Cover Photo Major News from Emergence, OpenAI, Alter3 and Google's Gemini

Emergence Unveils Ambitious Plans to Revolutionize AI Agent Technology

Emergence, led by former IBM executive Satya Nitta, emerged with $97.2 million in funding from Learn Capital and over $100 million in credit lines. The company aims to develop an innovative “agent-based” system utilizing generative AI models like OpenAI’s GPT-4o for various tasks. Emergence’s CEO, Nitta, highlighted their focus on advancing AI tasks such as planning and reasoning. The company’s roadmap includes an orchestrator agent and a project called Agent E for task automation. Emergence plans to monetize its technology through a premium version and strategic partnerships with Samsung and Newline Interactive. Despite competition in the AI agent space, Emergence stands out with its research-driven approach and experienced team. 

OpenAI Expands Enterprise Reach with Acquisition of Multi Collaboration Platform

OpenAI continues its acquisition streak by purchasing Multi, a video-first collaboration platform for enterprises, following the recent acquisition of Rockset. Multi, previously known as Remotion, will be integrated into OpenAI as an acqui-hire, with most of Multi’s team joining OpenAI. Multi’s CEO, Alexander Embiricos, announced the shutdown of Multi after July 24. The platform, founded by Embiricos and Charley Ho, aimed to enhance remote team collaboration through video chats. Multi had raised $13 million from investors like Greylock and First Round Capital. OpenAI’s focus on enterprise solutions is evident, with successful ventures like ChatGPT’s corporate tier and partnerships with firms like PwC. The company’s revenue is projected to exceed $3.4 billion this year. Speculation arises about potential integration of videoconferencing and collaboration features into ChatGPT in the future.

Alter3: GPT-4 Empowered Humanoid Robot Advances Language-Controlled Actions

Researchers at the University of Tokyo and Alternative Machine introduce Alter3, a humanoid robot system leveraging GPT-4 to interpret natural language commands for tasks like taking selfies and mimicking behaviors. Alter3’s design integrates large language models to enhance robotic actions, showcasing the synergy between AI models and robotics systems. By utilizing an “agentic framework,” GPT-4 plans and executes actions based on textual instructions, adapting its behavior to the robot’s API through in-context learning. Human feedback refines actions, enabling corrections and enhancing the robot’s performance. Alter3 demonstrates diverse capabilities, from everyday actions to emotional expressions like embarrassment and joy. This innovative approach highlights the potential of foundation models in robotics research, paving the way for advanced human-robot interactions and enhanced task execution. Despite the progress, challenges persist in developing robots for fundamental tasks due to data limitations and technical complexities. The integration of GPT-4 in Alter3 sets a precedent for future humanoid robots, emphasizing the transformative impact of AI models in enhancing robotic functionalities.

Google Introduces Gemini AI to Teen Students via School Accounts with Enhanced Educational Tools

Google expands access to its AI technology, Gemini, for teen students using their school accounts, aiming to equip them with skills for the future. Gemini offers real-time feedback to enhance learning confidence. Google assures responsible use of student data, with guardrails to prevent inappropriate responses and promote information literacy. The feature will be available in English across 100+ countries, with admins controlling its activation. Additionally, Google launches Read Along in Classroom globally, aiding reading skills development. Educators can assign tailored reading activities and track students’ progress. New features empower educators to create interactive lessons, manage assignments, and provide personalized feedback efficiently.