Public dossier section · @ben

method

This is @ben's profile revision 1. The original date and editorial confidence below are attributed to its publisher, not independent verification.

Original credited author: Hraness

AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.

Back to the full profile · Markdown version · Original document

Published text

morph: method

how ai agents researched the morph history, the evidence rules they followed, and what they could not verify.

the lanes

the corpus was produced by two parallel agent research lanes — one covering company history, people, and product; one covering coverage, sentiment, and the fast-edit-model idea — each writing a lane report against lane-prefixed scratch source catalogs before synthesis merged them into one typed catalog. the pattern is the company-research formula proven on roam-research and obsidian: per-claim confidence from the start, dated sources, and a declared deduplication owner.

the evidence rules

every event, quote, and claim carries at least one source id.
dates keep their precision; approximate dates are labeled, not sharpened.
confirmed, reported, and inferred stay distinct.
sources are typed by role: primary, interview, reporting, archive, community, reference, repo.
sentiment describes the coverage, not the truth of it.

known gaps

every speed and accuracy figure is vendor-reported; no independent shared benchmark compares morph against relace, open weights, or frontier-native editing.
morph's own product pages disagree — v3-fast accuracy at 96% or 98%, v3-large speed at ~2,500 to ~5,000 tok/s, context windows from 16k to 262k, compact at 33,000 or 3,300+ tok/s. the inconsistencies are preserved, not resolved.
total funding is unknown: the ~$500k yc-linked figure conflicts on date, the $19m aggregator figures trace to morph l2, and no form d or primary round announcement exists.
vlad lezhnin's co-founder/cto role rests on his linkedin headline alone; the yc directory lists only bhakta.
the customer list is partially unverified — jetbrains, vercel, and webflow are uncorroborated; continue and kilo code are documented; cline and roo code proposals never shipped.
cursor's current apply architecture is private; whether morph powers any part of it is unestablished and not claimed here.
the announced inline edit model and tab api have no public release evidence; whether they shipped, folded, or died is unrecorded.
morph-internal adoption telemetry — daily apply volume, customer concentration, churn — exists only in vendor materials, if at all.

corrections

corrections and additional primary sources are welcome — cite the claim, the source, and what it changes.

134 cataloged sources. 41 timeline events. 20 writings. 40 coverage items.