Skip to main content

MAP context record

LumenUs helps automate tedious paperwork in times of grief | TechCrunch

Source
https://techcrunch.com/2026/10/09/lumenus-helps-automate-tedious-paperwork-in-times-of-grief/ · About this source
Published
2026-10-09
First observed
Oct 9, 2026 17:02 UTC
Last observed
Oct 9, 2026 17:33 UTC (2 observations)
Related MAP entity pages
No MAP entity page keyword appears in the headline

The seven frames

Origin

This content is hosted on techcrunch.com.

The current article identifies Amanda Silberling as the author, drawn from the page's metadata.

A publish date of 2026-10-09 was found in the page's metadata.

Attribution

The page contains 76 outbound links, including references to strictlyvc.com, facebook.com, twitter.com, linkedin.com, reddit.com.

No structural attribution phrasing (e.g. "according to," "officials say") was detected in the headline or description. This reflects only the headline/description text — attribution may still be present in the body content, which this record does not deeply parse.

Specificity

Approximately 914 words were extracted across 27 paragraphs of body content. (Heuristic: text pulled from tags in the fetched HTML, not a full reader-mode content pipeline.)

4 date-like references were found in the body content, including: 2026, October 9, 2026.

42 distinct capitalized name-like sequences were found, including: Book Exhibit Table Now, REGISTER NOW, TechCrunch Desktop Logo TechCrunch, Mobile Logo Latest Startups, Venture Apple Security AI — a structural proxy for named entities (people, places, organizations), not verified named-entity recognition.

Timing

The headline and description lean toward past-oriented framing (based on language such as "had").

This is a keyword-based structural proxy, not verb-tense or temporal-expression parsing.

Framing

The content follows a "standard article" presentation pattern.

The headline is structurally characterized as: moderate length, declarative.

Context Change

MAP holds 2 feed observations of this URL (first Oct 9, 2026 17:02 UTC, latest Oct 9, 2026 17:33 UTC). Differences, if any, are listed under What changed.

Information Relationships

76 outbound links were found in the content MAP extracted. MAP stores at most 10 referenced domains per record; 4 remain after removing the publisher's own site, social platforms and account or CDN hosts.

Domains referenced: strictlyvc.com, estateexec.com, lumenus.life, cnbc.com

What MAP observed, check by check

Frames observed: 7 of 7. A frame counts as observed when at least one of its rows below says Found. This is a tally of what was found on the page, not a rating of the page or of its publisher.

For each frame, what MAP's automated extraction found on the page it read, one check at a time. “Not found” means the item was not in the page metadata or text MAP read; it does not show that the page, the publisher or the reporting lacks it. “Not read” means MAP could not check the item. Rows about MAP's own observations or filters use their own plain words where nothing was found. MAP does not score, rate or grade pages.

Show all checks

Origin

Publisher domain
Found — techcrunch.com
Author or byline in the page metadata
Found — Amanda Silberling (from a meta tag)
Publish date in the page metadata
Found — 2026-10-09. MAP stores the date only, not the time.

Attribution

Outbound links in the extracted content
Found — 76 outbound links.
Attribution wording in the headline or description
Not found — No wording such as “according to” was found. MAP checks the headline and description only, not the body.

Specificity

Body text extracted
Found — About 914 words in 27 paragraphs.
Date-like references in the body text
Found — 4 date-like references.

Timing

Past, present or future wording in the headline or description
Found — Past-oriented wording. This describes the wording, not when the event happened.

Framing

Presentation pattern
Found — standard article. Default pattern when no other format matched.
Headline structure
Found — moderate length, declarative

Context Change

Repeat observations of this URL
Found — 2 observations, first Oct 9, 2026 17:02 UTC, latest Oct 9, 2026 17:33 UTC.
Differences between those observations
None listed — No differences were listed between MAP's like-for-like observations.

Information Relationships

Referenced domains after MAP's filters
Found — strictlyvc.com, estateexec.com, lumenus.life, cnbc.com. At most 10 domains are stored per record, and MAP then removes the publisher's own site, social platforms and account or CDN hosts from the list shown.

Outside the scope of a MAP record

  • Earlier events, votes, filings or contracts connected to the story. A MAP record does not look these up or link them to the story, and it does not judge what a story leaves out.
  • Coverage of the same event by other outlets. MAP does not compare coverage across outlets.
  • Later edits or corrections by the publisher, other than differences MAP happened to observe between its own repeat observations.
  • Who the people and organizations named on the page are. The name-like sequences in the Specificity frame are a structural proxy, not identified names.

See the Transparency Explanations page and the Methodology.

What changed

No differences were listed between MAP's like-for-like observations of this URL.

Other recent records from this source

Provenance

Record ID
b8cc1790-04f8-4862-b64d-e2d64008da39
This record observed
Oct 9, 2026 17:33 UTC
Ledger position
36730
Record hash (SHA-256)
6a31655d1c8a4757942b2329082f7d20fbb0196db2dd8bfae74f4d69ae2c626c
Previous hash
161024314bf57b9db5603f339afeab32dc18781da5d5f361dfb8c2298f26919d
Link to previous entry
continuous

MAP operates this ledger. The hashes are published here for reference; MAP does not publish the data needed to recompute them.

Continuity check: MAP runs a continuity check on the ledger every day (last run Oct 25, 2026 10:00 UTC). MAP does not publish its detailed results or the data needed to recompute the hashes.

A record reflects the page as MAP fetched it at one moment. MAP does not determine whether content is true, false, biased, or trustworthy. Methodology · JSON