Knowledge InSight

Knowledge InSight makes self-paced learning experiences for curious people who want to go deeper into a subject for their own sake. There is no enrolment, no cohort, and no deadline. You bring the curiosity; we provide the structure, sequence, and guiding questions.

Our programs are open access and self-paced. Nothing is graded. Quizzes score themselves and can be retaken without limit. Written work stays with you, in your own notebook or document; we never collect it.

Each program is organized as a small number of courses, each course as modules, and each module as lessons. A lesson is one sitting. The structure repeats so you can settle into it and spend your attention on the material instead of the interface.

Our programs are developed with the help of generative AI, and we would rather tell you exactly how than leave you guessing.

A learning designer works iteratively with an AI assistant to design the program structure, the course outlines, and the learning objectives. Once that scaffolding is settled, the designer uses AI to draft the learning content against those outlines. Every reading names its sources and links to them wherever they are freely available, so you can go to the original instead of taking a summary on trust. That is the point of the links: not decoration, but a standing invitation to check the work.

A fully vetted program has completed human-in-the-loop review: a subject-matter expert has validated every learning item and every source in it.

An in-progress program is published while that review is still underway. We release early because the material is useful early, and because readers catch things that reviewers miss. Every page of an in-progress program says so, and tells you how to report what you find.

We also leverage AI to help facilitate your learning journey. Each Guided Conversation is a prompt written and reviewed by the learning designer. It tells the assistant to ask rather than lecture, to keep its turns short, to treat your observations as evidence, and to correct a specific list of misreadings that the material tends to invite.

You run it in your own account, so the conversation belongs to you. We never see it. You can open the same prompt in Claude, ChatGPT, or another LLM of your choice.

Nothing in a Guided Conversation is graded or recorded. If the assistant says something that contradicts the readings, trust the readings and tell us. Remember that the LLM model you use will impact the quality of the conversation.

Nothing leaves your browser. Marking an item done, and any quiz score you record, is saved in this browser's local storage so the site can show you where you left off. It is not synced, not shared, and not visible to us. Clearing your browser data clears it.

The Guided Conversations open in your own AI account. We never see those conversations.