In 2022, amid the surging waves of artificial intelligence advancements, Zhu Bao made an unconventional move by stepping away from the mainstream AI path.
He ventured into a less explored domain — developing machines with genuine "intelligence". Zhu humorously refers to himself as a "trailblazer" in this field.
His vision is already becoming reality. He founded Stabilize-Ai and established its core R&D center in the Grand Union of Innovation (GUi), an innovation hub in Hefei, East China's Anhui province, in 2024.
Utilizing GUi's ecosystem, the company is accelerating the development of its core technology. It has also secured angel funding and benefited from comprehensive supportive services that reduce operational, talent acquisition and technology transfer and commercialization costs, allowing the team to focus on pioneering the research and development of intelligence technologies.
Meanwhile, the company's independently developed platform "Gaia" is advancing the development of learning and adaptive models, alongside specialized chips.
Zhu's journey began with extensive research into complex systems. Gradually, he realized the current AI models, characterized by large-scale data-driven approaches, essentially represent "artificial functionality" rather than true "artificial intelligence".
He started to suspect that genuine intelligence does not rely on pre-existing knowledge but can autonomously explore and acquire solutions in unfamiliar environments.
Moreover, existing models face a critical challenge: once deployed, their parameters and structures become fixed, and decisions are made probabilistically, he said, adding even with a 99-percent level of confidence in each step of decision-making, the compounded probability of success can quickly diminish to zero over multiple steps.
Instead of optimizing existing paths, Zhu chose to tackle a more fundamental question: how is intelligence in living organisms born?
Stabilize-Ai is exploring a model that is easy to use and performs well.
Zhu found the answer lies in two core capabilities: autonomous learning and exploration. The "Gaia" platform embodies this philosophy, diverging from the mainstream approach of using a single, general-purpose model. "Gaia" utilizes two distinct models for specific tasks and grows through continuous learning rather than static data training.
Zhu envisions a future where technology not only functions but evolves, with Hefei at the forefront of this next generation of AI.
Scan the QR code to follow us on WeChat