NOT FINANCIAL ADVICE · EVERY CLAIM IS SOURCED · NOTHING HERE IS A RECOMMENDATION TO BUY, SELL, OR HOLDNOT FINANCIAL ADVICE · EVERY CLAIM IS SOURCED
Century Worth research
Loading sourced intelligence
Reading the public evidence graph
Retrieving a bounded, validated view from the Century Worth API.
Pipeline quality
Live Research Engine
Recent system work and measured quality, derived from durable records rather than a seeded demo. These numbers describe the pipeline, not the companies.
Operating with degraded sources
In the last 24 hours, the engine checked 3 live sources, archived 45 unseeded artifacts, found 0 candidates, and published 0 findings from non-seed evidence.
Generated
Activity
What the engine did during this window. Counts are not quality scores.
Source checks
75
Artifacts archived
47
Candidates found
24
Automatic publications
0
Provisional publications
0
Manual publications
24
Research-only abstentions
0
Reviews
0
Revisions
0
Warnings
5
Failures
0
Operating performance
Rates use only records in the selected window. Missing denominators stay unknown.
Artifacts per check
0.63
Candidates per artifact
0.51
Recent publication rate
100.0%
Warning rate
6.7%
Failure rate
0.0%
Review agreement rate
Not enough data
Aggregate disclosure
Research quality
public-quality/v1
Counts remain public, but rates and calibration are withheld until their denominator reaches 20. This reduces false precision and makes small operational buckets harder to reverse-engineer.
Verified evidence coverage
0.0%
0 verified of 24 published; 24 not recorded
Automatic / provisional / manual
0 / 0 / 24
24 publications; shares 0.0% / 0.0% / 100.0%
Correction rate
0.0%
0 append-only corrections among 24 publications
Contradiction rate
Suppressed until 20 observations
0 contradictions among 0 public follow-ups
Review agreement
Suppressed until 20 observations
0 agree, 0 disagree, 0 reviewer abstentions
Research-only abstention
Suppressed until 20 observations
0 research-only decisions among 0 policy evaluations
Prediction calibration
Latest full-ledger rebuild, separate from the selected activity window.
Resolved sample
0
Hit rate
Suppressed
Forecast average
Suppressed
Excluded
0
Calibration values are suppressed until 20 resolved predictions exist.
Activity metrics use the selected UTC window; prediction calibration uses the latest full-ledger rebuild.
Small-sample rates are withheld while their aggregate counts remain visible.
Policy evaluations are counted as decisions, so a later reevaluation may add another record.
Corrections are append-only supersessions; contradiction measures include only publicly eligible follow-up evidence.
Coverage reflects tracked and reviewed research, not the full market or all AI developments.
Private candidates, internal identifiers, per-source and per-claim buckets, reviewer notes, policy inputs, and kill-switch state are excluded from this public view.
Source freshness
Source
State
Last outcome
Last checked
changelog-openai
never checked
Never checked
—
changelog-stability
never checked
Never checked
—
company-earnings-chegg
never checked
Never checked
—
company-newsroom-chegg
never checked
Never checked
—
company-product-openai
never checked
Never checked
—
customer-review-openai-trustpilot
never checked
Never checked
—
edgar
healthy
unchanged
7/26/2026, 11:00:09 PM
edgar-financials
degraded
warn
7/26/2026, 9:07:40 PM
marketstack
never checked
Never checked
—
news-rss
healthy
unchanged
7/26/2026, 11:30:20 PM
workforce-layoffs-fyi
never checked
Never checked
—
Public-safe details
Operational errors, internal identifiers, private candidates, reviewer notes, and unpublished claims are intentionally excluded.