Google has developed a powerful internal AI agent named "Agent Smith," distinct from the publicly available Gemini. Built on the internal Antigravity platform, this AI can automate programming workflows, access internal documents, and even autonomously deploy to production environments.

Unlike conventional AI assistants that answer questions or generate text, Agent Smith functions as an autonomous agent, capable of planning and executing entire workflows without constant supervision. Its capabilities are so advanced that employees can issue instructions from their phones, and Smith will work independently.

Due to its high effectiveness and rapid adoption within the company, Google was forced to restrict access to Agent Smith because its infrastructure could not keep up with demand. This limitation was a result of its success, not a security concern.

As anticipated by Google co-founder Sergey Brin, who predicted AI agents would play a central role in the company's operations this year, Agent Smith serves as a concrete realization of this vision. This case highlights that the most advanced AI tools deployed by major tech companies internally are often far more sophisticated than those released to the public.

👉 Read the full article on Gizmodo en Español

The explosive growth of artificial intelligence demands systems capable of processing enormous volumes of data at unprecedented speeds, pushing current GPUs and chips to their physical limits. However, an international team of researchers has discovered a radically different computing method based on light, potentially ushering in a new era for AI technology.

This innovative approach leverages physical phenomena of light to process information in ways current electronic chips cannot replicate. Light's ability to execute multiple operations simultaneously is expected to perform tasks that currently require thousands of electronic operations in a single step.

A team led by Yufeng Zhang from Aalto University's Photonics Group developed a method where digital data is encoded into the amplitude and phase of light, allowing interacting light waves to naturally perform mathematical operations like matrix and tensor multiplications, which are fundamental to deep learning. This enables multiple calculations to occur simultaneously without friction or electronic heating.

This light-based computational framework is projected to be integrated into photonic chips, enabling complex AI tasks with minimal energy consumption. Researchers estimate this technology could be incorporated into existing platforms within 3 to 5 years, significantly impacting fields such as robotics, medicine, defense, astronomy, climate simulation, and generative AI.

👉 Read the full article on Gizmodo en Español

The content of the provided URL was inaccessible. However, the article's title suggests that Nvidia-backed SiFive has achieved a significant valuation of $3.65 billion in the open AI chip sector.

This likely reflects the growing importance of open-source chip designs, particularly those based on the RISC-V architecture, within the semiconductor industry for AI and machine learning applications. Support from a major player like Nvidia indicates strong confidence in SiFive's technology and market potential.

👉 Read the full article on TechCrunch