Google Creative Lab has developed the Gemma Translator, an experimental, open-source device capable of real-time translation without an internet connection. This device listens to conversations, converts them to text, translates them, and then plays them back in another language, all without connecting to the cloud.

The core of the system is a Raspberry Pi 5 with 8GB of RAM, powered by Gemma 4 E2B, one of Google's smaller language models. The entire translation process runs locally, utilizing LiteRT-LM, Google's infrastructure for executing generative models directly on hardware.

The translation workflow involves three steps: speech-to-text by Moonshine STT, text translation by Gemma, and text-to-speech by moonshine-voice TTS. Google has made the project's code, installation scripts, and 3D printing files for the casing available on GitHub, showcasing the potential of on-device AI processing.

👉 Read the full article on Gizmodo en Español

OpenAI has acquired NextSlide, a startup specializing in presentation technology. This acquisition signals a strategic move by OpenAI to expand its Generative AI and LLM capabilities into business applications, specifically in the realm of creating and enhancing presentations.

By integrating NextSlide's AI-powered presentation tools into OpenAI's ecosystem, users may be able to generate professional presentations more efficiently using LLMs and other AI models. This move indicates OpenAI's aim to deepen the practical application of its AI technology and strengthen its presence in the enterprise solutions market.

👉 Read the full article on TechCrunch

China has commercially approved its first brain chip implant, named NEO, which allows paralyzed individuals to control devices using brain signals. This approval marks a significant shift for Brain-Computer Interface (BCI) technology, moving it from experimental lab use to commercial application.

The NEO device, a small implant placed in the skull, collects signals from the brain area associated with movement. It interprets imagined actions (e.g., closing a hand) and translates them into commands for external devices, such as a robotic glove, enabling people with severe paralysis to perform basic tasks.

This development is part of a broader national strategy by China to promote BCI as a strategic industry. China aims to establish a leading position in neurotechnology by accelerating its development and commercialization, contrasting with the more cautious approach seen in the US and Europe.

The strategy encompasses both invasive and non-invasive technologies, with applications extending beyond medical use to industrial contexts. The move towards commercialization also raises new ethical debates concerning privacy, data control, security, and potential non-therapeutic uses.

👉 Read the full article on Gizmodo en Español