docs: document form autosave + FE↔BE request lifecycle in ARCHITECTURE

New §2g explains field persistence (keystroke → model → 600ms snapshot
debounce; blur only marks touched, never saves) and §6a refreshes the stale
backend section with the real request lifecycle (NSwag client, httpClientFetch
seam, read/write traces) — both with relative links to the source files.
Add discovery pointers from the learning path (lesson 2.2 and the capstone).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/docs/learning-path.mdx

@@ -153,6 +153,11 @@ the *outcome*. Reducer = "what the new state is"; command = "go do it, then say
 happened." And **derive, don't store** anything you can compute — e.g. a wizard's visible
 steps are `visibleSteps(answers)`, not a stored field.
 
+A field's value lands in the Model on **every keystroke** (not on blur — blur only marks
+the field "touched"); a separate 600 ms debounce off the model snapshot autosaves the
+draft to the backend, an effect that lives *outside* the reducer. See
+`docs/reference/architecture/ARCHITECTURE.md` §2g.
+
 **Do:** run `/form-machine` for a toy single field (say a "nickname" field with a max
 length). Read the generated Model / Msg / reduce and its spec.
 
@@ -359,7 +364,8 @@ one is the **authority**?
 
 **Go deeper:** `docs/reference/architecture/0001-bff-lite-decision-dtos.md`;
 `docs/reference/fp-tea-atomic-design.md` Part 7 (the copy-paste recipes);
-`docs/reference/architecture/ARCHITECTURE.md` §4. For how contexts scale to a second
+`docs/reference/architecture/ARCHITECTURE.md` §4 (the recipe) and §6a (the full FE⇄BE
+request lifecycle, read + write, with file links). For how contexts scale to a second
 app and actor-based authorization, ADR-0002 (the advanced read).
 
 ---