Part of being a maintainer is helping to onboard new contributors, this is why many projects have a tag for “good first issue”. Teaching people how to use the library/tool is part of that.

The top comment on that site says you can already disable all of these features from within the settings. Is this true?

Kinda weird experience to be reading textual descriptions of memes and having to reconstruct them in my head. They had enough to say to not need to pad out their word count that way.

I was going to quote this part as well - nice bit of malicious compliance.

Interesting stuff.

I assume this also needs high precision timers, which are disabled in many cases. I’m also assuming that the reason the file needs to be quite so large is to ensure that caching doesn’t ruin the measurements.

I read this disgraceful comment yesterday, and I’ve dug through my history to reply to it today.

@rako, this unacceptable. Let’s remove the mention of AI to see if you can get some perspective… Imagine this exchange:

P1: I’ve been cooking for the homeless but it’s taking up a lot of my time and energy. Is it ok to use shop-bought meals?

P2: You being weary of cooking is belittling the homeless! People like you are what’s wrong with society.

I hope you can agree that this is unfair, and unhinged. It’s also not mischaracterising what you wrote.

@Gonzako you don’t seem to have minded rako trying to shame you, but they were way out of line.

Are you referring to the photo of a baby eating meat from a bone?

I was watching a video the other day that was showing WiFi sensing using TOMMY which uses channel state information (CSI) - something this article describes as the previous approach. That was already quite impressive, although not nearly as powerful as this teaser hints at.

We’ll have to wait for the hang on, the Taipei conference was last year so this is old news. Here’s the paper: https://publikationen.bibliothek.kit.edu/1000185756

Scotland Yard are saying that unless they’re allowed to spend £50m on Palantir, they’ll have to reduce officer numbers? I can’t see their logic.

Yup: https://deadeclipse666.blogspot.com/2026/04/public-disclosure.html

This Chaotic Eclipse/Nightmare Eclipse is the same one whose opening post read:

I never wanted to reopen a blog and a new github account to drop code…

But someone violated our agreement and left me homeless with nothing. They knew this will happen and they still stabbed me in the back anyways, this is their decision not mine.

I’m guessing there’s plenty more to come.

Kinda funny that they’re targeting Microsoft and yet using GitHub to share the PoCs.

My first thought was that Benn Jordan did a great bit of video journalism on this, but it’s already linked from the article, although without any other mention of it.

Odd that piefed.social gets counted as a “Mastodon-ish” server instead of being counted with the Lemmy stats.

Isn’t picts-rs just something that runs on one server and just handles converting and resizing images? The new remote media storage stuff could run that, but the point is to allow more centralised servers so not every instance needs to individually store, process and transmit media.

Not just media, things like searching the Fediverse can be handled by a dedicated server, shared (or not) by multiple servers. Also spam detection, link preview generation, etc.

Zaatar is both the name of a herb (Hyssopus officinalis) and a spice mix, containing the herb along with sumac, salt, etc. Mostly references to zaatar are about the mix as that’s what normal people actually use.

The herb is quite strong, is like a slightly bitter mint, and is also used in Chartreuse and absinthe.

We will implement a new Fediverse Auxiliary Service Provider (FASP) that will allow sharing storage and media processing between servers.

This is pretty big too, as the cost and legal risks of hosting this user content is high. They’ve clearly thought about the media moderation problems too:

We will build a reference implementation of a Automated Content Detection service, again as a new Fediverse Auxiliary Service Provider with an open protocol.

This will allow server owners to opt-in to use external tools to scan content for spam, illegal materials, etc in order to help them fight bad actors; they could self-host these tools if they choose to do so, or share the infrastructure with other servers for better efficiency.

Tell me more about these bunkers, let’s Streisand this.

I would like to see more finely-grained feedback than just “no”. Particularly as it forces the voter to justify why they’re downvoting.

Some posts/contents are bad because they’re spam, abusive, or against the rules, or in the wrong place - downvoting these is useful crowd-sourced moderation, particularly for things that don’t warrant a mod report.

Some are just indisputably wrong and it’s a useful signal that the poster is talking rubbish. Then there’s the grey area.

Sometimes it’s just that the poster’s opinion is unpopular, and people are downvoting to suppress views they disagree with - I think that’s not a good thing for a discussion and encourages echo chambers. I like the idea of one of the downvote options being “I disagree” and that basically not doing anything.

I believe Piefed let’s you restrict downvotes to people that are subscribers. I think that’s a good solution to idiots downvoting from all just because they’re not interested in or understanding the post and/or the community it’s in. I also wonder if people even understand that they’re not training an algorithm, just dragging people’s posts down.

You mean because businesses don’t pay VAT? In that case it makes sense to not include it.