Open your phone.
A shooting.
A war.
A racist video.
A political scandal.
Someone screaming at a stranger.
Someone else screaming about the person screaming at the stranger.
Thousands of replies telling you America is finished.
Now put the phone down and go buy coffee.
Someone holds the door.
A cashier says good morning.
A guy walks his dog.
Two people laugh in a parking lot.
Nobody is fighting about the election.
Which world is real?
Both of them.
And that might be the problem.
Your Feed Is Not a Random Window Into Earth
Imagine every city in America sent you its worst moment of the day.
One murder from Chicago.
One racist incident from somewhere else.
One insane road-rage video.
One teacher caught saying something horrible.
One political protest that turned ugly.
Every event could be 100% real.
Put them all into one endless feed and watch for two hours.
How do you think America would feel afterward?
Dangerous.
Angry.
Broken.
But you would not have watched an average day in America.
You would have watched a highlight reel of things extreme enough to travel.
And research suggests negative information really does travel differently.
A 2024 study looked at more than 95,000 news articles and about 579 million Facebook and Twitter posts.
Negative news links were about 1.91 times as likely to be shared as non-negative ones in that dataset.
Not because positive things never go viral.
They do.
But negativity has an advantage.
And Then the Algorithm Enters the Room
This is where people usually say:
"See? Algorithms are brainwashing everybody."
Slow down.
Algorithms absolutely choose what gets shown to us.
But people also choose what to click.
What to share.
What to rage at.
What to watch twice.
So the machine is learning from humans while humans are learning from the machine.
That loop is much more interesting than "evil algorithm."
And we now have unusually strong evidence that a feed can actually change some political opinions.
In a 2023 experiment published in Nature in 2026, researchers randomly assigned active U.S. X users to an algorithmic feed or a chronological one for seven weeks.
Turning the algorithmic feed on increased engagement and shifted some political opinions in a more conservative direction in that specific group and time period.
But here is the part that stops this from becoming a neat political story:
The researchers did not find a significant change in how much participants disliked the other political side, or in whether they identified more strongly as Republican or Democrat.
So the feed changed some opinions.
It did not simply flip a "become polarized" switch.
And the researchers warn that the results come from active U.S. X users in 2023. They are not a magic formula explaining every platform, every person or every election.
The Scariest Part May Be Simpler
Your feed does not need to contain fake information to give you a distorted picture of reality.
Selection alone can do that.
A security camera pointed at a dangerous intersection can show you real crashes all day.
That does not mean every road in America looks like that intersection.
Social media can work the same way.
The post is real.
The outrage is real.
The tragedy is real.
But the feeling that this is what life looks like everywhere, all the time?
That can be created by what gets selected.
And once you realize that, doomscrolling starts to look different.
Maybe you are not watching the world fall apart in real time.
Maybe you are watching a machine assemble the world's most attention-grabbing moments into one screen — and your brain is trying to treat that screen like reality.
One Question to Leave With
If your feed showed you an average day instead of the most shareable moments, would you still believe the world was becoming this angry?