Eric Zelikman, a former researcher at xAI, is set to raise $1 billion for his new venture, Human&, at a $5 billion valuation. The company aims to pioneer “human-centered AI” — technology designed to complement and strengthen human abilities rather than substitute them. (Source: Forbes)

• The founding team includes Georges Harik (Google’s 7th employee) and senior engineers from OpenAI, Anthropic, Meta, and DeepMind.
• Human& is developing a new AI training method that emphasizes reasoning and empathy, helping systems better understand human intent and collaboration patterns.
• Zelikman’s earlier research in “chain-of-thought” reasoning — training models to break problems into logical steps — directly influenced OpenAI’s o1 reasoning series.

Why it matters: In an era where most AI labs are chasing superintelligence, Zelikman’s Human& represents a philosophical shift — prioritizing AI that works with humans rather than against them. The funding discussions underscore investor confidence in post-foundation model innovation, where human-aligned AI could shape the next decade. The potential $5B valuation also highlights the soaring appetite for pre-revenue, high-concept AI ventures leading the next wave of frontier research.