Skip to lesson
Exit
Developer Marketing & Product Fluency1 / 3

1 min lesson

The surfaces few candidates name

Explain the practical point behind "What separates a candidate who actually lives in the product is the underknown surfaces."

Step 1 of 3

The surfaces few candidates namedepth that reads as real usage

Naming Tab, Agent and is table stakes. What separates a candidate who actually lives in the product is the underknown surfaces. Drop one or two of these unprompted and the panel registers you as a real user, not someone who skimmed the marketing site.

Learn more

Full explanation

Outcomes, not adjectives

Plan and Debug modes

The mode dropdown changes the toolset, not the model. Plan mode researches the whole codebase and outputs an editable Markdown plan with zero edits until you switch to Agent; editing that plan costs no tokens.

Debug mode walks you through reproducing an issue and adds instrumentation. Knowing mode = toolset (Agent / Plan / Debug / Ask) is the kind of detail that proves you drive the product daily.

@-mentions, including past chats

@-mention guarantees a specific file, doc or symbol lands in context instead of relying on semantic search to find it.

It isn't just files: referencing a past chat summarizes it and reads the transcript, so a new agent inherits the summary without carrying all the prior context forward.

The built-in browser

Cursor ships a full browser inside the editor that renders your running app. Give a plain-language directive ("increase the padding on the overview heading") and watch it render while Cursor surgically edits only the necessary files.

Agents can self-test in it: clicking around, taking screenshots. It's the front-end demo beat that lands hardest with engineers.

Canvas

A recently shipped output surface that renders visual components beyond Markdown or Mermaid, like a coverage report or an accessibility result.

You direct an agent to "output the results in a nice canvas." Almost no candidate knows it exists, so naming it signals depth.

Knowing the surfaces is table stakes. The harder skill is translating each one into an outcome a developer feels.

Surface
Tab
The job it does
Finishes the edit you were about to make.
The outcome you sell
Less typing, fewer trips to docs, momentum held.
Surface
Agent
The job it does
Takes on a whole multi-file task.
The outcome you sell
Tedious changes done while you stay on the hard part.
Surface
The job it does
Targeted edit in the current file.
The outcome you sell
A refactor without breaking flow or context-switching.
Surface
Codebase context
The job it does
Grounds answers in your real repo.
The outcome you sell
Fewer hallucinated APIs, so engineers actually trust it.

Outcomes, not adjectives. "Time saved" and "stay in flow" are the language engineers accept.