速递｜五大厂前员工联手创业“AI微信”，4800万美元种子轮押注“社交智能”新基础模型架构

Core Insights - The article discusses the emergence of Humans&, a startup founded by former employees of major AI companies, aiming to bridge the gap in AI's ability to facilitate complex human collaboration and decision-making processes [2][3][4] - Humans& has raised $48 million in seed funding to develop a new model architecture designed specifically for social intelligence, moving beyond traditional AI applications like information retrieval and code generation [3][4] Group 1: Company Vision and Goals - Humans& aims to empower humans in the AI era, shifting the narrative from "AI will replace human jobs" to one that emphasizes collaboration between humans and AI [3][4] - The startup is focused on creating a product that enhances communication and collaboration, potentially replacing existing platforms like Slack and Google Docs [4][5] - The company intends to develop a model that interacts with users in a more human-like manner, understanding the value of questions rather than just optimizing for immediate responses [5][6] Group 2: Product Development and Market Position - Currently, Humans& has not launched a specific product but is working on a collaborative ecosystem that integrates AI with productivity tools [7][8] - The startup's approach involves training models through long-term planning and multi-agent reinforcement learning, aiming to optimize outcomes over time rather than providing one-off answers [8][9] - Humans& is positioned in a competitive landscape, not only against collaboration tools but also against major AI players like Anthropic and OpenAI, which are also focusing on enhancing human collaboration through AI [9][10] Group 3: Future Prospects and Challenges - The company believes it can fundamentally change how humans interact with AI models, aiming for a long-term vision rather than seeking immediate acquisition [10][11] - Despite the promising outlook, Humans& faces significant challenges, including the need for continuous funding to support the development of its new models and the competition for resources with established AI companies [9][10]