Our research and development in Artificial General Intelligence (AGI) focuses on understanding how intelligence functions within the human brain by developing software models inspired by neuroscience, particularly emphasizing the neocortex (the brain region responsible for higher-order thinking) its fundamental structures called cortical columns. Unlike other AI approaches, such as large language models (LLMs), our models maintain sufficient complexity to capture the key cognitive mechanisms, yet are abstract enough to remain computationally feasible.

Short-term goals:

• Develop a software framework for testing and exploring brain-inspired cognitive models.
• Gain deeper insights into human-like intelligence through simulation and experimentation.

Medium-term goals:

• Create a foundational model approaching human-level intelligence.
• Validate and refine our foundational model through practical applications in various cognitive tasks and real-world scenarios.
• Develop models to better understand neurological disorders, contributing to possible therapeutic advances.

Long-term goals:

• Build an AGI-level model based on our neuroscience-inspired approach.
• Develop personal, compassionate, understanding, and good-willed assistants based on this technology. Our first priority will be to support people with disabilities, including those affected by neurological disorders, because they may need such assistance most urgently to live independently and participate fully in modern society. These assistants are intended to help people navigate everyday challenges, improve communication, learn, work, and live with greater confidence, dignity, and autonomy. We also aim to make such assistants available more broadly, so they can support all people in everyday life. Beyond non-commercial scientific, educational, and medical uses that clearly benefit people, any other application of this emerging technology will be carefully restricted and considered only on a case-by-case basis, subject to individual agreements between the creators of the technology and interested parties.
• Explore advanced concepts such as consciousness, self-awareness, and “free will,” recognizing these topics as complex and not yet fully scientifically defined. While these properties may not be strictly necessary for basic human-level intelligence, they could be essential for fully understanding human behavior. We believe that our models will be designed to potentially exhibit these higher-level properties within simulations.

Recognizing that AGI represents a profoundly disruptive technological shift for humanity, we aspire to contribute to the creation of safe and beneficial AGI systems through informed and responsible research.