bgtlover boosted

"Fediverse: Peopleverse!"

That's the way I'll campaign for #Fediverse and advocate it to others.

Notions of #Metaverse and #Web3 have in common that they lack humanity.

People, humans, mankind.. that's fedi right there.

There's the #SocialHub slogans of fedizens #UnitedInDiversity and #ReimagineSocialNetworking to help with that. Weaving a #SocialFabric together.

I'll also use #WEB0 #SmallWeb and #OpenWeb where appropriate to highlight specific characteristics.

Rise of the #Peopleverse

bgtlover boosted

Devices are general purpose computers. Yet manufacturers, vendors, and internet platforms exercise monopolistic control over equipment and restrict #softwarefreedom. We respond: users should have access to alternative services and content in their devices

bgtlover boosted

Oooh, just came across an article in the newspaper mentioning Jitsi, the FOSS video-conferencing service! 🥳

Original link:

(works in :firefox:​ reader mode)



Working with and being slightly frustrated by over here.
Does anyone know how I can boost or favorite a post, reply to a thread and so on? I've tryed to long press the message to no avail, it just opens the message in a separate window as it does with a normal click, activating the more options button does nothing but pops up a menu with a single item, open in browser, where I have to be logged in for things to work. So, does anyone know how I can better operate with talkback?

bgtlover boosted

Asahi Lina got the Linux GPU driver for ARM Macs written in Rust to start working 🚀

it seemns the world has gone crazy, yet again. This time, some researchers managed to make a chat AI bot that would replicate the linux operating system in a conversation. Its answers were, while funny at times, what I would expect from an avatar embodying linux and the free software movement. Check it out here:

bgtlover boosted

it seemns matrix patched some end to end encryption vulnerabilities in the matrix js, matrix android and matrix ios sdk. For those of you using element, I recommend you update asap. If you are a package maintainer of element for a linux distro, consider packaging the newest version containing those fixes, some of the vulnerabilities are critical and could potentially impact users. For more information, see here

bgtlover boosted

Learning async Rust. This is an *excellent* talk on what the Rust compiler does to turn an async function into a state machine -

bgtlover boosted

The merge date for Rust in the Linux kernel keeps getting pushed out, previously it was 6.0 now we're pushed back to 6.1. Who wants to bet we see another delay

bgtlover boosted

Some days ago @distrotube suggested an alternative way of donating to Linux: Seeding their Torrent.

I generally have to keep one of my computers up and running for analysis, usually I don't have high network or disk IO usage, so I put that into good use, downloaded torrents from some of the distros I thought can use seeding, and the rest is history.

I didn't know which distros are in demand, but now based on 4 days running, it seems OpenMandriva, @EndeavourOS and @mxlinux are in demand.

bgtlover boosted
bgtlover boosted

Secondly you shift the issue over to the content blockers instead of the browser, now users will see that the plugins aren't working like they're supposed to be so there's no point even keeping it installed anymore.

Show thread
bgtlover boosted

I have only played with #columnq for a few minutes, and I can't tell how much I love this tool.

As somebody who plays with tons of JSON datasets, and still receives tons of CSV files at work, this is exactly what I needed to make my life easier.

So far, performing some complex queries on JSON or CSV datasets required me to set up some temporary relational database and dump the data into it just so that I could query it. Or spinning off a Jupyter notebook or a Python script that could parse the data and run the queries I wanted via code.

All these workarounds are probably a thing of the past now that e.g. getting the countries with a highest number of cities from the dataset is something as simple as this command line:

columnq sql --table ./cities.json \
'select country_name, count(*) as cnt from cities group by country_name order by cnt desc'

bgtlover boosted

Friendica (@news) is a very flexible Fediverse server type with lots of useful options and a Facebook-style web interface with no character limits (or at least I couldn't find one!).

Friendica lets you follow Fediverse accounts from Mastodon etc., but it also lets you follow RSS and Twitter feeds, or turn RSS feeds into Fediverse accounts that people on other Fedi platforms can follow.

There are also extra features like calendars and groups.

The flexibility does make it a bit trickier to use, but once you get the hang of it there are things you can do on Friendica that aren't possible on other Fedi server types.

There's no dedicated Friendica app, but if you have Android you can use Friendica through Fedilab (@apps), and of course it has a web interface that works on all platforms.

There's more info at and you can find servers to join at or

#Friendica #Fediverse #ActivityPub

bgtlover boosted

TIL that in Rust you cannot destructure something that implements Drop.

impl Drop for MyStruct {}

let MyStruct { field1, .. } = my_struct; // cannot move out of...

It makes sense. Ran into it while refactoring some really ugly code into pretty, idiomatic code now that I know how to do it.

bgtlover boosted

Do you write code for free and open source projects? Would you like to learn the basics of the Rust programming language? I’m offering to teach the basics of Rust to free and open source software programmers, for free.

#Rust #RustLanguage #FOSS #Training

bgtlover boosted

When people say they have nothing to hide, let's reply:

* Post your passwords here
* Write down everybody you hate
* Tell the world who you voted for
* List all the videos you watched

- Medical history.
- Sensitive conversations with employers, children, spouses.
- Billing and banking information.
- Purchase information.
- Web search history.
And more...

I need privacy. Not because my actions are questionable, but because 'your' judgement and intentions are.

And it doesn't matter whether or not my communications are weighty or frivolous in nature - they are mine and only for those whom I've chosen to share them with. Not the government's, nor goog's, nor my ISP's, but mine.

Also, what is legal where I live today, may not be legal tomorrow and my discussion of it may incriminate my future self. It has happened in the past many times to others in other places.

does anyone know how to boost messages, reply to threads and stuff like that with tusky? Maybe I'm missing something because there might be invisible, inaccessible parts to the interface, but all I see when tapping on a post is a thread opening, even if the post is not in a thread, but then the options menu only shows open in browser, nothing like boost and so on.

bgtlover boosted

#AI #SmallAI #SelfTrain

One of the arguments in favour of surveillance capitalism is the great usefulness of cloud-based ML predictions.

After all, who can deny the usefulness of photo apps that automatically recognize faces, detect your speech or help you making sense of the deluge of information in a social feed?

The argument usually goes like this: these features require large neural networks, which in turn require a lot of computational power to train the models, and a lot of memory and disk storage in order to load and save those models.

You can't do such things on small devices that run on batteries. Therefore your phone *HAS* to send your data to #BigTech servers if you want to get those features. Otherwise, you just won't get those features.

Except that... What if this whole argument is bollocks?

#POET (Private Optimal Energy Training) proves that you can run both the training and the predictions locally, without compromising neither on precision, nor on performance.

After all, the really expensive part of training is back-propagation. POET breaks down the back-propagation performance issue by quantizing the layers (so real-number large tensor multiplications get reduced to smaller multiplications of integer tensors, without sacrificing precision too much), and a clever way of caching the layers that are most likely to be needed, so we don't have to recalculate them, without caching everything though (which would be prohibitive in terms of storage).

The arguments in the paper sound very convincing to me. The code is publicly available on Github. I haven't yet had time to test it myself, but I will quite soon - and try to finally build an alternative voice assistant that can completely run on my phone.

Show older

Linux geeks doing what Linux geeks do...