Can AI compare how two people think?

What comparison means here

The question Rodin answers is not “do these two people agree?” It is “do they think the same way?” These are different questions, and the difference matters.

Two people can disagree on every conclusion they reach and still share the same reasoning structures — the same mental models, the same tolerance for unresolved tension, the same tendency to reason from first principles rather than authority. When they meet, the disagreement is generative. They can actually argue.

Conversely, two people who agree on everything may think in fundamentally different ways. Their agreement is coincidence, not architecture. There is nothing to build from it. Rodin is designed to find the first kind of connection — structural kinship — not the second.

How Rodin compares

1. 01

   Semantic similarity

   512-dimensional embeddings capture the meaning of your writing at scale. Two thinkers who keep returning to the same conceptual territory — even with different vocabulary — surface as close in this space.

2. 02

   Fingerprint overlap

   Jaccard similarity scores compare overlap across all 8 fingerprint layers — themes, open questions, mental models, intellectual DNA, blind spots, core question — with learned weights that reflect how diagnostic each layer is for genuine intellectual kinship.

3. 03

   Cognitive topology

   A 12-dimensional cognitive signature comparison captures how two people structure thought: epistemic confidence, argument density, abstraction level, temporal orientation, conceptual leap distance, and seven other structural signals. This layer is computed entirely from linguistic analysis — no embeddings, no external API.

4. 04

   Final score

   70% semantic similarity + 30% cognitive topology. The weighting reflects a deliberate choice: shared intellectual territory matters more than identical reasoning style. But the topology component catches cases where two people think the same way about entirely different subjects — the most surprising and generative matches.

What the comparison reveals

A Rodin comparison surfaces the themes both people keep returning to — not the ones they consciously identify as their interests, but the ones that appear across their writing regardless of ostensible subject. It shows the questions both are stuck on, the mental models both reach for, and whether their blind spots overlap or complement.

The most interesting comparisons show complementary blind spots — where one person's gap is another's strength. This is not similarity in the ordinary sense. It is a structural fit. Two thinkers whose cognitive architectures are complementary can do things together that neither could do alone.

An example

Consider a systems biologist and an urban planner who both use complexity theory as their primary conceptual lens. They write about entirely different subjects. Their credentials have no overlap. A keyword search, a LinkedIn recommendation algorithm, or a professional network would never connect them.

Rodin sees the shared cognitive structure immediately. The same mental models. The same tolerance for emergent behavior as an explanation. The same discomfort with reductive accounts. The fingerprints are structurally close — and the Cognitive Topology Analysis confirms it. They are intellectual kin working in different vocabularies.

Try it

Create your intellectual fingerprint by pasting your writing into Rodin — notes, essays, journal entries, an Obsidian vault export. Once your profile exists, visit the Compare page to see how your fingerprint aligns with other thinkers in the network.

The comparison is not a score to optimize. It is a map of intellectual proximity — useful for finding people worth talking to, collaborators worth seeking out, and conversations worth having.

Frequently asked questions

Can I compare myself to someone specific?

Yes. If they have a Rodin profile, visit the Compare page and select both profiles. The comparison runs across all three layers and returns a detailed breakdown — not just a score, but which fingerprint dimensions overlap, where your cognitive signatures align, and where you diverge.

What does a high similarity score mean?

It means you structure thought in similar ways — shared frameworks, similar reasoning patterns, overlapping open questions. It does not mean you agree on conclusions. Two people with a very high Rodin similarity score can hold entirely different views on every question they both care about. The score measures cognitive architecture, not opinion alignment.