FocalZero retweeted
BREAKING: Claude Code found 14,219 photos of my face on my Android that I never took. Phone was secretly taking them almost every time I picked it up. Even when I was just lying in bed. Even at home.
Not screenshots. Real frames from the front camera. On the train. In the car. At home. When I just grabbed it to check the time.
They were just sitting in a system cache. Timestamps and GPS on them. Nothing in Gallery. Nothing in Recently Deleted.
Plugged the phone into my laptop, made a full backup, told Claude Code to look for any front-camera files that weren’t in my photos. It pulled /data/system/users/0/faceunlock/cache and another folder inside Google Play Services. Over 14 thousand files. Normal ones and infrared from Face Unlock. Every single one had time and location.
Started scrolling. Almost everything from the last 15 months. Then I saw a file created 2 seconds ago. I was still looking at the folder. Phone kept taking photos while I was reading this.
Google and the phone makers say Face Unlock turns everything into a mathematical model and deletes the image right away. On this phone the raw frames were just sitting there. Some of the processes writing to that folder were from system diagnostics. Couple of old apps that still had camera permission could also get in.
Turned off diagnostics, revoked camera access from everything that doesn’t need it, cleared the cache. But how many of these frames already went somewhere else, I don’t know.
If you want to check yours, plug it in, make a full backup or use ADB, and ask Claude to look for front-camera files outside your Gallery. Path above is where mine were.
Claude Code found 145,763 photos I permanently deleted from my iPhone.
Three years of deleted photos and videos still sitting in a hidden cache. Any app with file access can read them.
I plugged my iPhone into my MacBook to run a backup. Claude Code was doing the security audit from my last post and flagged a directory I wasn't even looking for.
PhotoData/Thumbnails. Over 146,000 files. Photos, videos, screenshots. Every single one was something I deleted.
Photos from 2023, 2024 and 2025. The ones I deleted the second after taking them. The ones I cleared from Recently Deleted. The ones I deleted three separate times because I wanted to be sure they were gone. All still there.
Passport photo I took for a visa application in 2023.
Screenshots of my bank app with full account numbers visible.
Private photos I sent someone and immediately deleted after a breakup.
Screenshot of my crypto seed phrase I took temporarily in 2022.
Photo of my ID and credit card I took to remember the numbers and deleted 5 seconds later.
Screenshots of my ex's private Instagram stories I saved and permanently deleted in 2023.
Some were full resolution. All organized by date going back to 2021.
When you delete a photo on iPhone, iOS removes it from your camera roll and Recently Deleted. It keeps a compressed copy in a Spotlight search index cache so your phone can find things when you search. Apple never clears it. Not after 30 days. Not after a year. Never.
They told you it was permanently deleted. It wasn't.
I checked which apps on my phone had Files and Folders access. 23 apps. Weather apps. QR code scanners. PDF readers. A flashlight app I downloaded in 2022. Any one of them could have been reading every photo I ever deleted this entire time.
Went from 23 apps to 3 in about two minutes.
Full step by step below. Takes 2 minutes. Save it so you don't lose it and send it to anyone with an iPhone.
🤯 牛逼!有人把一台完整的 RISC-V CPU 直接编译进了 Qwen3.5-4B 的权重里!
这叫 Qwenputer:
🔹用 Qwen 自己的 Gated DeltaNet 层当精确 RAM
🔹在模型前向传播里跑 C / MicroPython
🔹不需要任何外部运行时或工具调用
算术基准结果炸了:基础模型 48–90% → Qwenputer 97.5–100%(基本打平外部 Python 工具)
Qwen 的架构不只是能聊天,还能把计算机编进权重里。这波是真·模型即计算机。
代码/模型/Demo 都开源了,值得一看。
#Qwen #Qwen35 #Qwenputer #RISCV #AI #LLM #开源 #阿里巴巴
xitter.cf/nft_chen/status/211541…
🚨炸了!阿里 ATH-MaaS 发布 OvisOCR3:仅 0.8B 的统一小模型,一个 checkpoint 干翻整套文档处理流水线!
同尺寸SOTA硬核性能:
🔹OmniDocBench 96.74(逼近顶级专用模型)
🔹PureDocBench 82.31 第一,超 Qwen3.8-Max
🔹OCRBench v2 英/中双榜第一(65.9/61.3)
🔹KIE 全面领先,跨页表格 Relation F1 97.07 反超大模型
主要亮点:
1️⃣文档解析 → Markdown/JSON(LaTeX公式+HTML表格)
2️⃣通用OCR + 边界框 + 图表数据恢复
3️⃣关键信息抽取(按字段/Schema直接出JSON)
4️⃣图表定位+Caption + 流程图转Mermaid
5️⃣跨页表格智能重建(自动判断并拼接)
基于 Qwen3.5-0.8B + 多阶段蒸馏/RL,Prompt 一键切换任务,Apache-2.0 开源,1.7GB 级显存本地可跑,RAG/企业文档管线直接能上。
小模型真的能替代复杂流水线了!
ModelScope:modelscope.ai/models/ATH-Maa…
HF:huggingface.co/ATH-MaaS/Ovis…
报告:arxiv.org/abs/2610.33066
#OvisOCR3 #小模型 #OCR #DocumentAI #ModelScope #开源AI #端侧AI #阿里AI
FocalZero retweeted
Introducing Reality-Aware Rendering in @Simulon, a completely new way to create with the real world.
Place neural assets with your phone and act out the scene. When you stop recording, Simulon renders the take, reading everything in it (what you do, what you say and the place around you), and your assets respond to all of it.
The final is regenerated from your recording: a new performance of the scene, not a copy of the footage.
Available now on iPhone. Coming to Android soon.
Introducing the Comfy Creative Partners Program.
Six agencies and studios already delivering high quality client work with ComfyUI, each with its own specialty.
→ Asteria: AI film + animation
→ Black Math: design + experiential
→ Brick: architectural viz
→ Contender: custom models + team upskilling
→ Native Foreign: motion, VFX, 3D
→ Silverside: AI-native brand campaigns
Click the link below to learn more ⬇️
L'inférence
Sara Conor en parlait sans en connaître le terme
On veut être dev sans comprendre une ligne de code, on veut raconter des histoire sans écrire, on veut dieu sans confession.
L'agencement de la chaine a changé, mais la connerie humaine est un socle difficile à bouger.
FocalZero retweeted
動画内の人物をローカルで差し替えるComfyUI向けツールが注目を集めている。
元の服を反映させず動作のみを保つ工夫や手軽さを評価する声が多数。動作環境への関心が高い一方、一部で接続エラーの報告も。
#MiniMaxH3 #ComfyUI
URLはリプ⬇️
FocalZero retweeted
This is f*cking insane. Nvidia founder and CEO Jensen Huang:
"Nobody writes prompts anymore. The new job is building loops and graphs
The engineers who stopped prompting are already years ahead, and the rest will end up working for them"
In this 50-minute talk, Jensen explains why prompting is over and what the best engineers are building instead
Prompts → Agents → Loops → Graphs → AI Workforce
Using AI doesn't mean writing better prompts into a chatbot
It means loops that do the work, and graphs that run the loops without you
95% of people are still polishing prompts, the 5% who build loops and graphs will run everything
Worth more than any $2,000 AI course
Bookmark and watch it today
Then save the full harness engineering guide below before everyone catches up ↓
Je regarde tous les aspects, pas juste ce qui me plait. On est tous hypé par la tech, cela ne doit pas nous rendre aveugles pour autant. Un mimic encode ton style mais aussi tes névroses. Il va amplifier/automatiser les patterns récurrents, efficients tout comme les plus toxiques
Payer le cloud 0.6$ pour faire un combat entre une bimbo de 50kg et un monstre de 3m en mode super sayien, tout le monde sait faire!
Raconter une histoire c'est autre chose.
Il y a une demande monstrueuse et très peu d'opérateurs. Les enfants ne demandent pas, ils regardent.
Comfy API puts a straightforward UI on your most complex workflow.
Silverside, a creative innovation studio, builds custom AI tools for brands, and their developers build on Comfy API.
For one beauty brand:
→ Product shots + brand guidelines + trend data in one workflow
→ Open and closed models working together
→ A simple UI: pick a shot, add direction, run
→ On-brand images + 15s video out
Get your API Key ⬇️
FocalZero retweeted
人物写真1枚から4面キャラシートを生成するMiniMax H3ベースのComfyUIワークフローが公開。
正面・斜め・横・背面の4視点を横並びで出力。写真外の服装などをテキストで補完する機能やSeedVR2拡大に対応。
#MiniMaxH3 #ComfyUI
URLはリプ⬇️
FocalZero retweeted
SCIENTISTS JUST CREATED THE FIRST SYNTHETIC CELLS THAT GROW, REPRODUCE, AND UNDERGO NATURAL SELECTION
They literally engineered the building blocks of ARTIFICIAL LIFE from nonliving materials.
Playing God comes with major consequences.
We are entering dangerous territory.
FocalZero retweeted
BREAKING: Claude Code found 12,847 photos of my face on my iPhone that I never took.
It looked like the phone was taking them every time I picked it up. From the club. From home. At a red light while I was driving. Almost every time I took the phone in my hands there was a frame.
They were sitting in a system cache with timestamps and location data. None of them were in Photos.
I plugged the phone into my MacBook and asked Claude Code to scan for any front-camera files that weren’t in my Photos library or Recently Deleted. It flagged Private/var/mobile/Library/Caches/com.apple.faceid. Over 12,000 files. Mix of infrared frames and regular front-camera images. Each one had a timestamp and GPS coordinates attached.
I started scrolling. Most were from the last 14 months. Then I saw one from four seconds ago. I was still looking at the folder.
Apple says the TrueDepth system turns the capture into a mathematical model inside the Secure Enclave and discards the image. On this phone the raw frames were still in the cache. Several of the processes that wrote to the folder belonged to system diagnostics. A couple of old apps with camera permission also had access to the same directory.
I turned off diagnostics sharing and revoked camera access from every app that didn’t need it.
If you want to check your own phone, plug it in, run a backup, and ask Claude Code to look for front-camera files outside the Photos library. The folder path above is where mine were.
Claude Code found 145,763 photos I permanently deleted from my iPhone.
Three years of deleted photos and videos still sitting in a hidden cache. Any app with file access can read them.
I plugged my iPhone into my MacBook to run a backup. Claude Code was doing the security audit from my last post and flagged a directory I wasn't even looking for.
PhotoData/Thumbnails. Over 146,000 files. Photos, videos, screenshots. Every single one was something I deleted.
Photos from 2023, 2024 and 2025. The ones I deleted the second after taking them. The ones I cleared from Recently Deleted. The ones I deleted three separate times because I wanted to be sure they were gone. All still there.
Passport photo I took for a visa application in 2023.
Screenshots of my bank app with full account numbers visible.
Private photos I sent someone and immediately deleted after a breakup.
Screenshot of my crypto seed phrase I took temporarily in 2022.
Photo of my ID and credit card I took to remember the numbers and deleted 5 seconds later.
Screenshots of my ex's private Instagram stories I saved and permanently deleted in 2023.
Some were full resolution. All organized by date going back to 2021.
When you delete a photo on iPhone, iOS removes it from your camera roll and Recently Deleted. It keeps a compressed copy in a Spotlight search index cache so your phone can find things when you search. Apple never clears it. Not after 30 days. Not after a year. Never.
They told you it was permanently deleted. It wasn't.
I checked which apps on my phone had Files and Folders access. 23 apps. Weather apps. QR code scanners. PDF readers. A flashlight app I downloaded in 2022. Any one of them could have been reading every photo I ever deleted this entire time.
Went from 23 apps to 3 in about two minutes.
Full step by step below. Takes 2 minutes. Save it so you don't lose it and send it to anyone with an iPhone.
FocalZero retweeted
👋🏼ADIÓS AL WEBSCRAPING!!
Han creado PixelRAG.
En vez de convertir una web en texto...
hace una captura de pantalla.
Y luego un modelo vision-language lee la respuesta directamente desde los píxeles.
Brutal.
Porque el scraping tradicional rompe muchísima información:
→ PDFs
→ tablas
→ layouts
→ botones
→ columnas
Todo eso se pierde cuando conviertes una web en texto plano.
PixelRAG hace algo mucho más simple:
indexa la página tal y como la ve una persona.
El equipo ha creado un índice visual de Wikipedia entero:
+30 millones de screenshots.
Y aun así supera al mejor RAG basado en texto por un 18,1% en QA.
Tiene también un plugin para Claude Code.
Básicamente le da ojos a Claude.
Le pasas una URL, un paper de arXiv o tu web local...
y Claude no scrapea el DOM.
La mira.
Cómo funciona:
→ construye un índice FAISS
→ busca visualmente sobre píxeles
→ los embebe con Qwen3-VL-Embedding
→ renderiza webs, PDFs e imágenes en tiles
Todo open source bajo Apache-2.0.
Link al repo en los comentarios 👇🏻
Je ne juge pas si c’est bien, mais c’est la direction prise. Personne ne veut d’une id digitale contrôlée, auditée à distance, régulée par un acteur incapable de sécuriser l’ensemble, ni d’un avatar créé sans consentement. C’est déjà hors de contrôle, simplement masqué des médias
Anthropic 工程师自曝:我们内部已经不怎么写 Prompt 了!写的是循环。!
白天 Agent 自己往 memory/ 里记笔记,
夜里系统自动“做梦”,核对、整理、去重。
第二天起来,它比昨天聪明一点。
结果?一次通过错误少了 97%。
Prompt 是一次性的,但循环,是能持续进化的。
真正拉开代际的,不是你会不会提问,
而是你的 Agent 睡醒后,还记不记得昨天踩过的坑。
#AIAgent #Claude #Anthropic #AI进化 #Agent架构 #克劳德 #提示词 #循环 #AI记忆
xitter.cf/nft_chen/status/211137…
💥绝美!Opus 5.5 + threejs 三句话做出的游戏现在会跳舞了!优化版画面美到窒息!
Opus 5.5 只用极少提示词就搓出了这样一款完整可玩的动漫风探索游戏,还加上了跳舞功能。
🔸极简提示词生成:从场景搭建、角色动作到交互逻辑全部由Opus 5.5搞定。
🔸角色与场景质感:白T黑短裤长发少女,日式便利店外景、雨夜高架步道、店内货架通道,全部高度还原,光影、反射和氛围感极强。
🔸新增跳舞功能:一键触发舞蹈动作,动作来源为临时映像素材,角色可在场景中自然跳舞,互动感直接升级。
🔸版本优化:针对性修复了上一版频闪问题,画面更稳定流畅,完成度也更高。
可玩性亮点:支持自由行走探索,场景切换自然,UI简洁底部跳舞按钮,纯代码实现的浏览器级游戏体验。
这种“几句话→完整可玩小游戏+动作升级”的能力,直接展示了Opus 5.5在游戏开发上的爆发力。
谁还在手搓引擎?!
#Opus55 #AI游戏开发 #Claude #游戏制作 #AI生成游戏 #三句话做游戏 #动漫游戏 #游戏修复版 #跳舞功能 #Claude #Anthropic #Dario
xitter.cf/nft_chen/status/211014…
太夸张了!一个人指挥 50 多个 AI Agent,干了 20 个工程师的活!😱
一位曾在 Cursor 工作的 SpaceXAI 高级工程师,分享了她的 AI 自动化工作流:
每天早上喝杯咖啡,就能管理整个 AI 团队。
她的配置是:
1 个 AI 总管:统筹所有工作
3 个 AI 项目经理:分配任务、协调进度
40 多个 AI 员工:通宵自动执行任务
她还声称,97% 的 SpaceXAI 工程师已经在使用类似的工作方式。
她用一场 60 分钟的教程,从零教你搭建整套系统。
以前有人收费 5000 美元教的东西,现在直接公开了。
以后一个人就是一家软件公司,真不是开玩笑!🤣
建议收藏!


















