Meta AI News - Page 5 of 13

The model knows the numbers. It has seen them before, during training, embedded in the weighty folds of its neural architecture.

Sixty times fewer parameters actually doing work. That is the headline math behind OpenAI’s newly released privacy filter: a 1.5-billion-parameter...

The promise is infinite scale. The power to run it is not. America's biggest tech companies—Amazon, Google, Meta, and Microsoft—have committed more...

Anthropic’s CEO is sounding the alarm just as Claude gets its hands on the apps businesses live inside.

Washington's focus is locked on viral deepfakes. Yet the greater danger, analysts warn, isn't flashy content.

Mark Zuckerberg is building a synthetic version of himself for the office. This isn't a PR stunt or a customer service chatbot.

Google Research trained a foundation model on more than one trillion minutes of wearable sensor data pulled from five million Fitbit and Pixel Watch...

The exit of OpenAI’s robotics lead wasn’t just a resignation, it was a signal. The departure came as the company deepened its ties with the Pentagon,...

Spain just lit a match under the data world. A new law demands that any output, dashboards, BI reports, slide decks, created or substantially...

Meta plans to ship two new AI models this quarter, internally codenamed Avocado and Mango.

For eight years, Astryx lived inside Meta, quietly powering over 13,000 apps. Now it’s open-source, and the company behind it just gave the design...

For years, your Facebook profile picture has been a silent, static sentinel, a frozen version of you that never blinks or celebrates.

A couple of months ago, a cancer diagnosis turned one person’s world into a blur of appointments, scans, and medical jargon.

Meta’s internal AI leaderboard is a perfect, stupid idea. It ranks engineers by how many AI tokens they generate.

Mark Zuckerberg is building a personal AI agent to bypass his own chain of command. No more waiting for middle managers. No more layered briefs.

When a RAG system sounds utterly certain but its answer is a mirage, you’ve hit the most dangerous failure mode: high confidence paired with low...

NVIDIA Blackwell delivers 15 petaflops of dense NVFP4 compute, a staggering amount of raw power.

There’s a graveyard of AI agents out there. Ninety percent of them, according to Fetch.ai CEO Humayun Sheikh, never get used.

Meta’s Muse Spark does something head-turning: it matches the capabilities of Llama 4 Maverick while using more than ten times less compute.

ChatGPT briefly, bizarrely, fell for goblins. The whole fling started with a single broken piece of code.