Vidéo IA

Anthropic AI Writes 80% of Its Own Code: Recursive Self-Improvement

Anthropic AI Writes 80% of Its Own Code: Recursive Self-Improvement

Anthropic's AI, specifically Claude, now writes more than 80% of the code integrated into Anthropic's codebase, according to internal data from May...

Read
Anthropic IPO Warning: Pause Frontier AI Development Now

Anthropic IPO Warning: Pause Frontier AI Development Now

Anthropic has issued a stark warning: the world must hit pause on frontier AI development. This call comes as the company confirms its IPO filing, ...

Read
Australia Mandates AI Reporting After OpenAI Hack: What It Means

Australia Mandates AI Reporting After OpenAI Hack: What It Means

Australia is set to introduce mandatory reporting for artificial intelligence incidents following a hack involving OpenAI, as the government moves ...

Read
China's Robot Hotel: Full Automation with Pudu Robotics

China's Robot Hotel: Full Automation with Pudu Robotics

China is preparing a robot hotel where machines will handle almost everything, from reception and room service to cleaning and security. The projec...

Read
DeepSeek Gains Ground Among US Companies as AI Costs Bite

DeepSeek Gains Ground Among US Companies as AI Costs Bite

DeepSeek, the Chinese AI provider, has achieved something that seemed unthinkable just months ago: gaining traction among US companies. According t...

Read
Anthropic Launches Claude Fable 5.2 and Mythos 5.2 AI Models

Anthropic Launches Claude Fable 5.2 and Mythos 5.2 AI Models

Anthropic has launched Claude Fable 5.2 and Claude Mythos 5.2, its two most advanced AI models to date, designed to tackle highly complex technical...

Read
AI Debt Collectors: The End of the Repo Man

AI Debt Collectors: The End of the Repo Man

AI is replacing one of the most hated jobs in the world: the debt collector. In Spain, the sight of a man in a tailcoat and top hat following someo...

Read
How I Used AI to Build a Custom Android Debloater in 30 Minutes

How I Used AI to Build a Custom Android Debloater in 30 Minutes

If you've ever bought a new Android phone, you know the frustration of pre-installed apps you can't remove. I used AI to build a custom Android deb...

Read
AI Agents Booked Only 4 Uber Rides and 500 Eats Orders on Halloween

AI Agents Booked Only 4 Uber Rides and 500 Eats Orders on Halloween

During Halloween week, Uber processed about 320 million trips and orders, but only four rides and 500 Uber Eats orders came from AI agents, accordi...

Read
Alice Huang: The Woman Linked to Nvidia CEO Jensen Huang

Alice Huang: The Woman Linked to Nvidia CEO Jensen Huang

Alice Huang, the woman reportedly linked to Nvidia CEO Jensen Huang, has become the center of a growing controversy involving allegations of espion...

Read

SPONSORED

Apple's China-Specific AI Model: A Dual Strategy for iPhone
Apple's China-Specific AI Model: A Dual Strategy for iPhone
Apple has trained a large language model specifically for the Chinese market, marking a significant shift in its AI strategy for iPhone. According to Reuters, the company developed this model... Lire plus...
Data Center Boom in Quincy, Washington: A Revival Story
Data Center Boom in Quincy, Washington: A Revival Story
The data center boom in Quincy, Washington, offers a rare counter-narrative to the growing backlash against AI infrastructure. While many communities protest new data centers over energy and water consumption,... Lire plus...
China's AI Gains Ground as Mistral Hosts GLM-5.2 Model
China's AI Gains Ground as Mistral Hosts GLM-5.2 Model
China's AI is now impossible to ignore, and Mistral's latest move proves it: the French AI champion has begun hosting GLM-5.2, a model developed by China's Z.ai, on its platform.... Lire plus...
Z.ai GLM-5.3 Coding Gains Boost Cybersecurity Skills Fast
Z.ai GLM-5.3 Coding Gains Boost Cybersecurity Skills Fast
Z.ai trained GLM-5.3 to code better, but the model also became much stronger at cybersecurity tasks—an unexpected side effect with significant implications. According to Z.ai, as the company scaled post-training... Lire plus...
Anthropic's Claude Watermark: How It Works and Why It Matters
Anthropic's Claude Watermark: How It Works and Why It Matters
Anthropic's new watermark for Claude is an invisible mark embedded in the AI's generated text, designed to indicate its origin without altering the content. This move comes as part of... Lire plus...
Anthropic's Unusual Power System to Keep Control After IPO
Anthropic's Unusual Power System to Keep Control After IPO
Anthropic's unusual power system to keep control after its IPO is a bold attempt by its founders to preserve the company's public benefit mission. The AI lab, founded in 2021,... Lire plus...
OpenAI Claims Breakthroughs on Three Complex Scientific Problems
OpenAI Claims Breakthroughs on Three Complex Scientific Problems
OpenAI has announced that its internal frontier AI model has made progress on three highly complex scientific problems, spanning mathematics, theoretical computer science, and physics. The claim comes just weeks... Lire plus...
AI and Developer Mental Health: The Identity Crisis in Tech
AI and Developer Mental Health: The Identity Crisis in Tech
AI is taking a toll on developer mental health, with many programmers reporting an identity crisis as automation reshapes their roles. The issue has moved from private anxiety to public... Lire plus...
Terafab: Elon Musk's Gigafactory Larger Than Pentagon and Apple Park
Terafab: Elon Musk's Gigafactory Larger Than Pentagon and Apple Park
Terafab, Elon Musk's proposed chip manufacturing plant, would be larger than the Pentagon, Apple Park, and Giga Texas combined. The project, spanning roughly 100 million square feet (about 9.3 million... Lire plus...
Prime's AI-Generated Castle Walls Is a Smeared Mess
Prime's AI-Generated Castle Walls Is a Smeared Mess
Castle Walls, Prime Video's first fully AI-generated series, has arrived, and it's a Vaseline-smeared mess. Produced by Turkish studio Ay Yapım using a mix of generative video tools, the show... Lire plus...
AirTag Reveals AI Book Buying: Amazon Facility Exposed
AirTag Reveals AI Book Buying: Amazon Facility Exposed
An AirTag hidden in a book shipment has exposed how AI companies are buying up rare and antique books in bulk to feed their training data. The tracker led straight... Lire plus...
Ilya Sutskever's SSI: The Real AI Revolution Beyond OpenAI
Ilya Sutskever's SSI: The Real AI Revolution Beyond OpenAI
Ilya Sutskever, co-founder and former chief scientist of OpenAI, is now at the helm of Safe Superintelligence Inc. (SSI), a startup that aims to build a safe superintelligent AI. While... Lire plus...

SPONSORED