AI chatbots have made significant strides in answering queries, summarizing data, and solving mathematical problems. However, their current capabilities are limited to assisting individuals on a one-on-one basis. They lack the complexity required for managing collaborative tasks, such as coordinating teams with diverse priorities, tracking prolonged decisions, and maintaining alignment among team members over time.

Enter Humans&, a pioneering startup founded by industry veterans from Anthropic, Meta, OpenAI, xAI, and Google DeepMind. The company recently secured a substantial $480 million seed funding to develop a “central nervous system” for the evolving human-plus-AI ecosystem. While the company’s initial focus on “AI for empowering humans” has garnered attention, its true innovation lies in creating a new foundation model architecture geared towards social intelligence rather than just information retrieval or code generation.

According to Andi Peng, a co-founder of Humans& and former Anthropic employee, the current phase marks a shift from specialized vertical-focused question-answering models to a broader adoption where everyday users seek to harness the full potential of AI technologies.

The core proposition of Humans& revolves around facilitating the transition into a new AI era, dispelling the notion that AI will replace human jobs. The company’s timing is crucial as organizations move from chat-based interactions to more sophisticated AI agents. While AI models exhibit competence, operational workflows remain challenging, particularly in terms of coordination. Amidst these developments, individuals often feel overwhelmed and threatened by AI’s growing presence.

Despite being a nascent entity, Humans& has managed to secure substantial funding based on its visionary approach and the distinguished background of its founders. The company is yet to unveil a concrete product roadmap, hinting at potential applications in both enterprise and consumer settings.

Eric Zelikman, co-founder and CEO of Humans&, envisions a product that emphasizes communication and collaboration, aiming to enhance teamwork and interaction among individuals and AI tools. The team’s goal is to create a model that mimics human-like questioning patterns, fostering a more natural and engaging interaction process.

Zelikman emphasizes that the forthcoming model will prioritize user engagement and understanding, departing from the conventional approach of optimizing responses solely based on user satisfaction and accuracy.

Humans& adopts a co-evolutionary strategy, developing the product in tandem with the model’s advancements. The team aims to create an interface that aligns seamlessly with the evolving capabilities of the model.

Central to Humans&’s vision is the ambition to redefine the collaboration landscape by owning the collaboration layer rather than merely integrating with existing tools. The convergence of AI and team collaboration tools represents a burgeoning field, with startups like Granola pioneering collaborative features in note-taking applications.

Industry luminaries, including LinkedIn founder Reid Hoffman, advocate for a paradigm shift towards coordination and collaboration in AI development, recognizing the pivotal role of teamwork and knowledge-sharing in driving innovation.

Humans& aspires to serve as the linchpin across diverse organizational structures, leveraging its model-product hybrid to understand individual preferences, motivations, and skills, thereby optimizing collective outcomes.

The company’s training approach emphasizes human-AI collaboration, incorporating long-horizon and multi-agent reinforcement learning to enhance planning, decision-making, and user interactions over time.

Yuchen He, a Humans& co-founder and former OpenAI researcher, underscores the significance of memory retention in enhancing user understanding and engagement. The model’s training regimen encompasses advanced reinforcement learning techniques to foster multi-agent interactions and long-term planning.

Despite its promising trajectory, Humans& faces formidable challenges, including the substantial financial investment required for model training and scalability. Moreover, the company competes with established industry giants actively exploring AI-driven collaboration solutions.

The strategic positioning of Humans& as a pioneer in social intelligence models could potentially position it as an acquisition target for industry leaders like Meta, OpenAI, and DeepMind. However, the company remains steadfast in its commitment to charting its independent course and reshaping the future of human-AI interactions.

According to Zelikman, Humans& envisions itself as a transformative force in the AI landscape, poised to revolutionize user interactions with AI models and drive meaningful change in collaborative workflows.