How we decide what's worth reading

Every post we index gets read, sorted into one of four sections, and scored out of ten. People ask how that works often enough that it's worth writing down, partly because the rules are the editorial product and partly because disagreeing with them is useful to us.

The four buckets

Investment thesis. The piece names an opportunity, a market or a theme that cuts across sectors, and argues that's where value will be created. "Request for startups" posts, "why now" essays, roadmaps with actual bets in them. This is the smallest section. In a typical week, three to five posts out of a hundred qualify.

Industry analysis. A specific, named market explained: who the players are, what the data says, what changed. Earnings breakdowns, landscape maps, state-of-the-sector reports.

Operating insight. Something a founder, operator or investor can apply. Playbooks, how-tos, teardowns of how a company priced or hired or scaled, frameworks for moats and defensibility. The test is whether you could act on it this quarter.

Additional reads. The largest section, and deliberately so. Essays, investor philosophy, career and culture writing, observations that span ten industries at once. These are the pieces that don't fit a box and are often the most interesting thing in the issue.

The score

Zero to ten, for how much the piece actually says. Length is not the input. A 600-word post with one original idea beats a 4,000-word summary of things you already knew.

Roughly: 0-1 is housekeeping and marketing. 2-3 is announcements, single-company profiles, podcast transcripts. 4-5 is commentary that restates known information. 6-7 is a solid argument with a useful frame. 8-9 is original research, real data, or a playbook with specifics. Ten is something people will still link to in three years.

What gets cut

The exclusions are where the editorial judgment really lives, and they're stricter than most aggregators.

Funding announcements are out, including the essay-shaped ones. "Why we invested in X" written by X's investor is marketing with footnotes, however good the market section is.

Profiles of a single early-stage startup are out, for the same reason.

Podcasts and transcripts are out. A 12,000-word verbatim interview with sponsor reads isn't an article. This rule costs us some genuinely good conversations and we keep it anyway.

Link roundups are out unless a real share of the issue is the author's own analysis.

Paywalled posts are out. If you can't read it, we shouldn't recommend it.

Where it goes wrong

The hardest call is industry analysis versus additional reads. The test we use: does the piece analyse a named market, or would its observation apply to ten different sectors at once? "The state of AI in healthcare" is analysis. "Why does everyone hate AI" is an essay.

The second hardest is the thesis bucket, which attracts everything with a prediction in it. A prediction isn't a thesis. If you can't finish the sentence "the opportunity is ___, because ___", it belongs somewhere else.

We re-read the week before it goes out and move things. If you think we've filed something wrong, say so. We've rewritten these rules four times this year, usually because a reader pointed at a specific post and asked what on earth it was doing in that section.

You can see the sections in practice by browsing a topic or looking at the front page, which ranks by the same score.