Mesh

Why the X Algorithm Is Bad at Breaking News

Written By

TT Staff

Published

Aug 19, 2026

Time to Read

4 MINS

The X algorithm is good at its job. Its job is to keep you on X, and it does that by showing you posts likely to produce a reaction. That objective is reasonable for a social product and directly opposed to what you need if you are using the platform to find out what just happened.

Engagement and recency are different sorts

A ranked timeline promotes posts that have already performed. Performance takes time to accumulate, so a post needs to exist for a while before the system can tell it is doing well.

The consequence is structural rather than accidental. The moment a post is most valuable, in the first seconds, is the moment the algorithm has the least evidence it is worth showing you. By the time it has that evidence, the information is old. For time sensitive news, the ranking system is working against the use case by design.

What this looks like in practice

You open the app after something has happened. The first thing you see is a thread explaining what it means, posted an hour ago, with high engagement. Below that is a joke about it. Somewhere further down is the original report.

You have received the reaction before the event. That ordering is fine when you are browsing and actively unhelpful when you are trying to establish a sequence, because commentary about a thing reads very differently once you already know what the thing was.

The secondary problems

Recommended posts. Accounts you did not follow appear because they performed with people like you. Sometimes useful, always unpredictable, and impossible to rely on for coverage of a specific beat.

You cannot tell what you missed. A ranked feed shows a selection, and there is no indication of what was left out. If a beat writer you follow posted three times and you saw one, nothing tells you the other two existed.

Everything else is one tap away. Notifications, direct messages, and the recommendation surface all sit alongside the thing you opened the app for. That is the design working correctly and it is why five minutes becomes twenty.

Chronological is not a preference

People describe wanting a chronological feed as nostalgia. For monitoring it is a functional requirement, because order carries information.

Knowing that account A reported something and account B confirmed it four minutes later is a different piece of knowledge from knowing both said it. Sequence tells you who broke it, how fast it spread, and whether a denial came before or after a confirmation. A ranked feed destroys that ordering, and there is no way to reconstruct it from what you are shown.

Working around it

Lists are the built in answer and they work. Chronological, no recommendations, free. The limits are that you view one at a time and the rest of the app remains one tap away, which is covered in more detail in the comparison.

A ticker takes the same idea further by removing the app entirely. Chosen accounts, in order, on a screen you are not touching. The behavioural difference matters more than the technical one: you are no longer opening something, which means you are no longer deciding when to stop.

Neither fixes the underlying tension. The platform is optimised for attention and you want information. Both approaches work by opting out of the ranking rather than by improving it.

Common questions

Does the Following tab fix this?

Mostly, and it is the right first step. It is chronological and limited to accounts you follow. It still puts you inside the app, and it covers everyone you follow rather than the handful relevant to a given beat.

Is the algorithm getting worse?

It is getting better at its objective, which is a different thing. Optimising harder for engagement makes the mismatch with time sensitive monitoring more pronounced, not less.

Do I have to give up the normal timeline?

No, and most people should not. The ranked feed is genuinely good at showing you things you did not know to look for. The mistake is using it for the job it is worst at.

Personnel Log

TT Staff

The TweetTicker Editorial Team provides real-time insights for high-velocity data environments.