wild that artists now have to film their whole process just to prove they’re human. is a process video actually enough proof though? AI video is getting good fast. give it a year and this won’t work either
F*ck. Everyone’s sharing this, so I actually tried it on my Android.
Here’s the problem: that folder doesn’t open on a normal phone. Not through a backup, not through ADB. You need root.
And the original story was about an iPhone 3 days ago, with different numbers. Now it’s “Android, 14,219 photos”. The story is mutating as it spreads.
What you can actually check in 30 seconds: Settings → Privacy → Permission manager → Camera. See which apps have access and kill the ones that don’t need it.
Bir içerik üreticisi:
Claude Code, Android telefonumda benim çekmediğim tam 14.219 yüz fotoğrafımı buldu.
Telefon, elime aldığım hemen her an gizlice fotoğrafımı çekiyormuş. Yatakta yatarken bile. Evdeyken bile.
Ekran görüntüsü falan değil. Bayağı ön kameradan alınmış gerçek kareler.
Trende. Arabada. Evde. Sadece saate bakmak için telefonu elime aldığımda bile.
Sistem önbelleğinde öylece duruyorlardı.
Hepsinde tarih, saat ve GPS konumu var. Galeride yoklar. Son Silinenler'de yoklar.
Telefonu bilgisayara bağladım, tam yedek aldım ve Claude Code'dan galerimde olmayan tüm ön kamera dosyalarını bulmasını istedim.
Gitti, /data/system/users/0/faceunlock/cache dizinini ve Google Play Hizmetleri içindeki başka bir klasörü çıkardı.
14 binden fazla dosya. Hem normal fotoğraflar hem de Yüz Tanıma sisteminin (Face Unlock) kızılötesi kareleri. İstisnasız her birinde zaman ve konum kaydı var.
Fotoğraflara bakmaya başladım. Çoğu son 15 aya aitti. Sonra "2 saniye önce" oluşturulmuş bir dosya gördüm.
Ekranda klasöre bakıyordum. Ben klasörü incelerken bile telefon fotoğrafımı çekmeye devam ediyormuş.
Google ve telefon üreticileri, Yüz Tanıma sisteminin yüzünüzü matematiksel bir modele dönüştürüp görüntüyü anında sildiğini iddia ediyor.
Ama bu telefonda ham görüntüler klasörde aynen duruyordu. O dizine veri yazan işlemlerin bazıları sistem tanılama (diagnostics) servisleriydi.
Hâlâ kamera izni bulunan birkaç eski uygulama da bu klasöre erişebiliyordu.
Tanılama verilerini kapattım, gereksiz tüm uygulamaların kamera iznini kaldırdım ve önbelleği temizledim. Ama bu karelerin kaçı çoktan başka bir yerlere sızdı, hiçbir fikrim yok.
Kendi cihazınızı kontrol etmek isterseniz: Telefonu bilgisayara bağlayın, tam yedek alın (ya da ADB kullanın) ve Claude'dan Galeriniz dışındaki ön kamera dosyalarını bulmasını isteyin. Benim dosyalarım yukarıdaki dizindeydi.
F*ck. I thought this was a fake screenshot, so I went and checked. It’s real.
VRChat just posted it on their official forum: no “AI-generated” label and no AI filters. Not now, not planned.
Here’s the wild part. They ran a survey, 21,595 people answered in one week, and 95% said they want some kind of AI label. VRChat read that and said no anyway.
Their reasoning: some creators would lie about using AI, and labels would turn the community into AI detectives who harass people, even ones who never touched AI.
What they’re doing instead:
likes and dislikes on worlds, aiming for end of 2026
bigger world galleries in 2027
misleading thumbnails are now against the rules
a “Human-authored” mark is being considered but not promised
So the idea is: don’t tell people what’s AI, just let them rate if it’s good.
Honestly I get the harassment argument. But when 95% of your users ask for one thing and you give them a thumbs down button instead, that’s a choice.
This is going to be the template for every platform. Steam, Roblox, YouTube, all of them will face the same question.
Would you want a label, or do you only care if the thing is good?
F*ck, we’re actually here now.
Someone just ported Sonic 1 to the Game Boy Advance using AI. A 90s Sega game running on a Nintendo handheld. One person and an AI.
Think about what that used to take. You had to understand the original hardware, understand the target hardware, dig through old code, and grind through months of trial and error. That was a job for a small team of very specific nerds.
Now one person can just try it.
This is the part people keep missing when they argue about AI. It’s not really about replacing artists or writing your emails for you. It’s about the gap between “I wish someone would make this” and “I made this” getting smaller every month.
The people who win the next few years won’t be the ones with the most skills. They’ll be the ones with the most ideas and the guts to actually start.
So what’s the project you’ve been sitting on because you thought you couldn’t build it?
I JUST PORTED SONIC 1 TO THE GBA USING AI — THIS IS GENUINELY INSANE. 🦔
I fed it the original Mega Drive disassembly, got it running on GBA hardware — even added a new mode with the Spin Dash. 🎮
This being real is so exciting to me — we're literally living in the future. 🚀
F*ck, look how fast this actually went.
2016: AI couldn’t tell a chihuahua from a blueberry muffin.
2019: it could barely write a paragraph that made sense.
2022: ChatGPT shows up and everyone laughs at its wrong answers.
2026: it writes code, edits videos, answers emails and apparently does her job too.
Ten years. That’s how long it took to go from confusing dogs with muffins to replacing people with degrees.
Yelling at AI on Twitter won’t slow it down even a little. The people who’ll be fine are the ones learning to use it right now, not the ones cursing at it.
The muffin era is over. Adapt or get left behind.
I fucking hate every single person who made this possible fuck you and your AI and fuck the system that made people work their buts off for decades and decades going to school then college then find a job only for that job to be taken away from them with no plan B in sight.
F*ck it, I’ll say what everyone’s thinking.
People did everything right. School. College. Student debt. Years at a job they didn’t even like. And now companies are offering buyouts and cutting whole teams because AI can do part of the work cheaper.
Nobody voted for this. Nobody asked the people who are getting replaced.
I use AI every day and it’s genuinely useful. But “just learn AI” is not a plan for a 45 year old with a mortgage and two kids.
If AI is making these companies this much money, some of it should go back to the people it’s replacing. Real severance. Paid retraining. Time to switch careers. Not a “thanks for your service” email on a Friday.
The tech isn’t the villain. The way it’s being rolled out is.
Are you actually scared AI is coming for your job, or is this all overblown?