Tadi Channel – Telegram
Tadi Channel
796 subscribers
357 photos
12 videos
6 files
219 links
Random stuff I consider worthy of sharing. Mostly tech.
Download Telegram
https://grapheneos.social/@GrapheneOS/115164183840111564

Delaying access to security patches by 4 months to nichest OEMs and the public to satisfy sloppy biggest partners with bureaucratic update policies is a little bit crazy.
Government-funded threat actors are safely assumed to be in possession of patches in question as they're exposed to lots of individuals for the sake of not creating any inconvenience. It's hard to not agree with GOS on this one.
😐8👏1
On the left, a 2019 phone.
On the right, a 2025 phone Sony considers as a proper representative of modern midrange.
😐81😁1
Tadi Channel
Photo
♻️🌱 (it's also combined with Snapdragon 6 gen 3, which is a refresh of first gen)
😐8🔥1😁1
Tadi Channel
https://grapheneos.social/@GrapheneOS/115164183840111564 Delaying access to security patches by 4 months to nichest OEMs and the public to satisfy sloppy biggest partners with bureaucratic update policies is a little bit crazy. Government-funded threat actors…
https://www.androidauthority.com/android-risk-based-security-updates-3597466/

So this is great news for the laziest OEMs. If Google determines that no patch is important enough to make it a part of monthly SPL (despite being aware of vulnerabilities), an OEM is literally ennoscriptd to a free string update. Yes, an update that literally changes nothing except the security patch date – the date alone.

But wait, if Google knows what constitutes an important patch, why OEMs can't even name the less important vulnerabilities they decided to patch out of their own will? Oh, maybe they aren't as much unimportant as Google would want you to believe and malicious parties can now count on a longer exploit lifecycle... That would be unfortunate.
😐8👏2
Honestly, this is really ugly. As much as it's hard to swallow it, GOS is absolutely right in their Mastodon thread. Not relying on an OS vendor known to merge private patches as soon as possible now means being months behind instead of just one. Unless they start leaking and getting merged by more custom ROM projects, a person reasonably expected to be targeted no longer can trust any builds (including stock) that aren't known to utilize them.

In current state of things, yes, no important person should use anything but an iPhone or the supposed special edition of GOS. I'm mad that I have to shill them.
😐5👏4👍1
🗿
😐29🔥9🤯6👏52😁1
I present you a world record: industry's largest watermark of 34 million pixels.
😁19😐4
There's a certain Telegram feature used for most simple form of scams you could possibly imagine. Instead of taking time for growing multiple channels with something that interests their followers, a significant channel's owners post invite links to empty channels with manual acceptation, promising whatever kind of content gullible people would want to see.
Such a channel collects a number of join requests only to turn into a "legitimately followed" scam place, right before accepting all of the requests.

Why do I mention this? Telegram doesn't let you see or cancel the pending requests as a user. A just little bit tech savvy person who'd have been duped would realize that the request is taking too long for acceptation and possibly cancel it if such an UI was available. Telegram could also limit the maximum time a request can remain unaccepted before cancellation, for example to a month, for the sake of making the scammer life a little bit harder. Another option would be to require second confirmation that mentions the channel name from the moment of your request, you'd get an option to join or resign and you wouldn't count as a member before it.

Personally, I'd want it to be educational, so that old people would make the connection between scams popping up in their chat lists and the kinds of promises they were given. Their worldview could genuinely change once they start seeing money behind the "wholesome", "secret" and "alternative" and realize why they're specifically targeted.


(It's a month old post I decided to post now after seeing this scheme is alive and well, I felt like describing it is somewhat unproductive compared to directly filling a report, but official interactions tire me and maybe someone could make it reach the right place better than I can)
👍93😐1
I can't think of a good caption, just usual iOS sideloading experience. I still wonder how three letter agencies bear with this system while limiting the money trail, it seems impossible.
😁10😐1
Tadi Channel
According to Camera2 Test output (not the best, but sufficient), the report in question (pic attached) doesn't seem valid as the reported white levels are 1023. The guy is just likely excited by quality of 10-bit raw video and there's nothing to see there.…
So GN8 can absolutely do it at 16:9 in 30 fps (duh, I don't think any sensor vendor would dare to advertise DAG/DCG if the lightest realtime configuration couldn't be handled), 4:3 and whether it's anywhere to be seen on P10 is TBD
😐1
Tadi Channel
‘iPhone 17 Pro’ reappears in regulatory listings alongside Pro Max, Air Smh, now the high-end everywhere will consist of a basic model, same size pro model, a big specifically thin model and a big pro max/ultra? Apple effect is scary. But thanks for smaller…
Now that I think of it, most of the compact Pro influx won't be unlockable without bribery, luck or significant leaks. It's why "the best compact phone" to me is still OP13T, even with its OIS-less telephoto.
👍3😐2
🤯407🎉3😐2
I overlaid the two Xiaomi 17 Pro renders visible in Xiaomi teasers and it seems clear as a day that only the bigger one is getting a massive periscopic telephoto despite all of the space on both devices. What you can see on the smaller model is at best a 1/2" format.
😐8
Tadi Channel
setprop vendor.camera.force.sensor.mode 10 or setprop vendor.camera.force.sensor.mode 11 followed by killing of camera stack (no command at hand, but you should figure it out) will give you the 16:9 12 bit DCG mode in question on P8P.
On Pixel 9 Pro, a 4:3 most-likely-DCG has the ID 14. It's 12-bit and offers a lower amount of noise than default mode. This may suggest some greater intentionality behind exposing it later on P10P.
1😐1