RE: https://livellosegreto.it/@scartafaccio/117325989088505936

Scartafaccio organizza eventi belli in quel della Valceresio e io non ne sapevo niente…

Tanta invidia per chi era presente! Un ottimo modo per rivitalizzare il territorio arcisatese ⚡️

#Arcisate #valceresio #laghidinchiostro

@hacrods I think you need to answer one question: do you want to have access to your services in your homelab from the outside world? Or do you want the outside world to have acess to the services in your home lab?

In my case I decided I eventually want to access from the outside, because the latter option is always risky and pretty surely resource intensive when bots decided it's something worthy to have. That's how I went to use Tailscale first, and moving to headscale soon enough.

@hollie This is something I’ve grappled with myself and ended up withdrawing from more personal blogging, so I don’t know if I have a great answer for you 😬

I do keep feeling like LiveJournal’s friends-only posting was a model that worked well for me, and therefore that I wouldn’t mind using Dreamwidth… if anyone I knew used Dreamwidth. (Well, my sister uses it, but I don’t think she wants to mix her smutty fanfic internetting with "keeping up with family" internetting, lol.) I still feel like Fediverse-compatible LJ is a missed niche.

I feel like an option that emails people on your subscriber list is the most accessible, but I personally don’t love email newsletters 😂 Ghost’s private RSS feeds are cool, but yes, cost. If your current blogging platform allows for password-protected posts, that seems easy so long as you don’t mind communicating passwords.

Just my thoughts, anyway. Probably they don’t help you settle on anything at all 😂

@stefano
I understand you having tried to run it "natively" on Linux. I reverted to podman aka docker (running in a VM 😉 ). Wouldn't this be also a solution in your environment?

@grahamperrin OK. I understand your point.

The rule exists partly because we've had an almost endless stream of bots, malicious users, and similar accounts lately, to the point that I've had to restrict registrations even further. The intention is not to police opinions, but to prevent this kind of behaviour from making the Cafe a worse place for the people who are here.

There is also so much money circulating around AI right now that some people, or bots, are actively trying to get into communities like ours and subtly promote their paid services. Sometimes it is obvious, sometimes much less so. That is one of the concrete reasons that led to this rule in the first place.

And just to be clear about the report you mentioned: threatening to stab someone in the face is obviously not acceptable here. That user was not on BSD Cafe, but on another instance, so I cannot enforce our rules there. What I can do, and what I did, is suspend that user's interactions with our server. Violence is not accepted here, whether it comes from inside BSD Cafe or from outside.

There are also other circumstances, which I won't go into again, that made me think that asking people not to *promote* AI is a reasonable way to protect the atmosphere of this community.

That doesn't mean people cannot talk about AI, disagree with the rule, criticise it, discuss its benefits, or have completely different opinions from mine.

What I want to avoid is every discussion turning into another occasion to repeat that AI will save the world, computing, humanity, or that replacing human skills with it is simply inevitable and everyone should accept it.

Until these technologies can be used in ways that are genuinely ethical, safe, transparent, and autonomous, I don't see promoting their current ecosystem as very different from promoting closed technologies and dependencies that can potentially cause harm.

So, for me, the rule is not against the technology itself. It is about how that technology is currently being developed, promoted, and used, and especially about the idea that it should be pushed everywhere, often at the expense of human skills and judgement.

I would genuinely be sorry to see you leave, but if this is not compatible with what you want from an instance, then of course I respect your choice.

@mindpersephone here we do it the other way around:
When you SSH to a server the authorized_keys set the git env vars for you to set your identity
Which means that on shared accounts you always have your identity and not www-data ;)

I love the idea of multilevel defaults (global then env then localconf...)