Stories

X has a repetition problem. A single account will post the same story forty times in a day, rephrased each time. Forty accounts will post that same story once each, within twenty minutes of one another. Either way your feed reads as forty events when there was one, and the actual news is buried under its own coverage.

Stories is the answer to that. It watches everything MBER collects, works out which posts are about the same real-world event, and folds them into a single story: one headline, one summary, and the list of accounts covering it. Instead of scrolling past the same thing forty times, you read it once and see at a glance how many people thought it was worth posting.

How a story forms

Every post that comes through the pipeline is embedded and compared against the posts around it. Near-identical posts (wire copy, quote tweets, the same take reworded) collapse into one group immediately. Groups that look like they might belong to an existing story go through a second check, because being about the same topic is not the same as being about the same event: two posts can both be about a central bank without being about the same decision. That check asks, per group, whether this is the same specific event as one of the candidate stories. If it is, the posts attach to that story. If it is not, a new story starts.

The upshot is that the count on a story card means something. It is the number of distinct accounts that independently posted about that event, not the number of times one account hit send.

Stories improve as coverage grows

A story is written from its coverage, so it gets better as more of it arrives.

A story earns its headline and summary once two distinct accounts have covered it. As coverage widens, the write-up is regenerated from the full set of posts rather than patched: once five accounts are on it, and again with every four posts after that. The summary also lengthens with coverage, from a single line on a thin story to two paragraphs on a heavily covered one, with the second paragraph carrying the newest developments and reactions.

So an event that breaks with one post and ends the day with thirty is not thirty cards in your feed. It is one card that started as a line and grew into a briefing, with the timestamp and the count telling you how far it has run.

The world map

Stories are placed on the map by the countries their posts are tagged with. The badge on a continent sums the stories inside it.

• Click a continent to zoom in and see per-country counts.
• Click a country to filter the panels below to its stories.
• Click the ocean to zoom back out.

Not every story has a place. Plenty of what MBER tracks is not about anywhere in particular, and those stories stay unplaced rather than being guessed onto a country. The map is a filter, not a census.

Three panels

Below the map, stories split into three feeds by their primary category:

Panel	Covers
News	General news, politics, macro, everything outside the two below
Crypto	Markets, protocols, tokens, exchanges
AI & Tech	Models, launches, research, the industry around it

The split follows the same taxonomy the Categorization engine assigns to posts.

Story cards

Each card carries:

• The headline, written from the coverage.
• The summary, behind the chevron. Click the card to open it.
• The consensus stack: the avatars of the accounts covering the story, with the total count next to them. Each avatar is a link that opens that account's post about the story on X, so you can always get back to the source. Click the count to open every post behind the story in a side panel, sorted by MIX or grouped per account.
• When it was last seen, so you know whether the story is still moving.

Sorting

The same three-way sort as the Timeline, applied to stories:

Sort	Order
MIX	Engagement weighted against recency. The default.
TIME	Latest developments first.
SCORE	Highest engagement, ignoring recency.

note

Stories reads the same collected posts as the rest of MBER, so anything you see here also exists in the Timeline. Stories is the consolidated view of it, not a separate source.