Still works for me with a combination of Firefox, uBlock and a VPN. I assume sharing an IP with thousands of other people screws up their detection algorithm.
Every Invidious or Piped instance could easily implement IPv6 rotation, the smart-ipv6-rotator is fully FOSS and easy to set up. Invidious and Piped both have official documentation on how to set it up: docs.invidious.io/ipv6-rotator/
docs.piped.video/docs/self-hosting/#ipv6-rotator-…
Me too, haha. This is what happens when you treat arguments like chess pieces and not, you know, reason. I don’t believe this guy has any idea what he just said. I also don’t think he cares.
The image doesn’t quite work because youtube needs a MASSIVE solution that works, scales, doesn’t fuck up their infrastructure and on and on and on.
Meanwhile on the user front, all your doing essentially is just skipping parts of a video, that will always be infinitely cheaper easier to do. Challenging for sure, but the solution can be small.
Even something as stupid as delaying your video start by 1 min, pre buffering then skipping ads. It’s brute force and barbaric but the point is that Google can’t do shit against that.
My ultimate vision is AI that preloads videos you want and detects ads / sponsor segments and just skips them / cuts them out on your device.
Another way to think about it… YouTube has huge amounts of compute resources, but per user it’s an extremely small amount. Your phone has orders of magnitude more power to dedicate to you than YouTube does. Collectively, we have more processing power than YouTube.
I truly hope the new planet of the apes franchise is building to an actual planet of the apes remake. Time traveling astronauts and shattered statue of liberty. I feel like they’re already sowing the seeds for human mutants living underground worshipping an unexploded atomic bomb. Man those movies were weird.
Back in the day, when I installed my very first Linux OS, I had a wireless stick from Netgear. Wireless Drivers back then were abysmal, so I had to compile them from source (literally 15 mins after seeing a TTY for the first time). After I had found out how build-dependencies and such worked somehow and ./configure completed successfully for the first time, the script ended with the epic line:
Ah, I had one of those wireless sticks from Netgear as well, probably a different model but still a royal pain to get it working.
Luckily ndiswrapper has become a thing of the past nowadays.
Yes, it would depend on your flatpack usage. For me I only have like 5 programs compiled from source and one flatpack (bottles) because of the sandboxing
That’s not good. It breaks the system as there isn’t any change control with that unless your using something like Gentoo. Get your packages from the package manager.
Because it’s always so easy to compile everything you need from source! Just make sure to download, compile and install the dependencies first as well. Oh, and the dependencies’ dependencies. And the ones from them. And so on. Unless you’re lucky enough that there are already packaged dependencies available for you. Don’t know how to compile? No problem, just read the documentation. You can be absolutely 1000000% dead serious sure that everything you need to know is documented and extremely super duper easy to understand if you don’t know the source code or barely know how to code at all. And if not, maybe you can find the bits of information on the respective Discord server. It will probably be also very intuitive to know which build options you have to set in which way and which ones even exist. And that without causing conflicts with other packages you need to compile. Still got got problems with compiling? EZ, just open a bunch of issues on the respective GitHub pages. (If present. Otherwise, try to find another way to contact devs and get support, Discord for example.) Maybe, about six months later you’re lucky to get a response. And if not, don’t worry. Some will tell you, you should RTFM or are an idiot. Some will just close the issue because your platform isn’t supported anyway. Then you know, what you did wrong. Also don’t mind if your issue gets ignored.
If you finally managed to compile everything from source, congratulations! Now run the program and test if everything is working. If it’s not or if it is crashing, don’t worry! In developed and civilised countries you can just buy a shotgun and blast your own head away to end this suffering.
I just complie from source some lightweight programs that are too niche for repositories. I am in no way advocating for full source compilation of every program in your system, that’s a security and usage nightmare. Flatpack does have its use for sandboxing an environment. I personally use it for windows applications in bottles.
My workflow always definitely includes multiple weeks to debug random issues with building the tools I need to use. Totally a scalable and good solution to dump this work on the end user.
This doesn’t scale. If I have a bug and my package has about two dozen dependencies which can all be different versions, and the developer can’t reproduce my bug, I’m just screwed. Developers don’t have the time and resources to chase down a bug that depends on build time variables.
I think this is a picture before the most recent expansion. (They saw this picture and said “hmm not wide enough, too congested.”)
In the normal parts:
2 express/toll/HOV/carpool lanes
5 regular highway lanes
3 feeder lanes (in Texas, the highways tend to have “feeders” or “service roads” or “frontage roads” that run parallel to highways so that people can exit and enter, turn onto intersecting roads, and access local businesses, and Houston calls them “feeders”).
That’s 10 in each direction. But at any given time there might be merge lanes between the express and the regular lanes, between the highway and the feeder, or between the feeder and a turn lane. So at the widest point, around the major freeway intersection with another huge toll highway, they bump it up to one more of each type of lane, for 13 lanes in each direction.
There’s also a fair debate about whether the feeder lanes should count. After all, they have traffic lights and intersections to deal with. But on the other hand, driving on them is necessary to get on and off the highway lanes, so in a sense it’s part of the same highway.
Also of note this is part of the evacuation system for hurricanes. All lanes are outbown in an evacuation.
But it also kinda hides some of the crappyness. it was supposed to be on top of a giant underground flood diversion system which is a huge problem for Houston as you may have heard. Also the center was supposed to be for a commuter rail but we couldn’t get the bond passed so that might happen someday…
Yeah, that’s the fun part. Hooking into some auto-update mechanism would be useful to me.
But my stuff is mostly in the scratching-my-own-itch stage, so setting up a FlatHub account, Flatpak metadata, sandbox rules, probably an icon and screenshots and whatnot, and automating the build+releases, just to get auto-updates, yeah… no.
I could code a whole nother project in the time that would take.
Well, if you have any form of build script, makefile, or CI, then you can easily shove that into a flatpak-builder manifest and push the build repo anywhere you want. The default OSTree repository format can be served from any old webserver or S3 bucket after all.
I’ve done this for personal projects many times, since it’s a ridiculously easy way to get scalable distribution and automatic updates in place.
Hmm, okay, that doesn’t sound too bad.
Does the sandboxing get into the way much? Can a user tell it to poke a hole into the sandbox, to use some specific folder, for example?
I think, my real problem is that I don’t actually use Flatpak for any software I have installed. 😅
I’m not opposed to using Flatpak, but I disabled Flathub pretty quickly on my distro’s software store thingamabob, when I accidentally installed some proprietary software from it. Fuck that shit, no matter how much sandboxing I get.
In regards to sandboxing, it only gets as far in the way as you ask it to. For applications that you’re not planning on putting on FlatHub anyway you can be just as open as you want to be, i.e. just adding / - or host as it’s called - as read-write to the app. (OpenMW still does that as we had some issues with the data extraction for original Morrowind install media)
If you do want to sandbox though, users are able to poke just as many holes as they want - or add their own restrictions atop whatever sandboxing you set up for the application. Flatpak itself has the flatpak override tool for this, or there’s graphical UIs like flatseal and the KDE control center module…
So, like, dumb question. People here assumed that I mean AppImages, whereas I actually meant just a statically linked binary. Is that really the only reason why AppImage exists? So, that dynamically linked applications can be distributed like statically linked ones?
This is the problem those tools try to solve. They package everything else upon which software might depend that can’t simply be linked into a single binary.
Ah, interesting. So, it’s different from just statically linking against the latest driver lib every 6-12 months, because the Flatpak runtime gives you a bit of a guarantee that there won’t be breaking changes in the meantime.
Bingo, and if the latest mesa breaks your app for example, you can target an older one until it’s fixed instead of end users having to fuck around downgrading system packages
Yeah, alright, packaging assets makes sense. I’ve always been fine with just a .tar.gz, but having it be a singular file without compression is cool.
I guess, since AppImage emulates a filesystem, you can also have your application logic load the assets from the same path as if the assets were installed on the OS, so that’s also cool.
lemmy.world
Oldest