Imagine a world where AI doesn't replace us, but empowers us. That's the ambitious goal of Humans&, a new startup spearheaded by former xAI researcher Eric Zelikman, who is reportedly seeking to raise a staggering $1 billion in funding. This bold move signals a shift in the AI landscape, focusing on collaboration rather than competition between humans and machines.
Humans& aims to train AI models specifically designed to work seamlessly with humans. The company's vision involves creating AI that understands and responds to individual preferences, interests, and goals. The company is reportedly seeking a $5 billion valuation.
But here's where it gets interesting: the founding team is a veritable who's who of AI talent. Joining Zelikman are early Google employee Georges Harik, along with researchers from Meta, Anthropic, OpenAI, and Deepmind. This dream team brings a wealth of experience and expertise to the table.
While the specific product remains under wraps, Humans& has hinted to investors that its new training methods will require significantly more computational power than current AI training strategies. This is a crucial point, as the enormous resources needed for advanced AI development often deter startups, pushing them to build applications on top of existing models like GPT-5 and Claude Opus 4.1. This makes Humans&'s ambitious undertaking all the more noteworthy.
And this is the part most people miss: Zelikman's approach contrasts sharply with other AI labs focused on automating human tasks. He believes in building models that excel at collaborating with large groups of people, understanding their diverse goals and values, and coordinating efforts to make everyone more effective. He stated, "If we invest everything in autonomy and nothing in collaboration, all IQ and no EQ, the future will be a colder place."
Zelikman, who wrote a paper on training language models to reason in natural language, previously worked on data collection at xAI. Georges Harik, the 7th employee at Google and co-creator of Adwords and Adsense, is also involved. Furthermore, Noah Goodman, Zelikman's PhD advisor, and Andi Peng, who worked on post-training at Anthropic, are also part of the founding team. Ray Ramadorai, who worked on system design for Microsoft's large data centers, rounds out the core team.
Controversially, one investor reportedly passed on the funding round, citing the size of the ask as too large for an early-stage company.
What do you think? Is Humans&'s focus on human-AI collaboration the future of AI, or is the investment too risky? Share your thoughts in the comments below!
