits of what we can do I think the smaller networks might see most innovation. I mean I do already uze a lot of smaller models like image recognition that runs directly on my phone. And I wouldn't want to miss that. Because it's actually quite helpful even if the accuracy is relatively low.
Yep, that's my hope as well, and it also removes a lot of the concern of having to upload very private data to systems I don't control. I'm quite interested in solutions that can run on local systems.
@modulux And researchers are actively working on bringing model sizes down and use different techniques to train them so that they could be run on the edge.
There's a particular frequency in there around I'd say G5, which sounds like bluetooth when it's connecting to a device and hasn't quite connected yet.
the more I think about it the less it sounds like an advertisement and more like the background track of a hypercasual mobile game. Now just need all the sparkly UI noises...
@Brynify this one definitely sounds better than any other I've heard from a consistency standpoint but I really do wonder if we're able to get better results with the hardware we have... and obviously if we actually should.
Really like a lot of this. Agree about the sound, but you could definitely sample some of the ideas into your own music. Diffusion models are quite good at drums, and you can hear that in the snare for example. I've been playing around with this on the HF demo, and really like it when compared to Music LM. Need to try Collab soon. I'm in a discord server dedicated to Harmonai, an org working on similar tech, and some of the guys in there have done some really cool stuff with MusicGen.
Muchas de las concesiones para la explotación #hidroeléctrica van a caducar en los próximos años. Se me antoja fundamental que desde la #izquierda empecemos a presionar cada vez más para formar una empresa pública.
Porque sólo desde lo #público se puede gestionar para el interés general un recurso vital como el #agua, cada vez más escasa. Y porque así los beneficios de esa electricidad no irán a Iberdrola, sino a todas las españolas.
El próximo Gobierno deberá decidir el futuro de cientos de megavatios (MW) de energía hidroeléctrica cuyas concesiones han caducado o lo van a hacer en los próximos años.
Russian moon landings, week long traffic jams, a workforce replaced by automation and above all, too much leisure time!These are just some of the bold predic...
Tak nám prasklý zadní ráfek na silničce 😔 V e-shopech lze koupit pouze obě kola, nebo jen ráfek a nechat udělat výplet. To je zase dilema. Mít to rychle, ale zaplatit za kola 6500, nebo počkat 4 týdny a nejezdit, ale ušetřit 3000.
Ian Failes of before and afters has an utterly fascinating article about the tests to add motion blur to Phil Tippett's stop motion dinos for Jurassic Park. While I knew of those attempts, I don't recall ever hearing that they used MORF (used in Willow) to try to make the dinos more realistic. Eventually the CG tests paved the way for a #VFX revolution
I've built an experimental prototype web-app for browsing #FDroid repositories. It only comes with a very limited set of features, but I'm quite happy how it turned out so far.
We’ve created a prototype of a progressive web app for browsing F-Droidrepositories. It’s built withFlutter, which is really great for working in rapiddevel...
Today I released Footage; the simplest tool to crop, trim, rotate, flip, mute, and export your video to a format of your choice. It's now available on flathub! flathub.org/apps/io.gitlab.adh…
Oh good grief: A Ukrainian official confirmed reports that originated from a Russian milblogger that soldiers stood still for two hours for a commander's speech, leaving them sitting ducks for Ukraine's high mobility artillery rocket system, or HIMARS.
a more official announcement 4 my guest spot at los diablos #tattoo in #mexicocity shares appreciated (: . ( sorry this has instagram usernames… one day i hope the tattoo industry won’t be so cemented within that app… )
Cool, the Reddit CEO straight up said he was inspired by Musk, if that's any indication for the future of the platform: twitter.com/MattBinder/status/…
Well, not that surprising but I agree, it's probably the best indicator for Reddits future we could have ever gotten. What I wonder considering that Reddit always had a thing for some level of free speech (at the beginning even quite literal) is if he got further inspirations from Musk, the bs that idiot tries to sell as free speech is a lot more rediculess garbage than we are used to even from other social media CEOs.
"What’s not happening is us continuing to subsidize businesses built on taking our data for free. That’s not changing" Steve Huffman theverge.com/2023/6/15/2376286…
Reddit's data. Content that reddit didn't create. These digital feudal lords
Here at Accessible Android, we strive to create helpful resources for both Android users and developers. If you're an Android developer and learning about Android accessibility, we posted a guide on the basic steps for labeling buttons on your app. erisilebilirandroid.com/code-e…
In this tutorial, we'll provide you with some code examples that demonstrate best practices for button labelling on the Android platform, enabling you to create a more inclusive and accessible user experience.
@queenslight Thanks for the heads-up. We have more details about Accessibility Scanner and other testing methods on our Developer Resources page erisilebilirandroid.com/develo…
quora is now spectacular failures in auto-placed ads, then ChatGPT presenting you the most milquetoast take ever, then some other recommended questions and then fuck you that's what's up
Po více než 45 letech by mohl Čech znovu letět do vesmíru. Česká republika dostala v rámci soukromých misí Axiom nabídku poslat astronauta na oběžnou dráhu Země a na Mezinárodní vesmírnou stanici.
tak určitě by to byl zdroj určité prestiže nebo národní hrdosti, ale přeci jen to bude stát miliardu. a asi bych se ptal, jak tu miliardu utratit co nejefektivněji a v nějakém kontextu dlouhodobější spolupráce...
Otázkou zůstává, co z toho? Je to v zásadě komerční let, kterým si firma vydělává na své další projekty (primárně na stavbu vlastní vesmírné stanice). Pokud by to mělo pro nás nějaký přínos a ten v článku jasně definovaný není, tak je to zajímavé. Pokud by to bylo jen o prestiži, jsou to vyhozené peníze, kterým někomu jinému financujeme jeho podnikatelský záměr.
If you care about such things, vote whether TeamTalk, removes the Classic accessibility client, now that the regular one is as accessible as the classic one. github.com/BearWare/TeamTalk5/…
Hi to everyone, as you know, since TT5.8, we've stopped developing classic client. Now, thanks to Qt6 and also thanks to some adjustments we made in TT directly, the Qt default client is more acces...
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).
Hey Lagos! 🎉We're excited to announce that #GNOME is a sponsor at this year's biggest tech event, #OscaFest! Happening at Land Mark event centre Victoria Island, Lagos. Come check out our booth!!! festival.oscafrica.org/
This accessibility-alt-text-bot leaves automated reminders in a comment when a user shares an image without providing meaningful alt text. And you can add it to your GitHub projects. github.blog/2023-06-12-make-yo… #Accessibility #a11y #GitHub
The level of GET BACK TO THE OFFICE desperation in the right-wing press is hilarious, and you won't be surprised to hear that this one is content marketing by an office furniture firm
Discover topics like acab, defundthepolice, grimdark, batman, 39, and the magic of the internet at Imgur, a community powered entertainment destination.
If you're interested in using Typescript for GNOME apps, you can now use the new Typescript SDK Extension. I've also published a Typescript Template you can use as a starting point for your apps.
I've been working on both of these for a while now, and I think they're finally ready to be used by others in their applications. Please give them a try and provide any feedback you might have :)
this looks way nicer than horrendously full of boilerplate and macros bindings for Rust. Might as well try to use it, but I wonder how I would bridge TS with C or Rust libraries for the backend.
@mks_h The really neat thing with the TS "bindings" is that they're not really bindings. They're type definitions that are there solely to tell `tsc` about introspected types. If the backend code you want to use is introspectable, you can use it in TS code like you would with JS code. You can also generate type definitions from the .gir files with gitlab.gnome.org/ewlsh/gi.ts/.
WAW! That's supper, I knew the existance of nodegui, and also some project to use gtk with javascript, is more or less the same? (a bridge between node and libadwaita?)
@eickot not really - node isn't used here other than to install eslint and prettier. GJS is a JavaScript runtime that can directly use introspected libraries, and that's what I'm using here.
Reddit is messaging moderators of communities that closed to protest its API changes. It says anyone “hindering” reopening can be replaced with someone willing to bring the subreddit back online.
Malinux
in reply to bgt lover • • •The UX is bad, because it could be a little better. But I don't get why the encryption is bad. It's not breakable, so it sounds good to me.
I only know gnupg, and this gives the option of separate signing, and encryption keys, but I don't know why I'd want that.
I don't know why the article says everyone should rotate keys. I'm not aware of rampant hacks which pull keys from the ram.
The article keeps mentioning normal users, but I don't see why. It's a tool for engineers, like SSH. And the idea that it's complicated is overblown.
The only 'key-signing' party I attended was people swapping keys at a standard Linux meetup. It takes like 5 minutes.
The statement that keyservers are centralised authorities is false. They're neither. You can find mickey mouse's keys and that's fine, because keyservers shouldn't verify anything for you.
I'm not sure how the alternative signing works if there's no ring of trust. Atm I have some reason to trust keys three steps from friends
bgt lover
in reply to Malinux • • •Malinux
in reply to bgt lover • • •Michael Connor Buchan
in reply to Malinux • • •Malinux
in reply to Michael Connor Buchan • • •@TheFake_VIP Yea, Matrix definitely looks like a better standard-user communication experience.
Of course, I don't see anything replacing it for signing git commits.
Michael Connor Buchan
in reply to Malinux • • •Malinux
in reply to Michael Connor Buchan • • •