Google AI News - Page 9 of 33

Small AI models are useful. They're also famously stupid. We stick them in phones and browsers, but rarely ask them to think.

Look at Meta’s marketing now, and the word “smart” has vanished. They’re AI glasses, full stop.

**The old way of generating text is a bottleneck.** Token by token, each word waiting on the last.

Google's latest large language model is officially the best liar in the business.

For three decades, the "Conway knot problem" stumped mathematicians. They declared it unsolvable. Then, last week, Google DeepMind's AI cracked it.

Auto-Diagnose is not gentle. Of 517 feedback reports, 436 came back with a blunt directive: “Please fix.” That’s 84.3%, an overwhelming majority.

Elon Musk lost a bid to keep xAI's safety records sealed. The ruling came from a Delaware court on Thursday, the same week Musk took the stand in his...

Silicon Valley’s relentless focus on giants like OpenAI and Google often overshadows the audacious few who dare to build from the ground up.

Google Chrome owns the web. For more than a decade, its 60 percent market share has seemed like a law of physics, not a business strategy.

Google is advancing medical AI on two fronts, and making both openly available. The company’s updated MedGemma model now processes CT scans and MRIs...

Google's Project Genie can build a video game from a text prompt. The result will be bad.

Broadcom reported a $21 billion order from Anthropic for Google’s seventh-generation Tensor Processing Units.

Anthropic just gave its beloved Opus 3 a retirement gig: a blog post penned by the model itself. That’s not a gimmick, it’s a signal.

Salesforce just pulled the rug out from under the traditional CRM interface. The company’s new Headless 360 and Agentforce Layer aren’t incremental...

The voice you hear from an AI isn’t just a technical output anymore, it’s a relationship.

The AI image generation race just got a lot more interesting. Luma AI’s Uni-1 doesn’t just match Google and OpenAI on benchmarks, it outperforms...

Kagi built its name on being the anti-Google: a paid search engine that actually respects your privacy and doesn't drown results in ads.

Every new AI model claims to beat the last. This one might. Z.ai's open-source GLM-5 just posted a record-low hallucination rate.

Imagine a machine that meets the clinical criteria for autism, dissociation, and trauma-related shame all at once.

In a Swedish classroom, a teacher just reclaimed hours once lost to lesson planning, and students are gaining fluency in an AI-shaped world.