Unfortunately, this will also make aviation safety analysis more difficult for us.

I've tried Connect, Liftoff, Thunder, Voyager, and Infinity. It's just bugs galore. The back button will take you to the wrong activity, comments won't post (due to the app not the instance), some instances or accounts won't load, and a laundry list of UI or functionality bugs. Liftoff gave me the fewest issues of all of them, apart from using Lemmy in the browser.

But Sync really is polished, and the experience is night and day. I'm a big FOSS proponent, but I don't think it's wrong for a developer to want to make money when their income source suddenly disappears and they need to pay bills (and region-specific pricing is coming soon). The level of customizability, spit and polish in the other apps doesn't even come close. I'm sure they'll get there eventually, but I think it's a mistake to drive away passionate developers who want to help Lemmy succeed.

I've given pretty much every app or browser front-end a fair shake at this point, and I'm gladly choosing Sync for now.

I work in AI, and I think AI is overrated.

My favorite is Dracula, in part because they support so many applications.

Same. I've written a fish plugin, but other than that I just fish pretty much stock. It works and just gets out of my way.

Based on the data, people are still buying Samsung these days. Samsung consistently is in the top 3 by sales every quarter, and #1 in global smartphone market share goes back and forth between Apple and Samsung.

I recommend installing Obtainium. In Obtainium, click Add App and then paste the Codeberg URL for Infinity for Lemmy: https://codeberg.org/Bazsalanszky/Infinity-For-Lemmy

Obtainium will take care of installing and updating from various code repositories whether they're hosted on GitHub or Codeberg or some other place.

I use this on my Galaxy Tab and S23U all the time. It's really handy to be able to draw a diagram or take down a quick name or number.

As a federated learning researcher, I love to see articles introducing the public to the idea. But this article is really drawing a comparison between the fediverse and federated learning that doesn't make sense (to me).

Beyond the fact that data and compute are stored on separate servers, they really aren't similar. Federated learning avoids sharing data by sharing gradients or model updates with a central aggregator; raw data does not leave your device. The fediverse enables easy sharing of data between servers and avoids a central server.

Additionally, this article makes it seem like medical researchers were inspired by the fediverse, but the FedAvg paper was released in 2016—two years before ActivityPub was introduced in 2018.

I agree. The System UI has been crashing hard on my S23 Ultra recently due to the One UI 6.0 Beta, which means I can't open any app or even shutdown—nothing happens. Thankfully, you can force a shutdown by holding Volume Down and Power for 15 seconds.

There should always be a hardware force shutdown option.

It's not even enabled for Google Workspace accounts, but it's still forced upon you even though you can't use it.

For what it's worth I find the text prediction is really accurate for me (and it capitalizes I without issues).

This is a quality of life improvement I can get behind!

I wouldn't mind some shared content from quality communities, like the posts from the machine learning research community.

My family and friends primarily use Messenger, and we like the chat themes and custom like buttons and gifs and nicknames and other silly features. People are allowed to like different things.

I enjoy writing with fountain pens, and I've got to justify the numerous pens and inks I have. I also find it helps me with recall and focus. So I take notes by hand most of the time.

This exists!

The ASUS ROG Claymore II is an 80% "optical mechanical" with a detachable numpad. Mountain also makes several TKLs and an 60% (the Everest 60) with a detachable numpad. You could also make one using the KBDcraft Kit Adam and Kit Addams.

However, I agree with the others. I think you get the best build quality using a standalone mechanical numpad. If I was going to get a split design, I'd probably get an Ergodox or Keychron Q11 QMK or something similar.

The article and tweet say the watch will likely use the W940 and that the GW6 used the W930. Why does the Lemmy post say the opposite of the article?

Thanks!

I bought the Galaxy Watch 4 Classic for the bezel even though the 5 was out. I just needed a stopgap until the 6 came out. As long as the reviews are promising, I'm excited to upgrade. That said, the 4 is working just fine and I haven't run into any limitations.

Same. I chose programming.dev because it was close to my interests and seemed less likely to be drawn into federation drama. Losing Reddit already made me feel "homeless." I was worried about choosing a home instance, and then being forced to create a new account all over again because of admins squabbling amongst each other and instance drama.

I've got a few smart plugs, TVs, and a bunch of Hue bulbs. It's surprisingly easy to get to more than 20 devices.