method

the lanes

the corpus was produced by two parallel agent research lanes — company history and the people behind it, and coverage, sentiment, and the fast-editor idea — each writing a lane report against one shared scratch source catalog before synthesis. the pattern is the company-research formula proven on roam-research and antithesis: lane-prefixed source ids frozen upfront, per-claim confidence, and a declared deduplication owner. synthesis merged overlapping lane-minted ids into one canonical catalog.

the evidence rules

1. every event, quote, and claim carries at least one source id.
2. dates keep their precision — a year, a month, or a day — and approximate dates are labeled rather than sharpened.
3. confidence stays honest: confirmed, reported, and inferred are distinct claims, not stages of the same claim.
4. sources are typed by role — primary, interview, reporting, archive, community — so a claim's foundation is visible before it is read.
5. sentiment labels describe the coverage, not the truth of it.

known gaps

1. adoption figures are self-reported or investor-quoted — sequoia's "150k active developers each month," the 1.0 post's "hundreds of thousands of developers," and delta's "570 changes to main" have no independent measurement.
2. the series b valuation is undisclosed; total raised is reported as "over $42m" but no post-money figure is public.
3. the seed's exact date splits across trackers (april 6 vs april 20, 2021) and no lead investor is publicly stated.
4. the warp ip separation is undocumented — sobo carried the codebase into warp for ~three months; how it came back out is not on the record.
5. headcount varies by source — the september 2026 team page lists ~40; linkedin and tracker counts disagree.
6. extension counts vary by registry versus tracker (~1,119 listed versus ~1,562 tracked including unlisted); the vs code comparison is approximate.
7. revenue and customer counts are undisclosed — pro, business seats, and planned delta plans are priced, but the take is not.
8. collab adoption inside zed is only admitted in passing ("very few people leverage collab today"); no usage data was published before the features moved to delta.
9. the x/twitter corpus is partial — tweet-level material came through archives and interviews, not native search.

corrections

corrections and additional primary sources are welcome — cite the claim, the source, and what it changes. the catalog is designed to be audited, not just read.