Pre-Assessment
Before generating any content, Lugh runs a diagnostic Feynman session. This isn’t a quiz — it’s a conversation. The system asks you to explain what you already know about the topic, then maps your response against the curriculum’s learning objectives.
Why this matters
Two failure modes of traditional courses:
- Covering things you already know (wastes your time, kills momentum)
- Skipping things you think you know but have wrong (leaves misconceptions intact)
The pre-assessment catches both.
How it works
The prompt is open-ended: “Tell me what you already know about [topic].” No multiple choice, no leading questions. The learner talks (or types), and the system listens.
The response gets mapped against the learning objectives for each episode in the curriculum. Each objective falls into one of three categories:
Solid
You explained it accurately. The episode covering this concept gets condensed or skipped entirely. Time saved, no redundancy.
Shaky
You know the term or concept, but your explanation has gaps, errors, or common misconceptions. This is the most valuable category.
The episode covering this concept gets a corrective emphasis: “You might think X — and that’s close — but actually Y, and here’s why the distinction matters.” This is more effective than teaching from scratch because it has something to push against.
Example: A learner says “histamine is the allergy chemical.” That’s shaky — it’s not wrong, but it’s incomplete in a way that will cause confusion later. The histamine episode would open by acknowledging that framing and then expanding it: histamine is one of many mediators, and it does far more than drive allergic responses.
Blank
No prior knowledge detected. Full episode generated with no assumptions about background.
What the pre-assessment is NOT
It’s not a gate. You don’t “fail” a pre-assessment. It’s purely diagnostic — it shapes what gets generated, it doesn’t block anything.
It’s also not exhaustive. The system doesn’t need you to demonstrate knowledge of every concept. A few minutes of explanation gives enough signal to meaningfully customize the curriculum. Silence on a topic is treated as “blank,” not as failure.
Pre-assessment at different depths
The same pre-assessment conversation produces different mappings depending on the selected Learning Depths:
- At Depth 1, the system is checking for gross misconceptions, not nuanced understanding
- At Depth 2, it’s mapping which core concepts need full coverage vs. which can be compressed
- At Depth 3, it’s looking for subtle errors in mental models that would compound across the series
Re-assessment on depth change
If a learner completes Depth 1 and later returns for Depth 2, the pre-assessment incorporates what they learned in the first pass. The system already has their tutor session data — no need to re-explain from scratch.