AI chatbots are getting better at answering questions, summarizing documents, and solving math equations, but they still act a lot like helpful assistants for single users. They’re not designed to manage real collaborative work: coordinating people with competing priorities, tracking long-delayed decisions, and keeping teams aligned over time.
Humans&, a new startup founded by alumni of Anthropic, Meta, OpenAI, xAI, and Google DeepMind, thinks that gap is the next major frontier for the foundation model. This week’s company is on the rise seed round of $480 million to build a “central nervous system” for the human-plus-AI economy. the beginning”AI to empower humans” framing has dominated the initial coverage, but the company’s real ambition is more novel: building a new basic model architecture designed for social intelligence, not just information retrieval or code generation.
“It feels like we ended the first scale paradigm, where question answer models were trained to be very smart in certain verticals, and now we are entering what we believe is a second wave of adoption where consumers or average users are trying to figure out what to do with all these things,” Andi Peng, one of the human founders and former founder of TechCrupic.
Human focus & focus on helping people enter the new AI era, beyond the narrative that AI will take over jobs. Whether or not it’s just marketing talking, time is critical: The company is transitioning from chat to agent. The model is competent, but the workflow is not, and coordination challenges remain unaddressed. And through it all, people feel threatened and overwhelmed by AI.
The three-month-old company, like some of its peers, has been able to raise the amazing seeds behind this philosophy and the pedigree of the founding team. Human& still does not have a product, nor is it clear what exactly it is, although the team says that it can be a replacement for multiplayer or multi-user contexts like a communication platform (think). Slack) or collaboration platforms (think Google Docs and Notion). For use cases and target audiences, the team advises on enterprise and consumer applications.
“We build products and models that focus on communication and collaboration,” said Eric Zelikman, co-founder and CEO of human& and former xAI researcher, to TechCrunch, adding that the focus is on getting products to help people work together and communicate more effectively – along with AI tools.
“It’s like when you have to make a big group decision, often someone takes everyone to one room, so that everyone expresses a different camp, for example, what kind of logo they want,” said Zelikman, chortling with the team while remembering the time it took to get everyone to agree on the logo for the start.
Techcrunch event
San Francisco
|
13-15 October 2026
Zelikman added that the new model will be trained to ask questions in a way that feels like interacting with a friend or colleague, someone who is trying to get to know you. Chatbots are now programmed to ask questions constantly, but they do so without understanding the value of the question. They say that this is because they are optimized for two things: How much the user immediately likes the responses they give, and how well the model can answer the questions they receive correctly.
Part of the lack of clarity about what the product is may be due to humans and we don’t have an answer yet. Peng said that human& designed the product together with the model.
“Part of what we’re doing here is making sure that the model gets better, we can develop interfaces and behaviors that the model can turn into a product,” he said.
But clearly, human& is not trying to create a new model that can connect to existing applications and collaboration tools. Startups want to have a layer of collaboration.
AI-plus team collaboration and productivity tools are a hot field – for example, startup AI note-taking app Granola is raising the bar $43 million round at a cost of $250 million due to the launch of other collaboration features. Some high-profile voices have also clearly made the next phase of AI one of coordination and collaboration, not just automation. LinkedIn founder Reid Hoffman said today that companies are running AI wrong by treating it like an isolated pilot and that the real impact is at the work coordination layer – that is, how teams share knowledge and hold meetings.
“AI lives at the workflow level, and the people closest to the work know where the friction is,” Hoffman said write on social media. “They will discover what needs to be automated, compressed, or redesigned.”
That is the place where people want to live. The idea is that the model-slash-product will be the “connective tissue” in any organization – be it a business of 10,000 people or a family – that understands the skills, motivations and needs of each person, as well as how everything can be balanced for the good of all.
Getting there requires rethinking how AI models are trained.
“We’re trying to train the model in a different way that will involve more humans and AI interacting and collaborating together,” Yuchen He, human founder and former OpenAI researcher, told TechCrunch, adding that the initial model will also be trained using long-horizon and multi-agent reinforcement learning (RL).
Long-horizon RL is meant to train the model to plan, act, improve, and follow through over time, rather than just generating good one-off answers. Multi-agent RL trains for environments where multiple AIs and/or humans are in the loop. Both concepts are gaining momentum new academic work as researchers push LLM beyond chatbot responses to systems that can coordinate actions and optimize results through many steps.
“The model has to remember about itself, about you, and the better the memory, the better the understanding of the user,” he said.
Despite the stellar crew running the show, there are many risks ahead. Humans& need an infinite amount of money to finance the expensive endeavors of training and creating new models. That means it will compete with major established players for resources, including access to computing.
However, the highest risk is human& not just competing with the Understandings and Slacks of the world. It comes to Top Dogs from AI. And the company is actively working on better ways to enable human collaboration on its platform, even as it vows that AGI will replace economically viable work. Through Claude Cowork, Anthropic aims to optimize work style collaboration; Gemini is embedded in Workspace so AI-enabled collaboration is already happening on devices people already use; and OpenAI has lately been building developers around multi-agent orchestration and workflows.
Importantly, none of the major players is ready to rewrite the model based on social intelligence, which can provide humans and can become acquisition targets. And with companies like Meta, OpenAI, and DeepMind on the prowl for top AI talent, M&A is always a risk.
Human& told TechCrunch that it has turned away interested parties and is not interested in being acquired.
“We believe this is going to be a generational company, and we think it has the potential to change the future of how we interact with these models,” Zelikman said. “We believe in ourselves, and we believe in the team we have assembled here.”
This post was originally published on January 22, 2026.