Eventually I think sites will customize every URL for each user.
TikTok is quite sneaky. Sharing from their mobile app, you get: https://www.tiktok.com/t/[9digitCode]/
Only by opening the URL in a browser will you see: https://www.tiktok.com/@[user]/video/[19numerals]?_t=[alphanumericIdentifier]&_r=1…which can be sanitized.
Here’s how they took it a step further too: YCombinator.
Same with Reddit, FB Messenger, Instagram, TikTok… Some of them are harder to spot, like how Reddit now goes reddit.com/r/example/s/8913y4h93
Would be nice if social networks and messengers would automatically strip these parameters.
I started using URLCheck on Android and SO MANY links have some kind of tracker that you can drop and not lose any functionality. Things like Signal (and even Lemmy/Mastodon) could do something similar and throw up a little warning when it encounters a known tracker, then offer to clean the URL for you.
Another advantage is that the clean URLs are a lot more descriptive
I’m using FairEmail on Android. When tapping a link, thr app detects tracking parameters and offers to remove them. I really like that feature and wish other apps would offer something similar.
That was the first place I noticed it, thought it was really smart of them, someone would send me a meme or whatever and it would show their account at the top. Was impressed that they generate so many links, now they can see who knows who so easily
Exactly. A website has to download ALL the HTML every time. Sure, it can put all that in a JavaScript file and cache it but it has to be built each time. With an app, you (the devs) get to choose what to load, and it’s just usually a few simple things each time instead of constantly running a script.
Using Lemmy as a web app really sucked. Having an actual app with actual integration to a robust UI works.
Plus as an app developer you get to go through the user's contacts and files. Having an actual app locks you and allows you to be the product the app owners sell. Nothing else and certainly nothing of value for 99% of the apps out there.
That is 100% wrong. Did you read the tweets or even look at a single YouTube URL?
youtube.com/watch?v=FOO&si=BAR would be shortened to just youtu.be/FOO?si=BAR
The link to other people’s account is in the &si=BAR part. Probably standing for “share ID” or “source ID”or something. The shortened link is just the same as the long one with watch?v=FOO being included in the URL instead of the parameters.
I looked into this a bit more and here is the summary: This is meant to show off a candidate kernel feature that allows for running different schedulers in userland.
Task scheduling has become much more complex as CPUs have grown in size and have had new developments in architecture, so the need to develop more complex and robust schedulers is steadily rising.
The kernel feature is meant to lower the barrier of entry for anyone who wants to try getting into schedulers, as well as enable quicker development iteration, by removing the need to completely recompile the linux kernel every time you want to test your code.
Not quite running in userspace. To the best of my own understanding:
The new kernel feature is to allow writing schedulers in eBPF, a “language” the kernel runs in kernelspace that is heavily restricted.
For example, all eBPF programs must complete in bounded time, and the kernel’s static checker must be able to verify that before the program can even begin executing. eBPF is a rare language that is not touring complete.
“For scx_simple, suspending the scheduler process doesn’t affect scheduling behavior because all that the userspace component does is print statistics. This doesn’t hold for all schedulers.”
So, it may be that eBPF also makes it easier to write a truly userspace scheduler, but that’s not the primary purpose, and it’s not what is being done with scx_simple.
In all seriousness, streaming could be profitable for these companies if they just didn’t layer bullshit on top of bullshit on the backend. Dozens of k8s clusters, hundreds of stupid microservices, and engineering team to manage it all…it’s insane. This is BEFORE all the idiotic choices they make greenlighting new content that costs a ton of money.
If they just wanted to make money on streaming, they could without the extra bullshit, and just charged people for sending content across the wire. This is just not what any of these companies are doing anymore, and we’re paying for it.
Sounds like you could save them buckets of money by redesigning their whole backend. They’d probably pay you idiot money to do so, in terms of seven figures per year at least for the cost and performance enhancements you seem to think are easily achievable.
So you are tryimg to tell me that the operating costs of the CDN are the big reason they need money, not the fact they throw billions at crappy content productions? And that the little performance you would gain by stripping k8s would make a difference, especially wrt the huge additional administration effort and lack of automation this would introduce? Nah man, sorry but you are wrong here in many ways.
No, I’m saying that Netflix WAS a streaming service only, and then created a scenario where they constantly need more money and can’t support their own business practices by incurring an insane amount of debt by complicating the platform and spending a small country’s GDP in creating new content. They WERE making money, and then went off the rails and now need to constantly raise prices.
I’m not sure what you mean about the CDN, but aside from Amazon and Google, none of these streaming services are caching at edge for the actual video libraries. Instead, Netflix specifically has tiered replication of their entire library in various AZs around the world tailored to each regions most popular content. You can see this at work by watching your network traffic hit different endpoints for something like a popular new title versus an old and rarely watched title.
Netflix is still making money, and the cost of their tech is utterly dwarfed by the cost of creating and licensing content, so I’m not sure what your point is.
Sorry to burst your bubble, but Netflix is $13b in debt. That is THIRTEEN BILLION USD IN DEBT. It’s a fact. The streaming arm of the company was the only profitable part for about two years when they first started it.
As far as their tech goes, and you can Google if you want, they would make more money per customer if they dropped the other bullshit, and just focused on streaming content.
I can’t speak to k8s but there are reasons you need clusters to handle absurd amounts of data with high uptime and high networking performance.
The reason the streamers are creating their own content is licensing breakdowns and copyright law. I don’t know why everyone thinks they can out Netflix the Netflix streaming service, but they also think they’ll make more money trying to do it. It’s like Amazon deciding to somehow make their delivery system better than UPS. Or cheaper.
I think it either calls into question the supposed economies of scale and core business competency theories or they’re not doing it to save money / make money at all.
I’m not here to shoot down your comment in response, but you are confusing a few different things here. My point was just to demonstrate how the actual cost for streaming from a provider when you throw all the extra junk they do on top of it. You can have very reliable, distributed systems without all that mess, they just choose not to go that route. Selling a user subscription is just the beginning down a rabbit hole of upsells, price hikes, content lures, marketing gimmicks, data capture and sale programs, Ads (for a service you already pay for in most cases)…it’s expansive. Each one of those things has a team behind it making decent money.
If they just wanted to stream things they could with much less cost and effort, AND make money doing it. It’s been done before. They all just choose to go the route of squeezing their audience for every last ounce of possible monetization they can, which costs a ton of money.
Do you really think that k8s, which is free (unless you buy a support contract) is less easy to manage at that scale than, what - a bunch of VMs with Docker installed? PHP and a NAS? Is that REALLY the cost sink you think it is?
Brother, what works in your homelab will NOT survive Netflix traffic levels.
And streamers are raising prices to get rich. Not because “k8s hard” lol
Hey, friend. No need to get mad. I’ve worked on these exact systems I’m describing for years, which is how I know these things. They suck. It’s not about k8s itself, you’re focused on the wrong part of what I’m saying.
You want to just stream files to people cheaply and quickly, these companies don’t do that anymore, and that’s the choice these companies have made. They’d rather scale their engineering resources to the microservices that nickle and dime the shit out of everyone. That’s just the facts.
Ok but marketing teams and micro transactions teams to get rich have fuck all to do with k8s or clusters to run lots of data quickly over the network.
Maybe you know some magic way to provide high availability for petabytes of data pushed at hundreds of Gigabits per second to millions of simultaneous connections. But I don’t. And I work with large data volumes for research that still are a fraction of Netflix data and throughput and we’ve hit the limit of single high performance storage devices necessitating a ceph cluster for storage and availability.
youtube.com/watch?v=FOO&si=BAR would be shortened to just youtu.be/FOO?si=BAR
The link to other people’s account is in the &si=BAR part. Probably standing for “share ID” or “source ID”or something. The shortened link is just the same as the long one with watch?v=FOO being included in the URL instead of the parameters.
If you’re on Android use grayjay, if you’re on Linux use freetube. You can follow channels on both of those.
If using grayjay and you can afford it please do pay for the license; you’re not technically required; it’s based on the honor system, but it helps the developers at FUTO work on it and it helps them donate to other FOSS projects.
If you use freetube please donate to them, even a dollar, the developers will greatly appreciate it.
Wow, that’s a huge difference. I wonder if it’s just a niche case that fixes Terraria specifically, it would be nice to see more benchmarks on other games.
The Video-Industry behaves just like the Music-Industry. They constantly throw dirt at the wall and look for what is sticking to it. If they find something that sticks, they throw further dirt at this point until the whole blob falls down and they start over. The saw what sticks to customers when it comes to price-hikes and now throw everything after this to see how long it will stick. Vote with your wallet.
Don’t know if any other browsers do it, but Firefox for desktop added an option when right clicking links to copy without URL tracker. I don’t know if it works on yt links, but it’s definitely a step in the right direction.
Heads up, I had to remove this extension on my browser because some websites would get stuck in a redirect loop because it’d remove the tracking stuff it’d use in a redirect chain. Took me months to figure out what was causing it
I made a PWA that can quickly remove tracking variables called Link Cleaner. If you install it through Chrome or another Chromium browser on Android, it shows up as a share target, so you can share links to Link Cleaner and then share again to the intended target.
nitter.net
Top