Notices by Fabio Manganiello (blacklight@social.platypush.tech)

I've had to take down my Piped instance at piped.platypush.tech.

As my Linode bills were getting more and more expensive, I've had to move some services inside of my home network - and Piped is among them.

Once I moved Piped within my home network, it really started taking a toll on my connectivity. Streaming videos to ~20 users simultaneously, with no Cloudflare in front, from a home DSL connection takes quite of a toll.

Since many have been using the service but nobody has bothered to contribute to so far, I've decided to take it down and move it to a domain that I won't advertise this time.

I like to self-host, and I'll always keep doing it. And I also like to give something back to society - many have been using my services for years, and I'm ok if others use them. I'll always offer them free of charge and I'll always respect everybody's privacy.

But there's a limit to everything.

One thing is to offer a search engine aggregator or an ebook hosting service for free. Another thing is to provide a media streaming service, running on a home connection, for free.

It cost me money to run Piped on a Linode instance (precisely ~$100/month extra on my Linode bills for a beefier 16GB machine), and it costs bandwidth and electricity to run it on my home network.

Free software (and services) should be free as in speech, not as in a free beer.

I've asked multiple times for users to contribute to keep the lights on, but haven't received a single penny - all while I got on average ~5 video views per minute that prevent me from even having a videocall on my own home network.

If you really love free software, but you can't run it yourself on your machines for whatever reason, please at least bother to financially contribute to the instances that you use. Otherwise you're just a freeloader who does more harm than good to the free software.

@santiago I'm now in the process of migrating to #Forgejo, which is supposed to be a fully FOSS fork and drop-in replacement for Gitea - let's see how long this lasts before pushing all of its users to their cloud offering too...

@santiago Sourcehut is also an option, but IMHO it's too minimal and elitist. It belongs to that "we're elite h4x0rs that use 1990s interfaces and only communicate over email and we only use command line interfaces, and we expect everyone to do the same" suckless-inspired purist school that eventually pushed a lot of people away from many interesting open projects. But if Forgejo also had to go down the enshittification road in the future (I trust nobody at this point), then I guess the options would be between Sourcehut and coding my own source forge.

I the beginning there was (mostly) Github.

Then it got acquired by Microsoft, it started taking down repos and accounts at every whim, it started violating FOSS licenses by training their closed models on our open code, and I started self-hosting my Gitlab instance instead.

Then Gitlab enshittified, its open offering started falling apart and serious bugs were left open for months, it started to aggressively push users to their cloud offering, it even embraced openly hostile practices (such as deleting repos and accounts that hadn't been active for a few months), and I migrated to Gitea.

Now Gitea is also enshittifying, providing features in Gitea Cloud that aren't available on the open core (one example, a stupid limitation on the maximum number of users allowed on a server, when it's my own fucking business to decide how many users I want to store on my own database on my own machine), and probably it won't take long before they start pushing their users to migrate to the cloud version by breaking even more basic features on their open version. So I'm in the process of migrating to Forgejo.

Luckily I have enough technical skills to migrate things around and even patching code. But just because I can it doesn't mean that I should. Especially when the trend is that of a code forge migration per year just because all of them have decided to enshittify and are trying to upsell me stuff that I either never asked in the first place, or that was given for granted and it suddenly became a premium.

I just want to have a service to upload and share my own open projects, I'm fine to run it on my own machines with no customer support involved, and I'm fine to even contribute to their codebase. I'm not asking much. I shouldn't go through so much trouble. I haven't been always so hateful towards the techbiz, but the techbiz has made me so embittered because business people keep breaking *our* tech on a daily basis and force me to invest all of my time and energies in seeking alternatives.

I seriously just want all these business parasites that have taken over my industry gone and dead by now. I fucking hate every single one of you.

@SallyStrange @santiago fair enough, I could have chosen a less controversial word. But the context remains the same - what I'm denouncing is ideological purity with no functional purposes.

To summarize:

1. If instance A defederates Threads, then no content from A will be relayed to Threads, and the other way around. Regardless of what instance B does. This is just how the protocol works. The "additional layers of security" argument just doesn't stand.

2. If Meta wants to scrape the Fediverse, they already have plenty of resources for doing it, even without going full in with an ActivityPub-compatible platform. And, if an instance is public, there's basically no way of stopping them - I don't think that a company that didn't care of harming teenagers or encouraging a genocide in Myanmar cares about our nobot tags or robots.txt rules.

So, from a functional perspective, defederating those who don't defederate Threads *is* a mere act of ideological punishment with no added value.

I understand that many are on the Fediverse because they were traumatized/bullied on other platforms, and they want this to remain a safe place. But friendly fire and fragmentation definitely won't help anybody other than companies like Meta. We already have technological solutions to ensure that our own instances are impermeable to the content of platforms/instances we don't like. We shouldn't be overzealous and add a layer of ideological retaliation on top of that.

#Threads is now officially defederated on this instance.

Meta has gotten countless chances of doing things the right way, it has screwed up each single of them, and I want each single byte coming from that morally bankrupt monstruosity out of my life, my house, my network and my instance.

Note however that, unlike other defederation talibans, I won't give a damn if you're an instance admin and you decide not to defederate Threads.

I'm not afraid of Meta scraping my instance through other federated instances that haven't blocked them. If they want to scrape this instance, they can already do so via API or even search engine results, no need for federation - and they obviously won't give a fuck about your robots.txt or #nobot hashtags, we're talking of a company with no moral compass after all.

If you want to give Threads a try, go for it. If you still need more evidence that Meta is irredimable and rotten to the core, go for it. Who am I to ban you/defederate you for holding a different opinion?

@SallyStrange I defederated Threads for very similar reasons, but I wouldn't go as far as applying the transitive property to anyone who doesn't defederate them. I feel like it can quickly escalate in a nasty twist of the "my freedom ends where somebody else's freedom begins" principle. On top of that, gratuitous defederation makes the Fediverse more fragmented - and that's exactly what Meta wants.

Back online! All the mods of this instance have been successfully migrated too.

I'm taking the instance down for a few minutes (hopefully) to finalize the upgrade to #Mastodon v4.2.0👷

The engineers who designed the #Voyager probes half a century ago even thought of the possibility that a wrong sequence of commands may point the antenna dish away from earth (like someone did a couple of days ago).

And they implemented a self-adjusting mechanism that a few times a year scans the positions of a few known stars to infer the position of the earth, and point back the antenna in the right direction.

50 years later, these wonderful machines are still working, tens of billions of km away from earth, with only 69 KB of RAM, and even a wrong sequence of commands won't put them out of use, while nowadays 4 GB of RAM aren't even enough to start VsCode or IntelliJ.

The more I understand how they were designed, the more I feel like an early Medieval engineer looking at the Pantheon or other marvels of Roman architecture. Some amazing skills, knowledge and attention to details have been lost from that generation to ours.

Sysadmins are rightfully freaking out because #RedHat is making RHEL closed source.

Alternatives like Alma/Rocky Linux are doomed with just a tiny fraction of the resources to play a compatibility catch-up game with the (now closed source) RHEL.

CentOS has become a community-driven rolling release project.

Ubuntu is doing weird things and it has decided to become a system built around snap - even if nobody else likes snap, and even if snaps make Ubuntu-based Docker images pure bloatware.

And Manjaro is a potentially nice project that however is guaranteed to break after installing a few AUR packages with dependencies that are too fresh for Manjaro's repos (and, in general, it's always one step away from breaking badly: https://rentry.co/manjaro-controversies), and it's managed by people who have even failed to update their own SSL certificates.

We often say that Linux won on servers, but Linux on servers definitely isn't in a good shape. If Linux runs everything and we can't afford it to go down, then it needs more public funding, or private companies will keep doing stupid things and kill the toy for everyone every time some investors knock on their door.

I could be the "btw I use Arch" guy here (and btw, I've been using Arch on most of my servers, desktops and embedded devices for more than a decade now), but even though I'm very happy with it, I also acknowledge that it's not for everyone. Not everyone wants/needs a rolling release system that requires frequent updates and runs the freshest versions of all the software. If you're ok with the rolling release approach (and, in all honesty, I don't think that a rolling release model is that bad on servers), and you want to run Linux on servers that are used by paying customers, then you may also go for CentOS Stream directly - you'll have to deal with newer/more unstable packages than RHEL, but most of the stuff released as RPM for enterprise solutions will still work.

Debian is of course another popular option, but I gave up on it a while ago, as I was tired of using packages that were 3-4 years old, and risk breaking the system with backports. Sure, you can use snap/flatpack/Docker images on it, but that you could also do on any other Linux distro, right? But if you're a guy who needs to run RHEL-like systems (e.g. stick to a distro with packages that barely change for a couple of years), then you may probably go for Debian.

What are you folks running on your servers? Are the recent decisions from RedHat/Canonical pushing you to consider changing? Are you one of those who's currently considering screwing Linux for *BSD for good?

The only question I have after a shallow read of the #ATProtocol (the "specs" behind #Bluesky) is: why?

Why an entirely different protocol? Why not just an extension on top of #ActivityPub?

They wanted a bit more centralized control on top of a decentralized network - not a bunch of geeks running their own servers in the bedroom, but a "marketplace of companies", whatever that means, in charge of distribution and moderation? Fair enough, they are businesses that needs to make money after all.

But then why didn't they go for an extension that simply supports self-certifying data (signed using whatever centralized chain of certificates they want) on top of ActivityPub? The Fediverse and Twitter (or whatever they are going to call whatever they're trying to build now) could have immediately tapped into one another. They could have simply shown posts not certified by their chain as "unauthenticated", and we would have had almost 100% compatibility from day one.

Why a whole new protocol? Why build a new ActivityPub that does the same things as ActivityPub, but just with a layer of self-certification on top? Why call things differently and break compatibility, instead of extending the open standards?

It takes time and resources to build a new protocol, validate it, get developers to adopt it and build stuff on top of it, and add yet more layers of bridges and mappers to/from what exists already. And yes, I know that the alternative approach that I am promoting is exactly that "embrace+extend" that we all know how it ends up. And I'm also aware that are valid arguments to keep the streams separate. But I'm not sure if the best alternative is to let the rest of the world build their own brand of our open protocols (R&D these two balls: these guys are simply repackaging a decade of work on open standards and they are trying to profit from it), while we either keep digging our niches or play a catch-up game with bridges and mappers to the outside world.

https://atproto.com/docs

#GRB221009A, the most powerful gamma-ray burst ever detected to date, was recorded just three days ago.

It most likely originated in a galaxy 2.4 billion light-years away, therefore, even if it got to our planet only now, the supernova that originated it most likely already turned into a black hole when multicellular life wasn't even a thing on Earth.

At the same time of the burst, #NASA confirmed a disturbance in the ionosphere on the band between 19.6 and 63.9 kHz reported by 8 radio stations around the world.

Let that sink in from a moment. A supernova that exploded in a distant galaxy in a time when jellyfishes weren't around yet, was so powerful that its echo created ripples in our radio communications today.

Now imagine what could have happened to planets that were much closer to the explosion.

These are the kind of things that really help me put everything in perspective.

https://gcn.gsfc.nasa.gov/gcn3/32744.gcn3

@codewiz I don't think that the term parasite is an exaggeration. If you look at the history of Microsoft, from Gates' initial letter to the hobbyists, to the BSD code that they got into the NT kernel, to their multiple attempts to cannibalize JavaScript and the web's open standards in order to ensure IE dominance, to that great theft of FLOSS code that is the WSL, to the acquisition of Github just to use years of open-source projects just to train their models and offer products that no other competitors can offer, you'll see the history of a company modeled around the embrace-extend-exterminate principle, a company that greatly benefits from open code and open standards, and used them to get where they are today, but doesn't want to give anything back - on the contrary, they make their best to ensure the demise of the projects and tools that ensured their success.

To me, this is the textbook definition of a parasite: it can't exist without a host environment feeding it, but at the same time it contributes to the destruction of that environment.

UEFI is a technological disaster that has stifled innovation and competition for years just to let Microsoft hold a monopolistic position on x86 hardware, while providing basically ZERO added value when it comes to additional user security.

And #Microsoft requiring Lenovo tablets to ONLY boot Windows proves that, behind of all their "Microsoft ♥ open-source" marketing bullshit, Microsoft remains a deeply evil and parasitic company with no added value to the world, a filthy wolf disguised as a sheep, and a cancer that needs to be eradicated from this planet.

Seriously Microsoft, you first ask the FLOSS community to forgive you for being absolute greedy monopolistic jerks for three decades, you promised that you were a different company now, you even managed to convince many of us that you've changed, and then you come up with this shit? How can Microsoft employees even sleep at night without feeling guilty for working for one of the greatest parasites in the tech world?

https://mjg59.dreamwidth.org/60248.html

#Twitter is launching #Bluesky, a.k.a. Authenticated Data Experiment.

It'll lead to a more decentralized social network, a more open-source platform, federation across instances, shared protocols, and users in charge of their own "Personal Data Repositories" that they can easily share and move around.

In other words, Twitter is reinventing #ActivityPub, and ignoring years of progress already made on the protocols and infrastructures of the #Fediverse.

I'm really wondering what's the point. If you're a company like Twitter, that has already been struggling to turn profitable for the past decade, what's the point of pouring even more resources into rebuilding ActivityPub from scratch instead of reusing what's already available?

https://www.theverge.com/2022/5/4/23057473/twitter-bluesky-adx-release-open-source-decentralized-social-network