If you find those few instances that have those few uploaders that interest you, the content can be good.
However it does certainly suffer from the fact that all federated instances can't pull in videos from other platforms in some reproducible way (so that all instances would be able to serve the same content).
I'm sure some YouTubers would allow mirrors if they didn't have to pick an instance and keep it running.
Servers can communicate their preference in terms of CAs they want. But the UX in browsers is unbelievably horrible for no good reason.
Not only is it difficult for an user to make a proper selection, it's also hard to fix a wrong one. The error pages are also terrible. There's no way for the site owner to request that when the navigation to the (auth) page fails, redirect back. Nope, no way to do error handling without some really clever iframe stuff and even then it's way too opaque.
I blocked HTTP connections from my local network years ago and you wouldn't believe how many driver installers and auto-updaters break. One should never trust a HW vendor's (auto-)update implementation.
> [...] it seems likely they also have enough to maliciously chug the hardware sufficiently to degrade capacity over time and otherwise impact system integrity. I hate the thought of some random website writing and overwriting random bytes in a tight loop in the background while I'm browsing elsewhere to find the cause of my slow disk subsystem.
Absolutely. Things like IndexedDB get fsynced super frequently. There's no way to tell Chrome that some web apps do not need to make it do the physical disk this often.
It's really difficult to reliably separate temporary and persistent browser storage. I tried at some point to reduce HDD noise. But given how neither Firefox or Chrome properly follow the XDG spec, it did not yield the results I wanted without a lot of handcrafted mounts.
In the end I'd guess you can also use some aspects of persistent storage to achieve similar results, even if the rest is actually tmpfs/RAM.
> As an online store you don't want to ask customers to manually input a payment reference into a SEPA transfer. It's all about ease of use (and safety).
How? With a SEPA transfer I can actually see who I'm paying. With a CC or equivalent it's a lottery.
LLMs managing the "coloring book" equivalent of something is not bullish for the "art" version of something.
The intent for most CTFs is to provide a meaningful challenge that concerns a single topic without introducing noise that wastes time. Of course a training exercise is easier to complete for an LLM.
The more you obfuscate a topic against LLMs the lower the educational value of a challenge.
The only things that works is novelty and obscurity. LLMs still suck with things mentioned in the footnotes of datasheets and manuals, things that deviate in subtle ways, unique constructions that alter something very very common. It's hard for LLMs to avoid common pitfalls in terms of making assumptions, while staying on track.
However it does certainly suffer from the fact that all federated instances can't pull in videos from other platforms in some reproducible way (so that all instances would be able to serve the same content).
I'm sure some YouTubers would allow mirrors if they didn't have to pick an instance and keep it running.