Humane Launches $699 Ai Pin, Harnessing Generative AI for Hands-Free Experience
Humane, a hardware startup founded by former Apple employees Bethany Bongiorno and Imran Chaudhri, has officially launched its first product, the Ai Pin. The company, which has raised $230 million in funding since its inception in 2017, aims to create stand-alone AI devices that leverage the power and popularity of generative AI platforms like OpenAI’s ChatGPT and Google’s Gemini. The Ai Pin, priced at $699, is the first in what Humane hopes will be a long line of devices designed to free users from the limitations of traditional technology. The voice-based, always-connected device utilizes multiple LLM-based systems, selecting the best platform for each individual task. The $699 entry point for the Ai Pin includes the device, an extra battery, and an AI charging case. Different color options start at $799, and various accessories are available, ranging from $29 to $49. Humane also offers a subscription service at $24/month, with three free months included with the purchase of a pin.
OpenAI Enhances ChatGPT with Upgraded GPT-4 Turbo Model for Premium Users
OpenAI has announced a significant upgrade to its viral AI-powered chatbot, ChatGPT, for premium users subscribed to ChatGPT Plus, Team, or Enterprise plans. The updated version of GPT-4 Turbo, one of the models powering the conversational ChatGPT experience, brings improvements in writing, math, logical reasoning, and coding, as well as a more up-to-date knowledge base. The new model was trained on publicly available data up to December 2023, providing it with more current information compared to the previous edition of GPT-4 Turbo available in ChatGPT, which had an April 2023 cut-off date. This ChatGPT update follows the general availability launch on Tuesday of new models in OpenAI’s API, most notably GPT-4 Turbo with Vision, which adds image understanding capabilities to the typically text-only GPT-4 Turbo model.
Google Unveils RecurrentGemma: Bringing Advanced Language AI to Edge Devices
Google has introduced RecurrentGemma, a groundbreaking open language model that brings advanced AI text processing and generation to resource-constrained devices like smartphones, IoT systems, and personal computers. This innovative architecture significantly reduces memory and processing requirements while maintaining excellent performance, making it ideal for applications that demand real-time responses, such as interactive AI systems and real-time translation services. Unlike state-of-the-art language models like OpenAI’s GPT-4, Anthropic’s Claude, and Google’s Gemini, which rely on the resource-intensive Transformer architecture, RecurrentGemma achieves its efficiency by focusing on smaller segments of input data at a time. It combines techniques from traditional recurrent neural networks and attention mechanisms to address the shortcomings of Transformers in situations where efficiency is critical. The introduction of RecurrentGemma marks an exciting new development in the evolution of language AI, bringing the power of advanced text processing and generation to the edge.
Meta AI Introduces OpenEQA: Advancing Embodied Intelligence in Artificial Agents
Meta AI researchers have released OpenEQA, a groundbreaking open-source benchmark dataset designed to measure an artificial intelligence system’s capacity for “embodied question answering.” This new benchmark evaluates an AI’s ability to understand the real world and answer natural language questions about an environment. The OpenEQA dataset contains over 1,600 questions about more than 180 real-world environments, such as homes and offices. These questions test an AI’s skills in object and attribute recognition, spatial and functional reasoning, and commonsense knowledge. The researchers believe that Embodied Question Answering or EQA is both a useful end-application and a means to evaluate an agent’s understanding of the world. The release of OpenEQA marks a significant step forward in the development of embodied AI, providing a comprehensive benchmark for evaluating the progress of artificial agents in understanding and interacting with the real world.
