Self-hosting your own media considered harmful according to YouTube

I don't think this is a job that requires an LLM but if an LLM took the order, made the plan to go through the relevant data(bases|lakes|platforms) and triggered the warnings, etc. I'd be very impressed.

I do agree here, but sometimes (let's say 10% of the time? less?) the squeeze does not continue -- see Apple. Perplexity/ChatGPT vs Google search right now.

> The rest of us, the unwashed users of the platform, do not hear about it or act upon it en masse. We'll occasionally see a post like this on HN or Reddit, shake our heads and call it a shame, there need to be alternatives and so on, then go right back into those platforms and forget that something happened but a few months later

Yup, wish I could add "- posted from Chrome browser" to my own response here but I use Firefox. I'm still going to watch YouTube.

I think the thing that might bring hope is that Google/YouTube doesn't actualy own the new paradigm of AI -- I can very much imagine a world where people just ask for videos/scroll through them, and YouTube isn't the site they do it on (in fact they don't do it on a "site", per say).

But then again, that's really calling for the death/dramatic reduction of the open/surfable internet. Is that what it takes?

> it's like Pixelfed but more ergonomic for video.

This is a huge problem IMO. Just like Mastodon/Bluesky (which seems to be working recently) and all these other things, the tech and experience need to be SUPER easy. I mean as-easy-or-easier than YouTube, etc for people to switch en masse.

> So the next stage, ideally, would be everyone kinda sharing hosting responsibilities, and if you like a creator, you just follow them. This has the benefit of possibly caching/mirroring all the videos, too. My Fediverse server was chewing through disk, one of the reasons I shut it down - but I was following 1400 news and journalist accounts, plus my ~100 or so gang of idiots. I was nearing 1TB on disk after about 16 months on my essentially single user instance.

Yeah the problem is people won't do this/can't be expected to do this, unless it's drop dead easy.

Really appreciate hearing about your point on the scaling curve for this tech though, clearly the tech has come really far, that sounds like much more than the average person and "only" 1TB and a single server is quite nice.

The best ever approach I have seen to this is PopcornTime. It took the world by storm (and IIRC people still use it/ it still exists in some form, they're just lower profile now), and it worked better the more people used it, because torrents (aka, the technology being a mature, perfect match for the usecase).

> I exported my follows and moved to an acquaintance's server and imported, the owner doesn't even blink. Who knows what they've got going for storage. > > Anyhow if you don't need to follow 1500 people, this becomes tractable. If it gets popular, someone will post how to cron the multimedia stuff to compress it as it ages, moving it to cold storage, whatever.

I could see this working if that acquaintance got paid for this. Tying money as an incentive to things is sometimes bad/not what you want, but having people think of computers and compute as an asset/tool for them to use is a step in the right direction IMO.

I'm not a crypto person (kind of wish I was, 10 years ago), but Filecoin was really interesting originally to me because it just made sense. The marketplace of data storage seems like something that could be easily democratized in this way (no need for the crypto bits, but the ease of payment was a legitimate use IMO).