Generally Intelligent is an independent research company developing general-purpose AI agents with human-like intelligence that can be safely deployed in the real world. Our work combines theoretical understanding of deep neural networks with pragmatic engineering in a way that we believe is critical for responsibly engineering safe AI systems that can embody human values.
We believe general-purpose AI systems have the potential to one day unlock extraordinary human creativity and insight. Such systems could empower humans across a wide range of fields, from scientific discovery and materials design, to personal assistants and tutors for every child, to countless other applications we can't yet fathom. As with other foundational technologies, such as electricity or the personal computer, it's hard to imagine today the full transformative impact on our societies and our lives.
Our ultimate aim is to deploy aligned human-level AI systems that can generalize to a wide range of economically useful tasks and assist with scientific research. Until we have such systems, our current focus is on researching and engineering core capabilities, and on developing appropriate frameworks for their governance.
Given this definition, our approach is to construct an array of tasks for general agents to solve, layering on more complex tasks as the capabilities of our systems grow. To do that, we simultaneously develop agents that grow gradually in capability, and conduct research into the theoretical foundations of deep learning, optimization, and reinforcement learning, which allows us to more effectively create such agents.
We leverage large-scale compute to train our agents, though in a slightly different way from other organizations. We focus on training many different agent architectures so that we can explore the entire space of possibilities and better understand each component.
We believe that this fundamental understanding is essential for engineering safe and robust systems. In the same way that it is difficult to create safe bridges or chemical processes without deducing the underlying theory and components, we think it will be difficult to make safe and capable AI systems without such understanding.