Here's the S20+. It's not a limitation of the subpixel array. In case of one, the colors would be switching sequentially like on a thin cable shot far away with a Bayer sensor. Here, the individual colors are evidently "centered" with a bias to the expected center of the pixel. This allows them to reliably drift in their own direction.

And pretty much yes, I believe that some of the people who thought PWM irritates them, may actually dislike misconfigured subpixel rendering. While fringing is one thing, some of the SPR configurations can result in blur. It's a fairly straightforward topic that probably isn't going away just yet.

Knowing the limits of your hardware before taking photos ❌
Getting a variable aperture ❌
Repainting unfocused faces ✅

Yeah...

https://youtube.com/watch?v=oP7xzAdCYQU

I wonder if they even paid him for using his personal brand, considering he's not in the video even for a moment. Too cheap to collab with a streamer?!

To be clear, Keymint remote key provisioning changes this to a really massive degree.

https://android.googleapis.com/attestation/status

This json will either disappear at some point, get redefined or grow to megabytes, since device-wide (all units running the same build during a given time frame) public listing of revocations would no longer be practical – the banned "serials" would supposedly have to be per each unique app installation. Yes, I'm still trying to comprehend it. No, collecting unit-unique public keys at the factory and uploading them to Google isn't the whole story.

Tldr: if whoever decides to play by Google's rules, attestation/PIA/screen proxying from a device you personally own may actually let you live in peace, but for all it's worth, it really seems that Google won't have to choose between revoking a whole device (bad for PIA adoption) and letting non-compliers in.

To benefit from the new level of control, Google will likely just break the promise of:
"Additionally, Google backend servers are segmented such that the server which verifies the device’s public key does not see the attached attestation keys. This means it is not possible for Google to correlate attestation keys back to a particular device that requested them."

If they won't, why all this effort?

https://linuxiac.com/linux-snap-users-warned-as-attackers-push-malware-through-old-trusted-apps/

Looks like Snaps don't have to be signed before uploading them and then are only signed by the store. A quite bad design.

As a matter of fact, I do wish someone would test this Google's promise, there's nothing to lose because the architecture for granular revocation is now in place and changing it would only be a change of policy, not of client side implementation.

For comparison, Flathub's process is Github-reliant (accounts are forced to have 2FA by the org)

I missed a very important event from years ago. Moto invented a charger that differs in its max power depending on how you put it in the socket.

https://www.theverge.com/tech/864595/nexphone-android-linux-windows-11

Don't underestimate this one. For reasons.

This image is cryptographically attested to be completely real
*adds a third eye on your forehead*

Inspired by Google?!

I used to say that the only sites without a numerical article ID or dates in URLs are an obvious SEO spam, but now it's more and more frequent. The syntax of your URLs has completely valid reasons to change over time and any measure that prevents link rot should be appreciated. It's great when a dead URL gives you a clue.

Btw, if you'd want to increase the chance that a conspiracy in your device gets found or simply inform the public about new or exotic device, proceed here and send a message with captured file to that channel (as long as your device isn't there yet):
https://news.1rj.ru/str/cam2caps/331

Word is spreading that you should absolutely stop updating your Snapdragon devices if you want a bootloader unlock on previously not unlockable models.