A new artificial intelligence startup, Humans&, is betting that coordination—not computation—is the next major leap for AI systems.

Founded by former researchers from Anthropic, OpenAI, Meta, xAI, and Google DeepMind, the company has raised an extraordinary four hundred and eighty million dollars in seed funding to develop a new class of socially intelligent AI models.

While today’s AI tools excel at answering questions and generating content, Humans& argues that they fall short in managing real-world collaboration.

The startup aims to build a foundational model capable of helping groups align priorities, track decisions over time, and navigate complex human dynamics across teams and organizations.

The company’s leadership describes its vision as creating a “central nervous system” for human and AI collaboration.

Rather than integrating into existing productivity tools, Humans& plans to own the collaboration layer entirely, potentially reshaping how people work together digitally. Although the company has not yet released a product, it has hinted at applications spanning enterprise and consumer use cases.

To achieve its goals, Humans& is developing a new training approach that emphasizes long-term planning and multi-agent interaction.

By combining long-horizon reinforcement learning with environments involving multiple humans and AI systems, the company hopes to teach its models how to remember context, understand individuals, and support collective decision-making.

Despite strong investor confidence, Humans& faces steep challenges, including the high cost of training large models and competition from established AI leaders already embedding collaboration tools into their platforms.

Still, the startup insists it is building a standalone, generational company and has ruled out acquisition talks as it pursues its ambitious roadmap.

Source: Techcrunch