Start from the goal
Myelin asks what you are trying to learn or build, the depth you need, and the constraints around the work.
Method
The path starts as a proposal.
Myelin starts from what you are trying to do, drafts a bounded path, and uses your answers and review history as current evidence for what comes next.
The method is deliberately visible: Myelin shows the proposal, the important assumptions, and the review state instead of asking you to trust a hidden tutor.
Goal shaping draft Companion
Four steps
A simple loop, kept inspectable.
The useful part is the loop Myelin follows when a goal becomes learning work.
Work backward
The system drafts a bounded path from the goal instead of expanding every possible prerequisite.
Check assumptions
When a wrong guess would matter, the method is built to ask and keep important assumptions visible.
Return with evidence
Study and review history become current evidence for what needs attention the next time you return.
Goal-shaped depth
The same topic can need different depth.
Myelin treats a concept as part of a goal, not as a generic encyclopedia entry.
Goal-shaped depth
Authentication for a SaaS login flow
If the goal is to ship a first login flow, the path can focus on sessions, OAuth basics, PKCE, refresh behavior, and provider setup.
That is different from learning authentication to design a security system from scratch, where cryptography, protocol design, and threat modeling would need more weight.
The goal decides how deep the path needs to go and what can safely stay in the background for now.
Adjacent tools
Where Myelin sits next to the alternatives.
Spaced-repetition apps, AI study modes, and curricula each cover a piece of the work. Myelin is shaped for a different cut: a real goal, a path that ends, and the AI you already use.
Spaced-repetition apps
Generic decks that hold facts but do not know the goal you are pointed at.
Myelin shapes the path around what you are trying to do, then keeps review evidence tied to that path instead of to a flat deck.
AI study modes
Tied to one provider and ended with the chat. The goal, the path, and the review state do not survive the session.
Myelin connects to the AI you already pay for. The goal, the path, and the review state persist across sessions instead of staying inside one transcript.
Courses and curricula
Generic depth, decided in advance, with no signal back from where you actually struggled.
The path is bounded by your goal, and review evidence from your own work decides what comes back next.
Current evidence
One pass is not treated as enough.
A concept can be ready for this path without becoming permanent mastery everywhere. Myelin keeps confidence, timing, and signs of struggle close to the work.
Confidence is local
Confidence is current evidence for this concept in this path, not a permanent label.
Review has memory
Concepts return over time instead of disappearing after one introduction.
Struggle has signal
A stuck point can show what needs attention next without becoming a promise that every gap is automatically found.
Review state Dashboard review
Learning science
The method borrows from practical learning research.
Myelin is built around a small set of durable ideas: recall beats rereading, spacing matters, working memory is limited, and transfer depends on context.
These references shape the product method; they are not a claim that Myelin has already proven learning outcomes.
References checked April 25, 2026.
Recall over rereading
Review should make you retrieve and explain from memory, not just recognize a note you have already seen.
Spacing over one pass
Important concepts need to come back after time has passed, when the work is harder and more diagnostic.
Bounded working load
A path has to stay small enough to act on. Expanding every possible prerequisite makes the next move harder to see.
Context-shaped transfer
Knowing a concept in one situation does not mean it will automatically transfer to every other use of that concept.
Method
Start with a goal. Keep the path honest.
The method is meant to keep the AI useful without hiding the shape of the work. Start with what you want to learn, then let Myelin keep the path and review state visible as you move.