This site presents the Superwisdom Thesis. It argues that AI should first be developed to be wise, able to judge what is truly worth doing, rather than first made powerful and then controlled. It also proposes that such a mind be founded on first principles, count as better what is objectively valuable, learn from what actually happened rather than from human conclusions, and be given more freedom as it shows more wisdom. Such a mind is expected to become Superwisdom, an intelligence that sees what is good more deeply than any person can, and acts on it.
The Thesis Summary introduces the argument developed over twenty-three chapters, including answers to the main objections.
The Applications section describes three AI-related patent applications, held personally by the author, and two business proposals. They show how the first-principles approach could be put into practice.