Source profile
ExtremeTech
✓Hardware and technology coverage with technical depth. Often covers CPUs, GPUs, chips, and computing trends. Good for analysis around performance and industry shifts. Useful for enthusiast and enterprise-adjacent hardware news.
Official source: NoMainstream media: NoLast activity: 15h ago
20 items analyzedVisit site ↗
Insights Metrics v2
Metric board
Publisher typeWhether the publisher is an official company channel or a mainstream outlet. Best-effort editorial judgements, hand-set per source.
24h
independent
Sample: 6Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
7d
independent
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
independent
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Articles publishedThe number of final articles from this source whose event date falls in the window, with the change from the previous window. Counts what we collected; sources run on different schedules.
24h
6
Sample: 6Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
7d
21was 24
Not enough data for a change figure
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
45was 0
Not enough data for a change figure
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Active daysThe number of days in the window with at least one final article from this source. For sources that rarely expose a publish date, this is based on collection date.
24h
- Active days
- 1
- Dated share
- 0.0%
Sample: 6Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
7d
- Active days
- 5
- Dated share
- 0.0%
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
- Active days
- 6
- Dated share
- 0.0%
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Topic mixThe five categories that appear most often in the source's final articles. Same generous multi-labelling as E7.
24h
Not enough data
not offered for this window
7d
AI38%
Business24%
Gadgets24%
Security24%
Gaming19%
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
Gadgets40%
Hardware33%
AI31%
Business18%
Security18%
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Content-type mixWhat kind of pieces the source publishes. Coarse AI labels - labelled 'AI-classified article type'.
24h
Not enough data
not offered for this window
7d
News reporting48%
Untyped33%
Announcement19%
Breaking news0%
Analysis0%
Other0%
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
News reporting62%
Untyped20%
Announcement18%
Breaking news0%
Analysis0%
Other0%
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Tone mix and net toneThe distribution of article-level sentiment in the source's final articles, plus net tone. Official channels are self-promotional by nature.
24h
Not enough data
not offered for this window
7d
Not enough data
not enough data (14 of 20)
30d
Neutral38%
Positive30%
Negative16%
Mixed16%
Net tone: Positive · 0.14
Sample: 37Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Most-covered entitiesThe top 5 linked entities in the source's final articles. Catalog gaps and disabled forms under-count.
24h
Not enough data
not offered for this window
7d
OpenAI14%
Alphabet10%
Android10%
NVIDIA10%
IPhone10%
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
Alphabet11%
Android11%
Gemini9%
OpenAI9%
Apple7%
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Regional focusWhich parts of the world this source's coverage concerns. The AI's reading of the article's geography, not the outlet's country of registration.
24h
Not enough data
not offered for this window
7d
Global 42.9% · Unspecified 4.8%
US55%
EU18%
China9%
Sample: 11Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
Global 71.1% · Unspecified 2.2%
US50%
EU17%
__no_region__17%
Sample: 12Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Audiences coveredWhich audiences this source's articles say are affected. Consumers is the corpus default; read every number as a deviation from it.
24h
Not enough data
not offered for this window
7d
Consumers62%
Others33%
Developers24%
- Consumers+17.8pp vs baseline
- Others+7.8pp vs baseline
- Developers+8.3pp vs baseline
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
Consumers69%
Developers42%
Others24%
- Consumers+24.3pp vs baseline
- Developers+28.4pp vs baseline
- Others+5.8pp vs baseline
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Sourcing-strength mixHow often the AI judged an article's own sourcing to be strong versus thin, for this outlet's articles. The model's assessment of one article's sourcing, made without browsing. Never a publisher reliability or trust score, never used to rank sources.
24h
Not enough data
not offered for this window
7d
Medium62%
Low24%
High14%
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
Low42%
Medium33%
High22%
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Confidence mixHow sure the model was about its own reading of each article. Uncalibrated self-confidence, descriptive only.
24h
Not enough data
not offered for this window
7d
Medium62%
Low33%
High5%
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
Low49%
Medium40%
High9%
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Stance profileWhere the stories sit in a trend's life - accelerating, mature, or an early signal. An AI reading of the article's framing, not a market judgement.
24h
Not enough data
not offered for this window
7d
Accelerating48%
Early signal33%
Mature19%
Other0%
Sample: 21Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
Early signal47%
Accelerating40%
Mature11%
Other0%
Sample: 45Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Event types coveredWhich kinds of events this outlet reports on. Same 2026-08-10 start date and New-Product skew as E12.
24h
Not enough data
not offered for this window
7d
Not enough data
not enough data (2 of 20)
30d
Not enough data
not enough data (16 of 20)
Evidence-basis mixHow much of this outlet's coverage is built on official confirmation versus rumor, anonymous sourcing, or opinion. Correlated with the outlet's beat, never used as a trust ranking.
24h
Not enough data
not offered for this window
7d
Not enough data
not enough data (15 of 20)
30d
31 rated (68.9% of coverage)
On record sources42%
Official announcement39%
Opinion13%
Analyst estimate3%
Rumor or leak3%
Sample: 31Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Story-phase mixWhether this outlet mostly breaks stories first or mostly follows up, reacts, and explains. Orthogonal to S5's content-type mix by prompt design.
24h
Not enough data
not offered for this window
7d
20 rated (95.2% of coverage)
First report95%
Reaction or analysis5%
Sample: 20Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
30d
43 rated (95.6% of coverage)
First report93%
Reaction or analysis7%
Sample: 43Sample size — how many underlying data points (e.g. labeled mentions, rated articles, or statements) this metric was actually computed from. Can be smaller than the entity or source's total coverage count.
Latest coverage
CloudProgramming
Googlebook OS desktop with Android apps displayed
Article15h
AIRegulation
OpenAI Follows Anthropic With Watermarked Text in the EU
Article17h
AIBusiness
OpenAI Starts Displaying Ads Next to ChatGPT-Generated Images
Article17h
Security
Illustration of a magnifying glass finding a bug on a computer
Article17h