A LinkedIn post about outliners

Click a heading to fold it · drag the dots or press Alt+arrows to reorder

An outliner hands the writing back to you after a long conversation with AI

A LinkedIn post about the interactive-outline-map skill, illustrated with this map.

1Outliners in the early nineties

The opening, anchored in experience.

almost as hot as “thinking aids” as LLMs are today
I used ThinkTank for a while, and later MORE 3.0
The same promise then and now: the machine helps you think
2The iteration trap

Why a long conversation with a model makes writing harder.

Finished-sounding text makes your own version harder than at the start
Easiest next step: share the model’s text
Your own voice goes missing unnoticed
3The skill: from conversation to map

What the skill does, and why it holds itself back.

a way to turn conversations with AI into a compact form
Breaks it down into claims, tensions, examples and open questions
Stops one step before finished prose
Only branches move — the point is to start writing
Under 20 words a card: a finished sentence invites copying
4The slowdown is the point

Writing by hand from the map.

one more slowdown, and it gives the brain room to notice gaps in the reasoning
Fluent prose hides the gaps: it sounds done before the thought is
Putting it in your own words exposes a missing reason or a weak example
5A claim to disagree with

The edge is in the argument, not in the last line.

LLMs first sped up producing text
The bottleneck was building the argument, not producing text
The machine helps most when it leaves the last step undone?
6Open

Unresolved before publishing.

The draft is the model’s — do I write the post myself from this map?
Do the nineties carry the argument, or are they just a hook?
For whom: SaaS leaders, or experts who write with AI?
Link to GitHub in a comment, or not at all?