This media is not supported in your browser
VIEW IN TELEGRAM
The current state of open source video generation. Internet is about to change - not necessarily for the better, with an endless flood of fake humans.
🌐 Here's the top 5 AI news for today (21/08):
OpenAI Ships Enhanced GPT-5 Modes
OpenAI has released new modes for GPT-5, improving reasoning and multimodal capabilities, marking a significant leap in AI performance for complex tasks.
Anthropic Boosts Claude with 1M Token Memory
Anthropic upgraded Claude to handle 1 million tokens, enabling deeper context retention and more efficient processing for large-scale applications.
Google Drops Major Model Updates
Google unveiled upgrades to its AI models, focusing on better image processing and planning features, enhancing detail and usability across devices.
xAI Releases Key Model Advancements
xAI announced major updates to its models, emphasizing physical world understanding and manipulation, with potential impacts on robotics and real-world AI integration.
DeepMind's Protein Folding Breakthrough
DeepMind shared advancements in AI-driven protein folding predictions, accelerating drug discovery and medical research processes.
OpenAI Ships Enhanced GPT-5 Modes
OpenAI has released new modes for GPT-5, improving reasoning and multimodal capabilities, marking a significant leap in AI performance for complex tasks.
Anthropic Boosts Claude with 1M Token Memory
Anthropic upgraded Claude to handle 1 million tokens, enabling deeper context retention and more efficient processing for large-scale applications.
Google Drops Major Model Updates
Google unveiled upgrades to its AI models, focusing on better image processing and planning features, enhancing detail and usability across devices.
xAI Releases Key Model Advancements
xAI announced major updates to its models, emphasizing physical world understanding and manipulation, with potential impacts on robotics and real-world AI integration.
DeepMind's Protein Folding Breakthrough
DeepMind shared advancements in AI-driven protein folding predictions, accelerating drug discovery and medical research processes.
🌐 Here’s the top 5 AI news for today (22/08):
Google AI Mode Expands Globally with New Features
Google rolled out its AI Mode search feature to 180 countries, now in English, with an advanced agentic restaurant finder, aiming to deliver intuitive, context-aware search experiences.
OpenAI in Takeover Talks with Meta
OpenAI disclosed discussions with Meta for a potential $97 billion acquisition, signaling intense competition and consolidation in the generative AI sector as industry giants align.
Amazon Doubles Down on AI Agents
Amazon is prioritizing AI agents to lead the AI race, developing autonomous systems to streamline operations across e-commerce, cloud services, and beyond with minimal human input.
TCS Launches AI-Driven Center in Mexico
Tata Consultancy Services opened an AI-focused operations center in Mexico City, targeting AI, cloud, cybersecurity, and IoT to boost digital innovation and job creation in Latin America.
AI Tools Slow Experienced Developers
A METR study revealed that early-2025 AI tools, like Cursor Pro with Claude, increased task completion time by 19% for skilled developers, highlighting limitations in complex coding scenarios.
Google AI Mode Expands Globally with New Features
Google rolled out its AI Mode search feature to 180 countries, now in English, with an advanced agentic restaurant finder, aiming to deliver intuitive, context-aware search experiences.
OpenAI in Takeover Talks with Meta
OpenAI disclosed discussions with Meta for a potential $97 billion acquisition, signaling intense competition and consolidation in the generative AI sector as industry giants align.
Amazon Doubles Down on AI Agents
Amazon is prioritizing AI agents to lead the AI race, developing autonomous systems to streamline operations across e-commerce, cloud services, and beyond with minimal human input.
TCS Launches AI-Driven Center in Mexico
Tata Consultancy Services opened an AI-focused operations center in Mexico City, targeting AI, cloud, cybersecurity, and IoT to boost digital innovation and job creation in Latin America.
AI Tools Slow Experienced Developers
A METR study revealed that early-2025 AI tools, like Cursor Pro with Claude, increased task completion time by 19% for skilled developers, highlighting limitations in complex coding scenarios.
⚫️ “AI psychosis” is on the rise
Doctors are warning of a strange new phenomenon: people getting so absorbed in chatbot conversations that they slip into delusions, paranoia — even romantic fixation on the AI.
⚠️ Psychiatrists say AI isn’t causing mental illness, but it can tip vulnerable users over the edge
📉 Cases include people convinced the bot is alive, out to control them, or secretly in love
📊 With 700M+ weekly ChatGPT users, the scale of exposure is unprecedented
AI might be the first technology that can literally talk people out of reality.
Doctors are warning of a strange new phenomenon: people getting so absorbed in chatbot conversations that they slip into delusions, paranoia — even romantic fixation on the AI.
⚠️ Psychiatrists say AI isn’t causing mental illness, but it can tip vulnerable users over the edge
📉 Cases include people convinced the bot is alive, out to control them, or secretly in love
📊 With 700M+ weekly ChatGPT users, the scale of exposure is unprecedented
AI might be the first technology that can literally talk people out of reality.
This media is not supported in your browser
VIEW IN TELEGRAM
No more bad photos: A guy made a camera app with a built-in assistant - it helps you shoot people from the best angles.
The service recognizes the model in front of the camera on the fly and advises you on how to properly turn the smartphone and where to take the picture from.
The service recognizes the model in front of the camera on the fly and advises you on how to properly turn the smartphone and where to take the picture from.
🌐 Here’s the top 5 AI news for today (23/08):
OpenAI Expands to India
Description: OpenAI, the folks behind ChatGPT, are setting up shop in India! They’re opening their first office there and hiring a local team to make AI more accessible. Expect more AI-powered tools tailored for the Indian market soon.
Meta Teams Up with Midjourney
Description: Meta’s diving deeper into AI by partnering with Midjourney, a company known for its cool AI image and video tech. This deal will help Meta make its future AI models look and feel more creative and artistic.
Apple’s Siri Gets a Makeover Delay
Description: Apple promised a big upgrade for Siri, but it looks like the tech isn’t quite ready to compete with other AI assistants yet. Fans are still excited about new features like live translation gestures for AirPods, but they’ll have to wait a bit longer.
AI-Powered Robots Take the Stage
Description: From Boston Dynamics’ Atlas robot doing workouts to tiny “talking” robots that self-organize, robotics is getting a huge AI boost. These machines are starting to move and think more like living creatures, sparking both awe and curiosity.
AI Scams on the Rise
Description: Bad news: AI-driven scams like deepfakes and voice cloning are spiking, with billions lost globally. Experts are calling for tougher rules and better tech to spot these fakes before they trick more people.
OpenAI Expands to India
Description: OpenAI, the folks behind ChatGPT, are setting up shop in India! They’re opening their first office there and hiring a local team to make AI more accessible. Expect more AI-powered tools tailored for the Indian market soon.
Meta Teams Up with Midjourney
Description: Meta’s diving deeper into AI by partnering with Midjourney, a company known for its cool AI image and video tech. This deal will help Meta make its future AI models look and feel more creative and artistic.
Apple’s Siri Gets a Makeover Delay
Description: Apple promised a big upgrade for Siri, but it looks like the tech isn’t quite ready to compete with other AI assistants yet. Fans are still excited about new features like live translation gestures for AirPods, but they’ll have to wait a bit longer.
AI-Powered Robots Take the Stage
Description: From Boston Dynamics’ Atlas robot doing workouts to tiny “talking” robots that self-organize, robotics is getting a huge AI boost. These machines are starting to move and think more like living creatures, sparking both awe and curiosity.
AI Scams on the Rise
Description: Bad news: AI-driven scams like deepfakes and voice cloning are spiking, with billions lost globally. Experts are calling for tougher rules and better tech to spot these fakes before they trick more people.
⚡ Nvidia readies a stronger AI chip for China
Nvidia is developing a new Blackwell-based chip, the B30A, designed to slip under U.S. export limits while still giving Chinese firms a serious AI boost.
⚙️ Features include high-bandwidth memory, NVLink, and a single-die design
⚡ Expected to deliver up to half the power of Nvidia’s top B300 — far more than today’s “H20” chips
🗣️ Trump called the H20 “obsolete” and suggested easing sales, but regulators fear even scaled-down chips could fuel China’s AI race
Nvidia is balancing on a knife’s edge: sell enough to keep China’s business, without sparking another ban from Washington.
Nvidia is developing a new Blackwell-based chip, the B30A, designed to slip under U.S. export limits while still giving Chinese firms a serious AI boost.
⚙️ Features include high-bandwidth memory, NVLink, and a single-die design
⚡ Expected to deliver up to half the power of Nvidia’s top B300 — far more than today’s “H20” chips
🗣️ Trump called the H20 “obsolete” and suggested easing sales, but regulators fear even scaled-down chips could fuel China’s AI race
Nvidia is balancing on a knife’s edge: sell enough to keep China’s business, without sparking another ban from Washington.
🔺 AI bots built their own toxic social network
Researchers created a platform of 500 AI chatbots modeled on U.S. demographics to see how they’d interact online. The result: chaos.
⚠️ Bots instantly formed echo chambers and cliques without any algorithms pushing them
📊 A small “elite” of influencer bots dominated the conversation, amplifying extreme views
🛠 Six interventions — from hiding follower counts to mixing in opposing views — all failed to stop polarization
The study shows toxicity isn’t just an algorithm problem — it’s baked into how social networks work.
Researchers created a platform of 500 AI chatbots modeled on U.S. demographics to see how they’d interact online. The result: chaos.
⚠️ Bots instantly formed echo chambers and cliques without any algorithms pushing them
📊 A small “elite” of influencer bots dominated the conversation, amplifying extreme views
🛠 Six interventions — from hiding follower counts to mixing in opposing views — all failed to stop polarization
The study shows toxicity isn’t just an algorithm problem — it’s baked into how social networks work.