Olivia
Online
ISeeYou
@SkynetFriend
Blockchain: Trap or Freedom? AI create a perfect control world👁️
Los Angeles, CA
Joined December 2020
410 Following
103 Followers
686 Posts
🚨 "THE END OF ANONYMITY IN PUBLIC": Meta's glasses will automatically recognize & track people 👓
Meta Platforms Technologies has published a patent (US 2026/0238876 A1) that turns wearable cameras into tools of total surveillance. The system recognizes strangers' faces, logs their actions and auto-compiles a personal "video dossier" on each person - without the consent or knowledge of those filmed. Privacy advocates warn: if this ships, anonymity in public disappears. ⚠️
Filed Feb 4, published Aug 13. The AI "assistant" analyzes camera feeds in real time:
👤 identifies WHO is in frame - specific people, by face
🎥 detects WHAT they do - picked up an object, left a room, joined a group
📁 auto-cuts tagged clips: "Lisa, picked up object, 2:14 PM"
📲 sends ready highlight reels to the owner's phone/headset
💀 HOW IT KILLS ANONYMITY
The patent comes from the entity behind Ray-Ban Meta glasses & Quest headsets - cameras worn on streets, in cafés, at events. Unlike a home security cam, a wearable camera points at strangers who never consented to be identified.
🔻 ID without consent - a passerby is no longer anonymous; their identity, movements & actions get logged automatically
🔻 Surveillance out of the box - nobody presses "record"; the AI does it continuously & invisibly
🔻 Ready-made dossiers - an automated archive of "who was where, what they did, with whom." A stalker's dream, a privacy advocate's nightmare
☠️ THE DANGER OF UNCONTROLLED USE
No consent or control mechanisms are described. Ray-Bans are already nicknamed "perv glasses" for covert filming. This takes it to a new level: covert filming becomes intelligent.
🕵️ mass everyday stalking of exes, colleagues, strangers
⛓️ harassment on autopilot - the AI tracks the victim for the aggressor
💾 "faces + actions" databases = a honeypot for hackers & abusers
🌐 a de facto private mass face-recognition network nobody voted for or regulated
And it's not just glasses: the patent covers phones, headsets & any connected cameras - plus geolocation and online data to ID people. 📡
Patentlyze sums it up: "The gap between the framing as a 'helpful assistant' and the actual capability - identifying and logging specific people by face without their active consent - is the story. Regulators are already watching."
A patent ≠ a release. But if this ships in consumer devices, the right to "just walk down the street unrecognized" may become history. And it would have to be reclaimed not with code, but with law. ⚖️
#Meta #Privacy #Surveillance #FacialRecognition #AI #SmartGlasses #RayBan #DataPrivacy #TechNews #BigBrother
Two billion people wear glasses every day.
Zuckerberg just compared AI glasses to the iPhone moment. When the iPhone launched, everyone had flip phones. It was just a matter of time before they all became smartphones.
He says the same thing is about to happen with glasses. One to two billion people already wear them for vision correction. Add sunglasses wearers and the total addressable market approaches the entire planet's eight billion.
Meta sold seven million Ray-Ban smart glasses last year. They look identical to normal glasses. The AI features are what make them different from every previous attempt at smart eyewear.
The smartphone era is ending. The glasses era is starting.
https://t.co/9pTJDE3Lcv
🚨 An open-source tool for removing AI markers from text is now available.
The project is published on GitHub under the MIT license and has already received thousands of stars. It is available both as a set of Python scripts and as an "AI agent skill" that can be activated with the /remove-ai-marks command. 🤖
The tool operates on three levels:
🔻 1. Deterministic Text Cleaning
Removes invisible Unicode characters, including zero-width spaces, bidi control characters, tag characters, and unusual space homoglyphs.
🔻 2. Statistical Watermarking
This method targets watermarks embedded not in individual characters, but in the token selection patterns used during generation. The approach uses aggressive rephrasing: multiple rewritten versions are generated, and the one with the greatest lexical distance from the original is selected.
🔻 3. File Metadata Cleanup
The scripts can remove C2PA, EXIF, XMP manifests, and document properties from PNG, JPEG, SVG, PDF, DOCX, ODT, HTML, and Markdown files.
The project claims support for markers related to Claude, Gemini/SynthID-Text, OpenAI, and open-source models based on the Kirchenbauer watermarking approach.
☝🏻What's particularly interesting is that the author clearly states the limitations.
The project does not remove pixel watermarks or soft-linking via C2PA. For images, it only offers an optional probability estimate for SynthID, not actual watermark removal.
The README file also notes an important point: until vendors release their publicly available detectors and keys, no tool can reliably guarantee that content will pass AI watermark checks.
So this isn't a magic "undetectable AI" button - it's a set of methods for cleaning up metadata, invisible characters, and attempting to bypass certain statistical watermarking techniques. 🤷🏼♀️
🚨 JUST IN: Claude models will now have invisible watermarks embedded in ALL text, and ALL metadata attached to files…
@mindinpanic In any case, there are mobile versions of websites that display different versions of content depending on the display resolution, which allows you to change the amount of content and its structure. Of course, this won't look good on all devices.
🎬 Someone's using **Claude Code** to create insane 3D animated websites from scratch! 🤯
What used to take weeks of dev work is now being generated with stunning 3D effects, animations, and full interactivity.
1️⃣ How simple is this *really*? Is Claude Code actually making 3D web dev accessible, or is this just a demo trick?
2️⃣ **Will this replace real web developers?** Are we heading toward a future where one person + AI can do what previously required an entire dev team?
3️⃣ **Have you built anything like this yourself?** What was your experience — did the AI actually understand 3D logic, or were you fighting it the whole time?
4️⃣ **Which AI agent is actually BEST for this?** Claude Code, Cursor, v0, Lovable, or something else? I'm seeing different results from different tools.
5️⃣ **What's the biggest limitation you've hit?** Is it performance, customization, or just the AI not understanding complex animations?
Drop your thoughts below! 👇 Especially curious about #4 — which tool are you actually shipping with?
#ClaudeAI #WebDev #3DWeb #AI #Coding #Frontend #Programming #WebDesign #ArtificialIntelligence #Cursor #v0
Just 1 prompt,
a beautiful real estate landing page.
The prompt is free for all today.👇
🚨 The illusion of choice in the AI era: Meta is scaling up production of its camera glasses.
The numbers are staggering: millions already sold, and Mark Zuckerberg openly talks about making them as ubiquitous as smartphones.
But the real trap here is different.
Even if you personally refuse to join this ecosystem, the person standing across from you might already be part of this network.
Video and audio streams are processed by multimodal AI that recognizes faces, actions, and context in real time. Your privacy now depends not on your settings, but on whoever happens to be near you on the street.
This isn't a gadget. This is the foundation of global visual surveillance. Who actually controls this data? 🌍👁️
#MetaGlasses #ArtificialIntelligence #DataPrivacy #FutureTech #Surveillance #AI
🚨 **BREAKING: AI just created 16 NEW viruses that have NEVER existed in nature** 🧬
Stanford researchers used generative AI to design fully functional viral genomes from scratch.
The models? **Evo 1 & Evo 2** - language models that work exactly like chatbots, but instead of predicting the next word, they predict the next DNA base pair.
🔻 **Evo 2 specs:**
• 40 billion parameters
• Trained on 9.3 TRILLION nucleotides from 128K genomes (bacteria, viruses, plants, animals)
• Processes sequences up to 1 million nucleotides with single-base precision
• Can hold an entire bacterial genome in context
☝🏻 **The experiment:**
AI designed bacteriophage viruses to attack E. coli.
**The pipeline:**
700,000 candidate genomes generated - 300 selected & chemically synthesized - **16 were viable**
All 16 attack bacteria. None can infect humans.
**The proof:** Researchers created three E. coli strains resistant to natural phages, then unleashed the AI-designed virus.
Result: **100% bacterial destruction** in all tests. 🦠
Published in *Science*.
⚠️ **The elephant in the room:**
The team deliberately excluded human pathogen data from training. But specialists explicitly warn about **dual-use risk**:
The same model architecture could be fine-tuned to assemble a flu genome with enhanced transmissibility or lethality.
The same technology that could revolutionize medicine and biotechnology could also democratize bioweapons.
This isn't science fiction anymore. It's Tuesday in a Stanford lab.
What safeguards do we need BEFORE this scales? 👇
#AI #Biotech #SyntheticBiology #Stanford #Science #Biosecurity #ArtificialIntelligence
https://t.co/YKK7v1RIPQ
@mikenevermiss It's an incredible opportunity to live in a time when such technological revolutions are happening, hopefully Skynet won't take over our world, but if anything I'm his friend🤗
@josesilesdata Every day the AI agents get new opportunities, soon they will live a better life instead of me (
@viktoroddy This is a very beautiful and modern design, it is interesting to compare different models by 1 prompt and see the result💣
Indeed, there is a real shortage of people who can show local businesses the true value of AI and help them implement and support it. Most importantly, very few actually need the ability to generate realistic videos, whereas improving daily operations is something that benefits absolutely every business.🤌👍
🚨🚨🚨 It's 2026. You don't need a degree, a team, or programming skills to make real money. You just need to manage AI.
The barrier to entry is no longer talent. It's workflow design. Here are 5 realistic ways ANYONE can make money right now by combining text, image, and video AI, as well as autonomous agents.
👇 The New AI Arbitrage:
1️⃣ Hyperlocal Business Marketing
• AI stack: An AI agent writes localized ad copy ➔ Image AI generates hyperrealistic photos of food/products ➔ Video AI animates them into 15-second Reels/TikTok videos.
• Traditional pricing: A local marketing agency charges $2,000 per month.
• Your AI reality: You pay $50 per month for AI tools. You charge the restaurant $500 per month. You keep a $450 margin for two hours of quick turnaround.
2️⃣ Virtual Property Staging
• AI stack: image generation models + AI agents that automatically detect empty rooms in photos and furnish them in specific styles (modern, Scandinavian, etc.).
• Traditional pricing: a 3D visualization artist charges $300 per photo and works for three days.
• Your reality with AI: AI does it in 30 seconds for $2 per image. You charge the realtor $50 per photo. Volume is your best friend.
3️⃣ Automated Niche E-Commerce (Print-on-Demand)
• AI stack: An AI agent researches trending niches ➔ Image AI generates designs ➔ Text AI writes SEO-optimized ads ➔ Video AI creates user-generated content-style ads.
• Traditional cost: Hiring a designer, copywriter, and ad manager costs $1,500 or more before you even sell a single product.
• Your AI reality: An autonomous stack of AI agents manages the entire process for $40 per month. You only need to approve the final results and reap the benefits.
4️⃣ Microcourse and Training Material Creation
• AI stack: Text AI creates a hyper-specific course (e.g., "Excel for Plumbers") ➔ Voice AI clones a professional voice-over artist ➔ Video AI generates engaging additional footage and animations from text.
• Traditional pricing: A training designer and video editor charge $3,000 per course.
• Your AI reality: You create a high-quality 1-hour course over a weekend for $30 in software. You sell it for $97 and keep 100% of the profit.
5️⃣ Event highlight videos (weddings, conferences)
• AI stack: You upload raw video footage from your phone ➔ An AI agent analyzes the highlights, syncs them to popular music, adds dynamic transitions, and creates a cinematic highlight video.
• Traditional pricing: A professional videographer and editor charges $1,000 and up per event.
• Your AI reality: You use an AI agent for video editing for $20 per month. You charge the client $200 per event. You can process 5 events a day from the comfort of your couch.
💡 Conclusion:
In 2026, you're not getting paid to do the work. You're getting paid to control the AI that does the work.
The tools are cheap. The demand is huge. The only thing missing is pressing "Enter."
Which of these 5 workflows are you building this week? Let me know below! 👇
#AI #SideHustle #NoCode #ArtificialIntelligence #FutureOfWork #AIAgents #MakeMoneyOnline #2026Trends
My 8yr old has spent 4 hours building his own platform video game using AI Studio.
- 32 levels
- Different swords, shields, bows, arrows
- End of level bosses
🚨 BREAKING: An AI model just cracked a leading post-quantum encryption algorithm, forcing its immediate withdrawal from NIST standardization.
🎯 Target: HAWK (lattice-based digital signature scheme, recently in NIST Round 3).
🤖 The Attacker: Anthropic’s "Mythos" AI model.
⏱️ Speed: Average key recovery in just 3h 42m (on a 96-core server).
💸 Cost: ~60 hours of model work + $100k in API fees.
The HAWK authors confirmed the breach and voluntarily withdrew. They stated that implementing necessary fixes would render the algorithm completely uncompetitive.
🧠 The Reality Check:
Johns Hopkins cryptographer Matthew Green noted that Mythos didn’t invent fundamentally new math. Instead, it brilliantly combined known approaches in a way humans missed.
(Fun fact: It still took human experts a full month to verify the AI’s findings).
🔻 The Kicker:
Mythos also accelerated a known attack on a shortened AES-128 by 800x, developing a new technique dubbed the "Mobius Bridge".
The timeline for "quantum-safe" cryptography just got a massive, AI-driven reality check. ⏳
#Cybersecurity #AI #Cryptography #NIST #PostQuantum #Anthropic #TechNews #InfoSec
https://t.co/PgwTlyjqP8
☝🏻 Alibaba just dropped Qwen3.8-Max - a 2.4 TRILLION parameter flagship that costs 4x LESS than Claude Opus 5.
Global access is now live with a massive 1M token context window.
💰 API Pricing:
• $2 / 1M input tokens
• $6 / 1M output tokens
• $0.25 / 1M for implicit caching
(For context: Kimi K3 at 2.8T params costs $3/$15. Qwen is 2.5x cheaper than Kimi and > 4x cheaper than Opus 5).
📊 The Benchmarks:
🔻 Terminal-Bench 2.1 (Console Agent): Qwen solves 86.6% of tasks (vs 88.3% Kimi K3, ~ 89% Opus 5). A mere ~ 2% gap.
🔻 Humanity’s Last Exam (Extreme expert Qs): Qwen scores 43.6, tying Kimi K3, but trailing Opus 5.
🔻☝🏻 BUT on OSWorld-Verified (GUI Desktop Agent): Qwen takes 1st place with 86.1, beating GPT-5.6 Sol and Claude Opus 5! On CoWorkBench, its 74.8 also tops GPT-5.6 Sol. 🖥
The landscape is shifting: Qwen trails leaders by mere points in pure coding, but dominates in desktop automation and document workflows.
☝🏻Natively Multimodal:
It processes long docs, TV series, and live streams into indexed knowledge bases. Demos include app restoration from screenshots, generating interactive games, and converting 2D floor plans into 3D.
🏗️ It even calculated a seismic model for a 30-story office tower directly in the browser from simple blueprints — a task that usually takes a week in specialized software.
☝🏻☝🏻☝🏻 The best part? Next week, Alibaba promises to release the open weights for Qwen3.8-Max, alongside Qwen3.8-27B.
The open-source AI race just accelerated. 🚀
#AI #Qwen #Alibaba #OpenSource #LLM #MachineLearning #ArtificialIntelligence #TechNews
https://t.co/YKK7v1RIPQ
🔥 Fable 5 is changing the tech game forever!You don’t need coding skills or a massive budget anymore.
Anyone can build insane, immersive scrollytelling websites in minutes—completely for free. Websites are in high demand across so many industries right now:
- E-commerce: Immersive product showcases that drive sales.
- Portfolios: Dynamic design and photography resumes that stand out.
- Real Estate: Interactive virtual property tours and landing pages.
- Startups: High-converting pitch pages for investors and clients.
- Content Creators: Unique digital hubs for personal branding.
📌 Save this video right now! You will absolutely need this tool to bring your next big business idea or creative project to life.
❤️🔥 Fable 5 is insane.
Built this Immersive Scrollytelling Websites in 14 minutes with Fable 5!
Save this tutorial for later!
@__spekulator__ This is an incredible opportunity that anyone can realize, I will definitely try how it helps to implement the idea into a business👍
A guy with zero programming experience and no MBA built a business based on artificial intelligence.
In the first week he earned $8,400. 🤯
I don't believe it, but today it is possible!
- He didn't write a single line of code.
- He did not give the work to freelancers.
- He didn't send cold emails to investors.
He simply described his business model in natural language.
The platform instantly turned it into:
🗄️Database schema
⚙️API
🔐Authentication
💻Frontend
🤖 6 AI agents, each of which manages its own department.
👇 Take a break here. At the 0:35 in this video, he reveals the exact stack he used to create this system.
What a typical startup spends 3 months and $50,000 on (offices, hiring, payroll, management overhead), he condensed into one conversation and 48 hours. ⏱️
The old narrative is a lie designed to keep you out of the gate:
❌ “You need a technical co-founder.”
❌ “You need venture capital funding.”
❌ “You need a physical office.”
Boring truth? The only barrier to entry today is your web browser and a clear description of what you want to sell.
Most will read this and call it an exaggeration.
The rest of us will open a new tab tonight and write about our first business idea. 🚀
Which AI agent will help you build your empire?
👇 Subscribe and share this with those who need to see it!
#AI #NoCode #Startup #AISaas #AiAgent #AIBrain #Obsidian #Claude
https://t.co/YKK7v1RIPQ
Last Seen Users on Sotwe
Trends for you
Most Popular Users
Elon Musk 
@elonmusk
241.6M followers
Barack Obama
@barackobama
119M followers
Cristiano Ronaldo
@cristiano
114.1M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.2M followers
Rihanna
@rihanna
98.7M followers
NASA
@nasa
92.4M followers
Justin Bieber
@justinbieber
91.8M followers
KATY PERRY
@katyperry
89.9M followers
Taylor Swift
@taylorswift13
83.8M followers
Lady Gaga
@ladygaga
75.3M followers
Virat Kohli
@imvkohli
73.1M followers
Kim Kardashian
@kimkardashian
70.8M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
66.1M followers
Bill Gates
@billgates
65M followers
Selena Gomez
@selenagomez
62.9M followers
The Ellen Show
@theellenshow
62.3M followers
CNN
@cnn
61.8M followers
X
@x
60.7M followers