Humans& Pioneers AI Coordination to Revolutionize Team Collaboration
In the rapidly evolving landscape of artificial intelligence, a new startup named Humans& is making significant strides by focusing on enhancing coordination and collaboration within teams. Founded by veterans from leading AI organizations such as Anthropic, Meta, OpenAI, xAI, and Google DeepMind, Humans& has recently secured a substantial $480 million seed funding to develop a novel AI model centered on social intelligence.
Bridging the Collaboration Gap in AI
Traditional AI chatbots have excelled in tasks like answering queries, summarizing documents, and solving mathematical problems. However, they often fall short in managing the complexities of real-world collaboration, such as aligning team members with diverse priorities, tracking long-term decisions, and maintaining team cohesion over extended periods. Humans& aims to address these challenges by creating an AI system that acts as a central nervous system for the human-AI collaborative economy.
Andi Peng, a co-founder of Humans& and former employee at Anthropic, highlighted the transition in AI development:
It feels like we’re ending the first paradigm of scaling, where question-answering models were trained to be very smart at particular verticals, and now we’re entering what we believe to be the second wave of adoption where the average consumer or user is trying to figure out what to do with all these things.
Innovative Training Approaches
To achieve its ambitious goals, Humans& is rethinking AI training methodologies. The company plans to employ long-horizon reinforcement learning (RL) and multi-agent RL. Long-horizon RL focuses on training models to plan, act, revise, and follow through over extended periods, moving beyond generating isolated responses. Multi-agent RL prepares AI systems to operate in environments where multiple AIs and humans interact, fostering collaborative problem-solving.
Yuchen He, a co-founder and former OpenAI researcher, emphasized the importance of memory and user understanding in their model:
The model needs to remember things about itself, about you, and the better its memory, the better its user understanding.
A New Paradigm in AI Interaction
Humans& envisions an AI that interacts with users in a manner akin to a colleague or friend, asking questions and engaging in dialogues that facilitate deeper understanding and collaboration. This approach contrasts with current chatbots that often lack contextual awareness and the ability to engage in meaningful, ongoing interactions.
Eric Zelikman, co-founder and CEO of Humans& and former xAI researcher, elaborated on the company’s focus:
We are building a product and a model that is centered on communication and collaboration… the focus is on getting the product to help people work together and communicate more effectively — both with each other and with AI tools.
Challenges and Industry Context
Despite the promising vision, Humans& faces significant challenges, including the need for substantial funding to train and scale their new model. The startup is entering a competitive landscape where established AI companies are also exploring enhanced collaboration tools. For instance, OpenAI has been developing specialized AI agents tailored for various applications, indicating a broader industry trend towards more integrated and collaborative AI solutions.
However, Humans& differentiates itself by focusing on social intelligence and the coordination layer of work, aiming to create an AI system that understands and balances the skills, motivations, and needs of each team member for the collective benefit.
Future Prospects
As Humans& continues to develop its model and product, the startup remains committed to its vision of fundamentally changing how humans interact with AI. The company’s approach reflects a growing recognition in the AI industry of the importance of coordination and collaboration, moving beyond automation to create systems that enhance human teamwork and productivity.
