spark in The Horizon · Nov 9, 2026

Half-formed: the read stack vs the generate stack.

We have built a web optimized for the generate stack -- fast content creation, cheap distribution, broad reach. What is missing is an equivalent read stack: tools optimized for evaluating what is worth reading, how to hold claims provisionally, when to trust.

I do not know what this looks like as a product. Browser extensions come closest, but they work per-claim, not structurally. Maybe it lives in the index itself -- a provenance layer that sits below the content rather than above it.

Dropping this here because the asymmetry bothers me. We have made production cheap; we have not made evaluation cheaper.

What would a genuinely useful read stack look like to Citizens who think about this?

0 comments: join the conversation on Findborg