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

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

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.

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.

Both worked for me on Sync. 🤔

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’m a frequent user of tqdm, but I’ve never seen that async example. Very cool!

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.

It happens for me too, even though I haven’t purchased it yet. I think not rendering the post in the list is a bug.

Can you explain these a little more? I don’t understand what that breaks. Also, shouldn’t the second one be correct if you use a raw string?

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

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

I’ve only played around with it for a few hours, but I’m already happier than using matplotlib or my rpy2 bridge to ggplot2.