But python is more of a first class citizen in libreCalc.

You can use Numpy, Pandas, matplotlib, etc… In both excel and librecalc since recently and excel has a hacky runtime inside of it that often has to be restarted if leaving it idle for too long.

You can do 10x what you can do in excel + VBS with python packages. Dataframes alone for data analysis is night and day compared to VBS hacks.

It is certainly not open source, but it is a copy left license.

You could make the exact same argument for literally any open source project where someone forks it.

The “power” of open source is that in theory if the creator went rogue that the public can switch to the fork as the main app used.

I really fail to see if the dataset is open, how the data itself (of course not the web domain or front end) would be any different in a fully open source project. Maybe the transfer to a new app in the last case would be quicker, but that is about it. The dataset is the gold here.

Maybe a bit like kids being told they are “smart” and “gifted” turns a lot of them into workaholics with a fear of skill-based failure.and coping mechanisms around that.

Nutella can never even be inferred to be a nut butter lol. It is literally >50% sugar and palm oil. Only 13% hazelnuts. In 2017 they reduced the chocolate solids further and added more sugar and more skimmed milk powder. Almost half its calories is from sugar.

The traditional hazelnut butter was 71.5% hazelnut paste and 19.5% chocolate, but nutella doesn’t even have enough cocoa powder to have the label of “chocolate cream” in many countries.

Really? Holy shit I can switch to zen fully at work and at home and uninstall chromium. Webserial was literally the only thing I needed

Hey there!

This is a job for Handy. I recently started using it with the Parakeet recommended model. It works very very well. Whisper packaged with Wyoming that I run is pretty much unusable for dutch but Parakeet in Handy seems to work quite well. Direct input of text into whatever program you want. Downloads a small local model and works only on your computer. Push-to-talk or toggle, tons of customizations.

Note that this just speech to text and the models aren’t made for post processing. So it will dictate exactly what you say, umms and everything. If you want cleanup, formatting, etc… Then you have to rely on a much more general model, but the makers are experimenting with an opt-in for that for various external AIs.

26°C inside? That is rough.

Hell, Runkeeper was a thing with a social network before strava and still exists.

It even gives cadence which strava doesn’t do (at least not the free tier), but the Apple cult picked up Strava and it became the runner’s status symbol as Apple devices are yoany people in America/Sweden/etc… and then that spread over the rest of the world.

Personally I have only a hundred notes or so and really only the basic plugins and it still takes up to 10 seconds to load and become usable on my phone.

It is definitely not fast loading up but it is very fast in most other use cases I have seen. At work I use it with getting more towards 1000 notes and it works fine there usually, though there is some windows+ electron weirdness

• Domain 15€/yr
• Music 8.50€/month (trying to convince my girlfriend to move from Spotify to Qobuz

That is it I think. I don’t consider internet/water/electricity subscriptions as much as utilities.

Can’t really afford anything else right now because my girlfriend as a cafe owner only brings in around 1200€/month or so working >60hr/week and we have a full renovation to pay for.

I want to dedicate like 10-20€/month to FOSS I use once the situation is better.

Aything from yeyushengfan258

Like Afterglow (Photo) or Twilight

How is this different than Nearby Glasses except the license is less copyleft?

I completely overlooked the android app in the readme, thanks! I have a server set up so can I start local to try it out and migrate to a server later?

Also is there support for mealie through an authentication platform like authelia?

Any plans for releasing on F-droid?

Edit: Oof, you can’t search for foods when making a recipe or meal, hopefully that comes because otherwise it is a huge process to make any custom meal or recipe…

Damn, I was hoping that it would be an android app with Health Connect support 😅

Looks very nice though!

One thing that people often forget is TVs / stream boxes and Alexa/google assistants.

Smart TVs all have microphones that are recording (for “voice commands”) and the same with the remotes for stream devices (though probably not unless asked because of battery life), also google and Alexa’s are constantly always listening. They are simply spyware devices that parse everything you say and hand it to advertisers, insurance, governments, etc…

Phones also suffer from the battery life problem, so jury is still out on whether they listen to you because constantly recording audio would degrade battery life quite a bit (though maybe it is factored in). Phones absolutely do share location data and any phones discovered on the same WiFi network or in the same location if that data was available and from there will share entire search history of devices on the same network (well, that is on the data center end, phones likely just send what devices and for how long they were together). From there it is quite easy to advertised based on search history, unencrypted text data, etc…

For example in OP, OP had likely searched around about cars and checked out manufacturers websites and such before meeting with the parents (unless they were going in completely blind to dealerships), so it would have shared exactly the cars that they were looking at and linked it to the parents for advertising.

Still fucked, ethically wrong and legally grey, but “listening” is a bit of an inefficient way to do it, generally.

Can you provide one or elaborate on it?

Embedded developers have tried all manner or wizardry to simply track speed, not even position based just on an accelerometer/gyro, but the sample rate error drift is so large that putting a GPS module in there is 100x more accurate for deriving speed.

I would be interested to see how a browser, which almost certainly doesn’t get the full serialized data, is able to track just based on that which the wearables industry have been trying for decades with bad results.

Been using it (not often beyond basics for home server container administration) for years but I always get confused with searching vs search and replace vs global search

/<search> (n for next instances)

:%s/<search>/<replace>

:g/<search>/p (grep was named after this one IIRC, Global Regular Expression Print)

And often I just use :%s to search and highlight all instances and don’t press enter because I forget the others and I often have to search and replace for docker comppse snippets.

Yes, but also on the hardware level.

I don’t know enough about OS programming to know if it is the architecture or the (closed source, as mentioned) CPU design itself that is more difficult to implement.

Looking at the MCU space, even with a known architecture (like ARM), each processor has to be individually implemented in software and firmware which is a ton of work, and the only people who necessarily know how are the processor designers unless it is open source. But take that with a big block of salt, because I have never done it, just looking at industry practices.

I wish I could use Graphene, but the pixel’s pitifully small and expensive storage to force you into cloud storage really makes an SD card necessary. I have around 200GB on my phone + SD already(including OS), and I don’t want only 50 GB for more photos and music in the next 7 years WITH the more expensive option…

Isn’t persona the Palantir (AI powered genocide) affiliated company?