When I first started exploring claim-level discussion (highlighting individual sentences and voting on them), the idea felt abstract.
So I tried grounding it in real situations where existing tools struggle.
Here are three that keep coming up:
1. Policy & Civic Discussion
Problem:
Long proposals get reduced to binary support/opposition.
What breaks:
People agree with parts, disagree with others, but have no way to express that nuance.
What changes:
Each claim can be evaluated independently (agree/disagree, true/false), creating a map of where consensus actually exists.
2. Open Source RFCs & Design Docs
Problem:
Feedback lives in comment threads or PR discussions, hard to synthesize.
What breaks:
Important critiques get buried, and it’s unclear which ideas are broadly supported.
What changes:
Specific lines or claims can accumulate structured feedback, making it easier to see what holds up.
3. Research, Journalism, and Analysis
Problem:
Readers react to entire articles, not specific statements.
What breaks:
Claims that are contested or unsupported don’t get isolated clearly.
What changes:
Readers can attach agreement, disagreement, or evidence directly to individual statements.
What This Enables
Instead of:
- ranking posts
- arguing in threads
You get:
- structured evaluation of ideas
- visibility into agreement vs disagreement
- a dataset tied to language itself
I’m building this as an open-source project called Quote.Vote.
If this resonates, I’d love your feedback:
Where do you see this working — or breaking?
Top comments (0)