Did you try it? What kinds of results does it come up with? Usually if I tell AI to assume roles like those, it comes up with really over-the-top fiction stories. And what kind of framework and model do people use for pipelines like this one?

I’m not sure if this translates to the content creators. There’s many of them whom I really like to watch who do (or did) Youtube as a business model. Tom Scott being one example or Derek Muller (Veritasium). I’m subscribed to many more. Simplicissimus and their yet better second channel (in German). We wouldn’t have those without monetization. The platform of course went shit over time. Fortunately my Ad blocker still works and thanks to Sponsorblock my experience is fairly alright… But personally - I’m split on this question. We had quite the amount of entertainment before monetization but I think a large amount of quality content also arrived after that, and because of it. Those people would be working some office job today if it wasn’t to Youtube. And I (and the world) would miss out… On the other hand we got MrBeast, a lot of fake cooking videos…

One thing I did is connect to the smart home (Home Assistant) and the NAS running at home. Some internet service providers don’t provide proper IPv4 addresses any more so IPv6 is the most convenient way to connect. This doesn’t require a VPN provider, though.

As far as I know it uses the B.A.T.M.A.N. mesh protocol. On a channel within the regular 2.4GHz wifi spectrum. So no license needed unless it collides with laws for point-to-point beams. All people communicating to each other obviously need to agree on a channel. It comes with some hierarchy where I’m at. There are local chapters who make up some config and who also operate nodes and exit nodes into the internet. These are necessary because Germany has stupid laws.

There’s Freifunk as well!

Don’t MoE models just load into memory as every other model and it’s just that they pick a subset of numbers to multiply by each step so they’re faster? That’d make me think it’d need somewhere around 80GB of memory at 8bit or 160GB at full precision, or something like 50GB at the average llama.cpp Q4_K_M…

I’m only tinkering with the 8B variant, so speed is alright. I hadn’t noticed yet. But yes, seems English to German leads to weird results. German to English seems to be fine, though. At least for the 2 texts I put in.

I’m running the 8B version with KoboldCPP and pretty much the default settings of the Min-P sampler. That tends to work very well with almost all other models and I’m not aware of any specific recommendations for Apertus…

Did someone try it? What’s your experience?

I think I like it. But, I’m having massive issues with it becoming very repetetive after a while. And I’m not sure if it’s the model or my sampler settings.

Yes. I think several clients have open feature requests. The Stalwart documentation has a list of projects. There is one command line client as of now. But I’m not switching to a cli mail client or proprietary software, so I’ve postponed it. We’ll see where this is going.

I welcome these modernization attempts. Though in theory I’d love to see someone revamp email in its entirety, add encryption, signatures, chat and crack down on spam and phishing. Not sure if that’s ever going to happen, but that’d be great, too.

Yes. And I feel a technical solution like watermarking would make things so much easier… We could just decide whether we want to read AI or human text.

Unfortunately JMAP isn’t supported (yet) by a lot of email clients. I don’t think there’s a good open-source email suite for computers available… But I’ve tried Stalwart as well and it’s really sleek and seems to come with good defaults.

Lol. I think FIFA demand a pile of money from people integrating the live streams into their platforms. We should start the fundraiser soon.

And I think I’m going to boycott the world cup 2026, so make this the women’s world cup in 2027.

Even if this is the piracy community… I think this holds true in several aspects of life. Availability of information, Freedom of speech, political freedom, privacy… Unless we fight for freedom, it might just go away.

I don’t think it’s allowed to request or directly recommend or link to pirated content in this community. Can’t be too specific. See the rules in the sidebar.

Yes. And I’ve seen people do really weird (and worse) stuff, like re-post questions. And if you’re unaware of that, you’re bound to take 10mins out of your day to answer the product or Linux or life question and later find out you were ripped off. At least that’s how it feels to me if people fabricate that kind of activity. And it kills the mood to write comments for a while.

Alright. And for your information, the Peertube function is a bit broken. I think the Peertube developers did their best. But Youtube has a lot of datacenter IP address ranges blocked. And they do rate-limiting and force people to sign in after downloading a few videos. Plus yt-dlp (which it relies upon) and Youtube are playing this cat and mouse game… So it’s disabled on most instances because it doesn’t really work. I was able to make it work on my instance, but I had to jump through several hoops. Configure a SOCKS proxy and tunnel it over my home, residential internet connection. And I think I transferred my login cookies because some videos would be age restricted.

PeerTube has that built in. You can set up a channel and have it import or mirror a Youtube channel. For Lemmy there’s several bots and scripts. As other people said that’s what lemmit.online is about.

Be a bit careful when rolling this out. Several people don’t like it. They’ve left Reddit for a reason and this is drowning them in bot activity. And usually these posts are low engagement, Reddit users can’t see the comments, so you’re not getting a lot of answers. I think it’s good practice how we here have separated that to dedicated instances, so users can just have genuine conversations everywhere else.

Mostly I can’t be bothered, or Roblox won’t run, or some stereotypes about Linux being difficult.

Thanks! Learned something today. Last time I opened port 53 to the public it didn’t take long and I was sending out several Megabits per second in DNS traffic. Constantly. Mostly querying the same few things. But I guess I had it the wrong way round and that wasn’t the target. Or I’ve seen a different attack type… Guess I can now try again with the new knowledge.