X user "Jian" discovered that Manus appears to be using Claude Sonnet with access to 29 tools and the open-source software Browser Use. He found this by requesting the sandbox runtime code from Manus ...

Google is presenting AI functions for its Gemini assistant at the Mobile World Congress (MWC) in Barcelona. The company announced that subscribers to the Google One AI Premium plan for Gemini Advanced ...

The company's new Responses API combines features from its Chat Completions and Assistants APIs, offering developers a more flexible foundation for building agent applications with integrated web ...

A novel prompting method called "Highlighted Chain of Thought" (HoT) helps large language models better explain their reasoning and makes their answers easier for humans to verify. The approach works ...

Google's experimental AI mode pushes search toward a closed ecosystem, mirroring OpenAI's ChatGPT search and Perplexity while potentially delivering a fatal blow to the open web. The system, built on ...

The company plans to concentrate 85 percent of its capacity in the United States, with three facilities already operational in Santa Clara, Stockton, and Dallas. Additional centers will open in ...

As language models mature, Microsoft CEO Satya Nadella sees them becoming commodities. The shift marks a shift in how companies approach AI development, with systems integration and product ...

According to Reuters sources, Nvidia and Broadcom have begun testing Intel's cutting-edge 18A manufacturing process, putting it in direct competition with Taiwan-based TSMC's advanced technology.

Sony Music reports removing tens of thousands of AI-generated fake songs, highlighting growing concerns about artificial intelligence in the music industry. The company has taken down more than 75,000 ...

New research from OpenAI reveals how AI systems exhibit problematic reasoning patterns when "thinking" through tasks, warning against attempts to forcefully correct these behaviors. Large reasoning ...

Newly released data from Chinese AI provider Deepseek reveals that AI language models could, in theory, generate substantial profit margins—even at prices significantly lower than OpenAI’s. Deepseek ...

BIG-Bench, developed in 2021 as a universal benchmark for testing large language models, has reached its limits as current models achieve over 90% accuracy. In response, Google DeepMind has introduced ...