Humans& thinks coordination is the next step for AI, and they’re building a model to prove it | TechCrunch

AI chatbots are getting better at answering questions, summarizing documents, and solving math equations, but they still largely act like helpful assistants for one user at a time. They’re not designed to handle the messier work of true collaboration: coordinating people with competing priorities, tracking long-term decisions, and keeping teams aligned over time.

Humans&, a new startup founded by alumni of Anthropic, Meta, OpenAI, xAI, and Google DeepMind, thinks closing that gap is the next big frontier for basic modeling. The company this week raised a $480 million seed round to build a “central nervous system” for the human-plus-AI economy. The start-up “AI to empower people’Framing has dominated early coverage, but the company’s real ambition is newer: to build a new basic model architecture designed for social intelligence, not just information retrieval or code generation.

“It feels like we’re ending the first paradigm of scaling, where question-answer models were trained to be really smart in certain industries, and now we’re entering what we think is the second wave of adoption where the average consumer or user is trying to figure out what to do with all this stuff,” Andi Peng, one of the co-founders of Human& and a former Anthropic employee, told TechCrunch.

Humans&’s pitch is about helping people enter the new era of AI, and goes beyond the narrative that AI will take over their jobs. Whether this is just marketing speak or not, the timing is critical: companies are moving from chat to agents. Models are competent, but workflows are not, and the coordination challenge remains largely unsolved. And despite all this, people feel threatened and overwhelmed by AI.

The three-month-old company, like some of its peers, has managed to produce its surprising seed, complementing this philosophy and the pedigree of its founding team. Humans& still doesn’t have a product, and it’s not clear what exactly it might be, though the team said it could be a replacement for multiplayer or multi-user contexts like communication platforms (think Slack) or collaboration platforms (think Google Docs and Notion). In terms of use cases and target audiences, the team hinted at both enterprise and consumer applications.

“We are building a product and a model focused on communication and collaboration,” Eric Zelikman, co-founder and CEO of Humans and former xAI researcher, told TechCrunch. He added that the focus is on helping the product help people collaborate and communicate more effectively – both with each other and with AI tools.

“Just like when you have to make a big group decision, it often comes down to someone getting everyone into one room and letting everyone express their different camps on, for example, what kind of logo they want,” Zelikman continued, chuckling with his team as they recalled the time-consuming tedium of getting everyone to agree on a logo for the startup.

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 you’re trying to get to know. Chatbots today are programmed to constantly ask questions, but they do this without understanding the value of the question. He says this is because they are optimized for two things: how much a user immediately likes when an answer is given, and how likely the model is to answer the received question correctly.

Part of the lack of clarity about what the product is may be that people don’t exactly have an answer for that yet. Peng said Humans designs the product together with the model.

“Part of what we’re doing here is also ensuring that as the model improves, we can evolve the interface and the behavior that the model is capable of into a meaningful product,” she said.

What is clear, however, is that Human& is not trying to create a new model that plugs into existing applications and collaboration tools. The startup wants to own the collaboration layer.

AI plus team collaboration and productivity tools are becoming increasingly popular. For example, startup AI note-taking app Granola raised a $43 million round at a $250 million valuation as it launched more collaboration features. Several high-profile voices are also explicitly framing the next phase of AI as one of coordination and collaboration, and not just automation. LinkedIn founder Reid Hoffman argued today that companies are implementing AI incorrectly by treating it as isolated pilots and that the real impact is in the coordination layer of work – that is, the way teams share knowledge and run meetings.

“AI lives at the workflow level and the people closest to the work know where the friction actually is,” says Hoffman wrote on social media. “They are the ones who will discover what needs to be automated, compressed or completely redesigned.”

That is the space where people want to live. The idea is that the model-slash-product would act as the ‘connective tissue’ in any organization – be it a 10,000-employee company or a family – that understands the skills, motivations and needs of each person, as well as how all of these can be balanced for the good of the whole.

To achieve this, the way AI models are trained requires rethinking.

“We are trying to train the model in a different way, where more humans and AIs interact and collaborate with each other,” Yuchen He, co-founder of humans and former OpenAI researcher, told TechCrunch, adding that the startup’s model will also be trained using long-horizon and multi-agent reinforcement learning (RL).

Long-horizon RL is intended to train the model to plan, execute, revise, and monitor over time, rather than just generating a good one-time answer. Multi-agent RL trains for environments where multiple AIs and/or humans are involved. Both concepts are gaining momentum recent academic work as researchers push LLMs beyond chatbot responses to systems that can coordinate actions and optimize results over many steps.

“The model has to remember things about itself and about you, and the better the memory, the better the user understanding,” he said.

Despite the great team running the show, there are plenty of risks ahead. Humans will need endless large sums of money to finance the expensive undertaking of training and scaling a new model. That means it will compete with the big established players for resources, including access to computers.

The greatest risk, however, is that humans do not simply compete with the world’s views and rules. It comes for the Top Dogs of AI. And those companies are actively working on better ways to enable human collaboration on their platforms, even as they vow that AGI will soon replace economically viable work. Through Claude Cowork, Anthropic wants to optimize collaboration at work; Gemini is embedded in Workspace, so AI-powered collaboration already happens within the tools people already use; and OpenAI has been pitching developers on its multi-agent orchestration and workflows lately.

Crucially, none of the major players are about to rewrite a model based on social intelligence that gives humans an edge or makes them a takeover target. And with companies like Meta, OpenAI, and DeepMind looking for top AI talent, mergers and acquisitions certainly pose a risk.

Humans& told TechCrunch that it has already rejected interested parties and is not interested in acquiring it.

“We believe this will be a generational business, and we think this has the potential to fundamentally change the future of how we interact with these models,” Zelikman said. “We are confident that we will do that, and we have a lot of confidence in the team we have put together here.”

This post was originally published on January 22, 2026.