Humans& Wants to Own the Collaboration Layer in AI

[lingriiddd]

Humans&, though only three months old, has ambitious goals: it wants to redefine how AI interacts with humans in collaborative environments. The startup is training a foundation model for social intelligence, moving beyond single-user chatbots toward systems that can coordinate across teams and multiple agents.

The startup’s co-founders emphasize that the product and model are being developed together, ensuring the AI can evolve alongside the interface and user needs. Humans& faces competition from giants like OpenAI, Anthropic, and Google DeepMind, which are integrating collaboration features into existing platforms. However, Humans& is positioning itself as a generational company focused on rewriting how AI understands teamwork, rather than being acquired.

The $480 million seed round gives Humans& the resources to train and scale its model, but success will hinge on its ability to tackle real-world coordination challenges, a space many AI firms have yet to master.

[cryptohog]

multi-agent rl + long-horizon planning is tough. real-world coordination is way messier than lab environments. execution will make or break this.