Jag's AI Heat Map turns AI news into an interactive network of companies, models, products, people, hardware and ideas. Bigger nodes indicate more attention, while links show which entities appear together in stories. A Hacker News submission describes the project as drawing from more than 140 news feeds.
The useful difference from a chronological headline list is the ability to explore relationships. Readers can spot a prominent model, examine its connections and see whether its attention is building or fading.
Dropday's listing attributes the project to @jurlycat and dates the associated X post to October 25, 2026. The dashboard itself provides no publication date, so that listing is an announcement reference, not confirmation of an exact launch timestamp.
What the Map Measures
Every story contributes heat to the entities it mentions. According to the dashboard's documentation, those contributions are weighted by source and popularity signals, including Hacker News points, Hugging Face upvotes and GitHub stars. Heat then decays over time.
That makes the map an attention tracker. A prominent node means an entity is attracting coverage and community activity within the dashboard's sources. It says nothing directly about model accuracy, product reliability or commercial success.
The map's layout offers several ways into that activity:
- Size and position: Bigger nodes closer to the center indicate more discussion.
- Connections: Links join entities mentioned in the same stories.
- Shapes: Topics appear as diamonds, distinguishing them from other entities.
A link needs particularly careful reading. Two companies connected on the map may appear in a comparison article, a competitive announcement or a shared industry story. The connection alone says nothing about a partnership or technical compatibility.
I'd use these links as prompts for further reading. Their value is showing where coverage overlaps, without requiring readers to know the relevant names beforehand.
Choose Week for Persistence, Today for Fresh Movement
The Week and Today controls change both the observation window and how quickly older attention loses influence. Today means the latest 24 hours, not simply the current calendar day.
| View | Coverage window | Heat half-life | Useful for |
|---|---|---|---|
| Week | Seven days | 24 hours | Seeing which subjects have sustained attention across several days |
| Today | 24 hours | Six hours | Finding recent shifts that may be less visible in the broader view |
These settings come from the site's explanation of heat. A half-life is the time required for a contribution's heat to fall by half through decay.
Consider one story's initial contribution, with no additional activity. After 12 hours, about 71% would remain under the weekly view's decay rate. Under the daily view's rate, only 25% would remain. Those percentages are calculated from the documented half-lives, not measured dashboard results.
The practical consequence is that Week still favors recent activity. It is not an equally weighted total of everything published over seven days. Today applies an even stronger recency filter, allowing a fresh burst to stand out more quickly.
I'd start with Week to establish the larger picture, then switch to Today to look for changes. A subject that remains prominent in both views deserves a closer look because it has both accumulated attention and recent activity. That is a useful reading strategy, not a claim about its underlying importance.
Read the Markers Before Following the Color
Orange has several meanings on the map. A newly detected entity, an accelerating entity and the hottest topic use different visual treatments.
The dashboard legend defines them as follows:
| Marker | Documented meaning |
|---|---|
| Orange name with a pulsing ring | Breaking: at least two stories and the required six-hour attention surge |
| Orange ring | New: the entity was first spotted this week |
| Up arrow | In the general legend, heat is at least 1.5 times its level 24 hours earlier |
| Down arrow | The entity's heat has cooled |
| Orange-filled topic diamond | The hottest topic; its closest connections are outlined |
New is a discovery label, not a launch date. An established company or older AI product can be newly spotted by the dashboard. Readers checking whether something actually launched should open the underlying coverage.
The arrow rules also need a small qualification. The general legend specifies a 24-hour comparison for the upward marker, while the Today description says it compares with six hours earlier. The short documentation does not explicitly explain how the arrow threshold applies in that view. It also gives no numerical cutoff for the downward marker.
A cooling marker is therefore best read as declining attention. It is not evidence that a product has deteriorated or that an earlier report was wrong.
What the Breaking Threshold Actually Tells You
The breaking marker has a concrete rule: an entity must receive at least two stories, with attention in the latest six hours reaching 2.5 times that of the preceding six hours, according to the project documentation.
That combines a minimum story count with a relative-growth test. It helps separate a surge from a single isolated mention, but it does not make the marker an editorial judgment about importance.
A worked example shows the distinction:
- An entity rising from four attention units to 10 reaches the 2.5-times threshold.
- An entity rising from 100 units to 200 doubles its attention but falls short of that threshold.
These are illustrative units, not published dashboard scores, and both examples assume the story-count requirement is met.
The smaller subject could therefore earn a breaking marker while the much more heavily discussed subject does not. That is the expected behavior of a relative-growth signal.
My view is that the badge is most useful as a discovery cue. It can surface an emerging topic that would otherwise be overshadowed by large companies. Node prominence supplies the missing context: how much attention the subject has overall.
Follow the Signal Back to Its Stories
The map supports scrolling to zoom, dragging to pan and clicking a node for details. Those controls help readers navigate a dense network without treating its default view as the whole story.
The Articles button in the top-right corner opens a separate archive covering the last 90 days, grouped by source with newer stories first. Readers can select a source on the left and open individual headlines in a new tab. The also explains that the small red number beneath the button counts articles discovered in the last hour.
Frequently Asked Questions
3 questions
1Does a bigger node mean a better AI model?
No. Size reflects attention within the dashboard's weighted sources, not model quality or benchmark performance.
