Two Types of Gaps: How to Spot Them and What to Do Next
Beyond the Linear Path to Mastery
Some people visualize learning as a single spectrum toward mastery:
Unknown unknown → Known unknown → Recognition → Reconstruction → Mastery
In that framing, the unknown unknown starts at zero, and every step is just “further along the line.”
But I’ve found that even when a topic feels comfortably understood, small, targeted investigations — the kind I describe later in Level 2 gaps — can be massive unlocks.
They don’t just inch you forward; they:
- Give you a talking point you can clearly articulate
- Create a thinking point you can reason from
- Provide a new anchor for connecting related concepts later
That’s why distinguishing between different kinds of gaps matters — because the right kind of deep dive can reframe your understanding in ways that feel less like “incremental learning” and more like unlocking a new tool in your thinking toolbox.
Level 1: Unknown Unknowns
You didn’t even know this existed until you stumbled on it.
- Example for me: Consistent hashing, CRDTs (the first time I encountered them)
- How they show up: Flipping through a textbook, overhearing a term in a podcast, stumbling on a slide in a conference talk
- Your move:
- Don’t deep dive immediately.
- Capture it in a “to explore” list.
- Give yourself time to circle back once you have more context.
- Why: At this stage, you don’t even have the right questions yet — a little incubation time makes later study much more productive.
Level 2: Known Unknowns (Anchor Gaps)
You know the topic exists, but something inside it doesn’t make sense.
-
Examples:
- “What really is the difference between CvRDTs and CmRDTs since a union (merge) is an operation?”
- “Why mod by the number of nodes when the output space is fixed?”
- My own struggle with
v-modelin Vue: I knew it was an alternate pattern to “emit to parent / pass to child” and had known that for ages — but after a deep dive, only fragments stuck. Without regular reinforcement, I slid back toward the recognition side of the spectrum.
-
How they show up:
Typically I find that these show up while doing the deep dive. Some part of it just doesn't make sense. You can't get a clean mental model based on what you've read and watched.
Why This Distinction Matters
Level 1 and Level 2 require different approaches:
| Gap Type | First Move | Goal |
|---|---|---|
| Level 1 | Capture & revisit later | Build a backlog of potential studies |
| Level 2 | Investigate immediately | Anchor the understanding with clarity |
By separating the two, you avoid wasting time deep-diving into something you can’t yet connect — and you ensure that real confusion doesn’t slip through the cracks.
Next Step: Closing Level 2 Gaps
Level 1 is about discovery. Level 2 is about resolution.
If you want a repeatable way to close Level 2 gaps, check out my upcoming article: The Clarity Loop — a simple framework for turning “this doesn’t make sense” into “I’ll never forget this again.”
--- for article 2
-
Resolving the Conflict:
Wikipedia is honestly a a great resource here because you can get a birds eye view of the subject.
Basically you can easily see that they divide the concept into a, b, c.
- Why are these divisions made that way?
- What are the defining characteristics of the categories?
Essentially at this point we are trying to deduce:
- Hard v. Soft boundaries
- MECE?
- Set / Subset / Superset, etc. (and then based on what characteristics)
It is at this point that you might find example that don't cleanly map. This is when we need a sparring partner (enter the LLM)
- Why some stick and some don’t:
- If you immediately start using the concept in real work, the anchor gets reinforced and lasts.
- If you resolve the gap but rarely revisit it, the forgetting curve will erode the clarity — leaving you with recognition but not reconstruction.
- Some concepts are second-class citizens in your workflow. They’re valid and useful, but you almost always reach for another pattern first (e.g., preferring a CTE chain over
HAVING, or using emit-to-parent overv-model). Unless you deliberately promote them to first-class status for a while, the gap may never fully close.
This is a huge reason I have the technical articles written... they are references for future me, so that future me doesnt' fall into the same exact micro-confusions/questions. The resolution is persisted.