Planned learning path
A spiral path revisits core concepts in new worlds, with more independence and complexity each time.
Collect examples, choose labels and test simple classifiers. Discover coverage, noise and the difference between a pattern and a reason.
Combine objectives, observations, memory and tools. See how a poorly specified goal can produce a technically successful but unhelpful result.
Create edge cases, compare outcomes and investigate bias. Learn why average performance can hide an important failure.
Coordinate multiple agents inside a persistent simulation. Balance short-term performance with long-term effects and explain a design decision.
This is a product-development roadmap, not a finalized accredited curriculum. Sequence and terminology will evolve through expert review and family testing.
Evidence over activity
Planned assessment is embedded in creation: a learner's model card, test cases, revisions and reflection show more than a multiple-choice score alone.
Added night-water examples after the first test failed.
False alarms fell from 6 to 2 in the challenge set.
Will the agent work in cloudy water?
Founding families
Join the early-access list for product previews, parent research invitations and launch updates.
