Humans&, a new artificial intelligence startup founded by former researchers from Anthropic, Meta, OpenAI, xAI, and Google DeepMind, has raised a massive 480 million dollars in a seed funding round to build what it describes as a new generation of AI focused on human collaboration. The company says its goal is to move beyond single user chatbots and create a foundation model designed for social intelligence, capable of coordinating teams, managing long term decisions, and supporting collective problem solving.
The startup argues that while AI models have become highly capable at answering questions and generating content, they remain poorly suited for real world collaboration where people operate with competing priorities across multiple tools and timelines. According to co founder Andi Peng, the industry is entering a second phase of AI adoption, where the challenge is no longer intelligence alone but how humans actually work together with these systems in everyday settings.
Humans& says it is developing both a new model architecture and a product centered on communication and coordination, potentially replacing or redefining collaboration platforms such as Slack, Google Docs, or Notion. Chief executive Eric Zelikman said the model will be trained to understand group dynamics, ask meaningful questions, and support shared decision making over time, rather than optimizing only for short term user satisfaction or one off correct answers.
To achieve this, the company plans to train its model using long horizon and multi agent reinforcement learning, allowing the system to plan, adapt, and remember context across extended interactions involving multiple people and AIs. While the approach sets Humans& apart from major AI players, it also puts the startup in direct competition with established firms investing heavily in collaborative AI features. Despite the risks and high costs involved, the company says it is committed to remaining independent and positioning itself as a foundational player in the future of human AI collaboration.