So, while this is not exactly a typical "self-hosting" question as many users might not be using domains, I would be curious if anyone else has any experience with this....
I digged a bit into the source code and the apk. From looking at the code alone one can’t tell if the crash report is actually enabled, the build configuration depends on some unpublished file. But looking into the apk allows to reconstruct it. These are my findings:
the crash handler is compiled in and also enabled (BuildConfig.ENABLE_ACRA=true)
the crash handler is configured to dialog mode. According to the ACRA documentation (www.acra.ch/docs/Interactions#dialog) that means that user interaction is required for sending (a popup dialog with a cancel button).
the privacy policy is correctly stated (voiceinput.futo.org/VoiceInput/PrivacyPolicy) > 4.1. If the app crashes, you may be asked if you wish to submit a crash report. If you accept, your device information and crash details will be sent to us for the purposes of investigating the crash and improving the software.
There are scammers that are sending phishing links inside a Google docs drive as in this way it’s less likely to get blocked by spam filters. Like “your statement is ready and you’re eligible for a refund, click here to check”
By mass replacing all the links with a Google redirect, (it shows the original one but when you click it changes), Google now can block those scam domains
The side effect is that now the system can be used to collect all the visit stats. But maybe they’re not tracking visits as they can already get full browser histories from Google chrome users with account sync, and that’s enough data
Let users collaborate on the list of accounts and domains they want to block. E. g. have a special type of channel and every link that is posted and upvoted is blocked.
X, formerly Twitter, is no longer showing headlines on articles shared on the platform.
Instead, X is only showing the article’s lead image and the domain it will link you to.
Fortune reported in August that this change was in the works, and X owner Elon Musk confirmed that the switchover was “coming from me directly” and would “greatly improve the esthetics [sic].” That leads me to believe that the headlines will be disappearing from the web and Android eventually.
Musk has recently been encouraging users to post more content to X directly in an effort to help boost engagement on the platform (and perhaps also because he “almost never” reads “legacy news anymore”).
He’s also said that the platform’s algorithm “tries to optimize time spent on X,” meaning that “links don’t get as much attention” and that the “best thing is to post content in long form on this platform.”
In my opinion, removing headlines makes X harder to use — posts are less easy to parse at a glance — so I’m not sure if this change is going to lead to people posting more often like Musk hopes.
The original article contains 247 words, the summary contains 191 words. Saved 23%. I’m a bot and I’m open source!
We are facing constant problems with the desktop apps in O365, wheter it’s RDS servers that somehow are Azure joined by a user from login 1001 errors to modern authentication Windows that automatically disappear or other generic error 1001 logon bullshit. We have a tome of registry bullshit with shit like EnableADAL to...
I’ve run into a lot of issues stemming from people creating personal Microsoft with the same email as their O365 email(why Microsoft? Fucking why do we even allow this?)
So they’ll mess around and log into their personal on one app and their work email on another, and then everything goes to shit because I guess Microsoft doesn’t quite grasp that having two accounts to very similar systems that have the exact same username is asinine.
Anyway, make sure you remove every account from the system, including removing the device from any azure domain, rejoin, and hand hold users through signing it.
DNSSEC signs DNS records so you know they’re genuine and come straight from the authoritative nameservers for the domain.
DoH encrypts DNS traffic so nobody can eavesdrop on what domains you connect to, and masks it as HTTPS traffic so providers can’t block it to force you to use their nameservers.
Regarding adoption: you can give a user DoH in the browser without them having to know about it, but you can’t enable DNSSEC for a domain owner or nameserver admin without their explicit approval. This will naturally lead to some adoption disparity.
What’s the easiest way to get https while still using my given tailnet as domain for accessing stuff? The tailscale documentation suggest to download certs to the server and point each service to those certs, but that seems like more work than it should…?...
It’s possible to host a dns server for your domain inside your tailnet, and offer dns responses like: yourwebserver.yourdomain.com = tailnetIP
Then using certbot let’s encrypt with DNS challenge and api for your public dns provider, you can get a trusted certificate and automatically bind it.
Your tailnet users if they use your internal dns server will resolve your hosted service on your private tailnet ip and the bound certificate name will match the host name and everyone is happy.
There’s more than one way though, but that’s how I’d do it. If you don’t own a domain then you’ll need to host your own private certificate authority and install the root authority certificate on each machine if you want them to trust the certificate chain.
If your family can click the “advanced >continue anyway” button then you don’t need to do anything but use a locally generated cert.
There is so much work out there for free, with no copyright
There’s actually a lot less than you’d think (since copyright lasts for so long), but even less now that any online and digitized sources are being locked down and charged for by the domain owners. But even if it were abundant, it would likely not satisfy the true concern here. If there was enough data to produce an LLM of similar quality without using copyrighted data, it would still threaten the security of those writers. What is to say a user couldn’t provide a sample of Stephen King’s writing to the LLM and have it still produce derivative work without having trained it on copyrighted data? If the user had paid for that work, are they allowed to use the LLM in the same way? If they aren’t who is really at fault, the user or the owner of the LLM?
The law can’t address the complaints of these writers because interpreting the law to that standard is simply too restrictive and sets an impossible standard. The best way to address the complaint is to simply reform copyright law (or regulate LLM’s through some other mechanism). Frankly, I do not buy that the LLM’s are a competing product to the copyrighted works.
The biggest cost in training is most likely the hardware
That’s right for large models like the ones owned by OpenAI and Google, but with the amount of data needed to effectively train and fine-tune these models, if that data suddenly became scarce and expensive it could easily overtake hardware cost. To say nothing for small consumer models that are run on consumer hardware.
capitalists just stealing whatever the fuck they want “move fast and break things”
I understand this sentiment, but keep in mind that copyright ownership is just another form of capital.
Thank you for the kind words, but I’m probably not going to do this for meme communities for the following reasons:
lots of images means hosting a lot of content and a big storage bill.
lots of users with lots of comments/submissions, will be hard to mirror those without hitting API rate-limiting ceilings.
if the mirrored instance gets lots of bot accounts who post nothing but low-quality content, it will make the people here on Lemmy associate it with spam.
But if you really want those, you can get a domain name and I could perhaps host it for you on communick.com ?
Since my favorite reddit app came to Lemmy I’m really keen on getting more people into the fediverse to pump up the volume of content around here. Are there any initiatives that we can assist to get folks onboard?...
The thing that kind of sucks about lemmy is there isn’t really any protection against fascists on the site. One of the reasons it took me so long to get off reddit is because there you have access to tools that let you see if someone you’re interacting with is an overt and open fascist, but nothing like that really exists here. In fact, it’s even worse here because the fascists will aggressively downvote to the point where anything directly calling out white supremacy gets absolutely slammed. Now you have a bunch of reddit frogs coming over here and the only real hint that they’re going to cause trouble is if their username ends in @lemmy.world or @feddit.de
The domain block is a bare minimum, I never want the displeasure of having to deal with a feddit,de poster ever again. Another thing they need to do is make votes public so I can clean house of people upvoting blatantly abusive comments or partaking in downvote harassment. Third they need to add tagging and user-level vote counts so you can identify known trolls without needing to commit their usernames to memory. Those three changes would go a long way in fixing a lot of the biggest problems with lemmy as a whole.
EDIT: And blocking a user shouldn’t delete them completely from your client but rather hide them. That way you can follow their comment streams looking for people supporting them and wipe them out in the process. The current system gives every comment below the original carte blanche to say whatever and there’s fuck all you can do about it because as far as you know, they don’t even exist.
The feature that Liftoff has is automatic redirects. So for example if I am lemmy.one user and want to subscribe to !memes, clicking the Subscribe button on the Liftoff app will offer to redirect me to lemmy.one/c/[email protected]. On the web interface, you would have to manually go to that domain.
Also all the links in my comment automatically open in the Liftoff app.
GPUs from all six of the major suppliers are vulnerable to a newly discovered attack that allows malicious websites to read the usernames, passwords, and other sensitive visual data displayed by other websites, researchers have demonstrated in a paper published Tuesday....
I’m setting up a jellyfin server, and want to access it on the internet. I created a xxxxxxx.duckdns.org address for it. I have installed caddy with duckdns addon (first installed regular caddy, then overwrote /usr/bin/caddy with this custom caddy). My caddy file is as follows...
Evil.com phishes a user and asks for username and password for Good.com
Evil.com immediately relays those credentials to Good.com
Good.com asks Evil.com for TOTP
Evil.com asks victim for TOTP
Evil.com relays TOTP to Good.com and does a complete account takeover
The various physical dongles prevent this by using the asking domain as part of the hash. If you activated the dongle on Evil.com, it’ll do nothing on Good.com (except hopefully alerting the SOC at Good.com about a compromised username and password pair).
It was my home to my account. I didn’t catch any announcement it would be going down. 2 weeks ago the website started throwing errors (50* internal server errors) and since a week it is completely unreachable. There are still DNS records resolving the domain, but the site seems completely gone.
There’s another project, lemmy_handshake, which is an Android app (YMMV, I haven’t tried it.)
It’s not too difficult to use Oracle’s free tier and Lemmy-Easy-Deploy if you want to register a domain and set up your own single-user instance. I do that, knowing that it could poof at any time. I run lemmy-account-sync as a cron job nightly on the same Oracle instance, but hosting your own instance isn’t required for syncing.
I sync my main account with accounts on a few small instances. I chose them from the list of Lemmy nodes which are on the current version of Lemmy, that have active users. Small instances tend not to defederate other instances so much, if that is important to you. They are also less likely to be targets of DDOS attacks. They can also wink out of existence without warning, which may be the case with lemmy.villa-straylight.social.
I also sync my main account with accounts on a few of the larger instances. (I mainly use Lemmy Explorer to find Communities, but big instances are best if you just want to doom scroll “All.”)
Should I tire of self-hosting, or if Oracle decides to randomly delete my instance (a real risk), then I’ll just log into another instance.
You’ll lose some stuff (like post history, and private messages) but it will be better than losing everything again.
I hope he realizes that’s not how it works. Companies own domains they don’t use for this exact reason. Hell, most major companies own hundreds of misspelled versions of their domains across many TLDs for user safety.
Forgive my ramblings, but here’s the main differences I see, from a community perspective:
Bluesky’s for people who loved twitter circa 2015
Mastodon’s for people who loved the format but hated the way the platform made use of it. The community is FOSS-focused and anti-corporate.
Bluesky folks are anti-corporate, but they still want their social media to be on a single platform and tend to dislike federation
Mastodon folks tend to be in smaller circles and more tech enthused
Features-wise, Mastodon kills the algorithm in favour of chronological timelines and lists, while Bluesky embraces algorithms, allowing people to even make their own algorithms for the platform. Bluesky’s AT Proto uses “DIDs” to identify users, which are associated directly with a domain^[or subdomain]. This means that when federation does eventually happen, usernames will just be @my.domain.com instead of ActivityPub’s @actor.
Federation’s still not enabled so I have no clue how things will look and feel on that front, nor am I familiar enough with the protocol to make any claim about how versatile it is. ActivityPub is flexible enough to be a Twitter clone, a reddit clone, a blogging platform, a youtube clone, a twitch clone, a goodreads clone, or several other formats. AT Proto’s currently only proven to work for a Twitter clone.
When federation is live on the main node on bluesky there will still be some similar effects when you follow links from other servers, in that you’ll need to bring that over to your own server somehow to follow and interact.
With Mastodon when you follow another’s link you’re asked to specify your own instance, in Bluesky you’ll enter your domain based username and it will find your instance.
Also with the CDN like BGS caching servers being shared across instances you’ll be able to find more content from your home instance so it will feel more like Twitter. You can directly search for users on other servers.
If your running behind OPN/PFsense I’ve found the easiest solution for internal only SSL is to use the router to create the certificate chains. Yes you’ll have to import 1 CA cert on each end user device but only the one then you can crank out internal certs without and https warnings or domain constraints/challenges.
Highlights include Sliding Sync (instant login/launch/sync), Native OIDC (industry-standard authentication), Native Group VoIP (end-to-end encrypted large-scale voice & video conferencing) and Faster Joins (lazy-loading room state when your server joins a room).
NGinx Proxy Manager and Domain Routing to Docker Services on Specific Ports (kbin.social)
So, while this is not exactly a typical "self-hosting" question as many users might not be using domains, I would be curious if anyone else has any experience with this....
Where to get free cock.li invites?
I recently wanted to use cock.li as my main email service but I’m not sure where to get a free invite
This gives Google LESS access to your data! - YouTube (youtu.be)
Google adds tracking to Google docs documents? (fosstodon.org)
Title says it all. Somewhat interesting if true. I wouldn’t be surprised either way.
people that use reddit repost bots- are you okay? why do you suck?
Title
X stops showing headlines because Elon Musk thinks it will make posts look better (www.theverge.com)
Are we the only shop with constant login bullshit on Office 365 desktop apps?
We are facing constant problems with the desktop apps in O365, wheter it’s RDS servers that somehow are Azure joined by a user from login 1001 errors to modern authentication Windows that automatically disappear or other generic error 1001 logon bullshit. We have a tome of registry bullshit with shit like EnableADAL to...
Firefox rolls out ECH enabled by default in 118 (blog.mozilla.org)
ECH (encrypted client hello) is going or get enabled by default (already existed in a hidden setting) with version 118....
Https on tailnet?
What’s the easiest way to get https while still using my given tailnet as domain for accessing stuff? The tailscale documentation suggest to download certs to the server and point each service to those certs, but that seems like more work than it should…?...
Suing Writers Seethe at OpenAI's Excuses in Court (futurism.com)
Lemmy instances that are focused on mirroring Reddit content?
I’ve posted before about my fediverser project, and I am now looking to see who is interested in participating....
How can we boost Lemmy membership?
Since my favorite reddit app came to Lemmy I’m really keen on getting more people into the fediverse to pump up the volume of content around here. Are there any initiatives that we can assist to get folks onboard?...
Germans caught celebrating Oktoberfest with Nazi salutes (www.jpost.com)
Performing Nazi salutes has been illegal in Germany and Austria since the end of World War II and is punishable by law with prison time.
Reddit is dead. Long live the Fediverse. (lemmy.ml)
GPUs from all major suppliers are vulnerable to new pixel-stealing attack (arstechnica.com)
GPUs from all six of the major suppliers are vulnerable to a newly discovered attack that allows malicious websites to read the usernames, passwords, and other sensitive visual data displayed by other websites, researchers have demonstrated in a paper published Tuesday....
Need some help. Duckdns reverse proxy doesn't seem to be working.
I’m setting up a jellyfin server, and want to access it on the internet. I created a xxxxxxx.duckdns.org address for it. I have installed caddy with duckdns addon (first installed regular caddy, then overwrote /usr/bin/caddy with this custom caddy). My caddy file is as follows...
I will burn your servers to the ground, foul villain (startrek.website)
HyperTech News Report #0001 - Happy FOSAI Friday!
cross-posted from: lemmy.world/post/5549499...
HyperTech News Report #0001 - Happy FOSAI Friday!
cross-posted from: lemmy.world/post/5549499...
Does anybody know what happened to lemmy.villa-straylight.social?
It was my home to my account. I didn’t catch any announcement it would be going down. 2 weeks ago the website started throwing errors (50* internal server errors) and since a week it is completely unreachable. There are still DNS records resolving the domain, but the site seems completely gone.
X Announces It’s Shutting Down ‘Circles’ As of October 31st (www.socialmediatoday.com)
What does Bluesky have that Mastodon doesn't?
Is it just the momentum and word of mouth, or are there improved features as well?
Certificate management
How are y’all managing internal network certificates?...
Matrix 2.0: The Future of Matrix (matrix.org)
Highlights include Sliding Sync (instant login/launch/sync), Native OIDC (industry-standard authentication), Native Group VoIP (end-to-end encrypted large-scale voice & video conferencing) and Faster Joins (lazy-loading room state when your server joins a room).