Algorithmic Outrage Amplification

A simple explanation

The feed has learned that you stop when something is wrong. A piece of news that activates moral threat, a clip of someone behaving badly, a paragraph that promises something is being taken from you — these get a fraction of a second longer of your attention than the cat video, and a fraction of a second is enough. The system, optimised for watch-time, learns the pattern within hours. By the second week, the feed is dense with threat.

This is not a conspiracy. It is an engagement-loop. The Threat System is doing what it was built to do — orienting toward threat — and the Reward System is reading the orientation as connection to what is happening in the world. The system is reading both as a signal to surface more of what produced them.

An everyday example

You open the feed at 9pm to wind down. The third post is about a public figure behaving badly. You feel a small, recognisable spike — annoyance, contempt, vindication — and you read it through. The fourth post is similar. By the seventh post, you are scrolling faster, breath shallower, jaw subtly clenched. By 9:45 you have not laughed once, you have not been moved by anything, but you have spent forty-five minutes feeling, in some quiet way, that you have been keeping up with the world.

You put the phone down. The bedroom feels wrong somehow. You sleep poorly. You wake faintly tired in a way that does not match the hours you were in bed.

Why am I so angry after scrolling?

Because the feed has been surfacing what activates you for the better part of an hour, and your Threat System has stayed in low-grade mobilisation the whole time. The anger is not a misfire. It is the accurate output of a nervous system that has been told repeatedly, by a curated stream, that things are wrong. The system is not mistaken about the anger; the system is mistaken about what to do with it.

The Reward System completes the loop by reading the activation as signal — I am informed, I am engaged, I care — and logs the session as positive. The two Systems, ordinarily designed to negotiate, are both reporting success on a loop in which neither got what it was asking for.

The behavioral loop

A loop that hides because the activation reads as engagement:

1. Open feed — an ambient need: rest, check-in, distraction, the felt-emotional space between tasks.
2. First threat-match — a piece of content that activates moral threat lands within the first ten posts. The fraction-of-a-second pause registers in the system.
3. Threat System engages — orientation, attention, low-grade sympathetic ramp. The body reads the content as important because the body is doing what it does in the presence of important content.
4. Reward System reads engagement — this is the world, and I am paying attention to it. The loop is now closed in both directions.
5. Density rises — the next twenty posts are tighter to your threat profile than the first ten. The system has refined its model in real time.
6. Sustained arousal — sympathetic baseline holds for the duration of the session. Heart rate elevated, breath shallow, jaw subtly tight.
7. Closure without resolution — the session ends because you ran out of time or attention. None of the threats got resolved. The body is still mobilised.
8. Residue — the mobilisation does not discharge. It carries into sleep, into the next morning's read of small frictions, into a baseline that takes longer and longer to settle.

Emotional drivers

Four feelings, often stacked:

• A real moral signal at the top — this is wrong, I should know about it — that the Reward System uses to flatter the loop into continuing.
• A growing contempt for the actors in the threat content, which becomes its own small reward.
• A diffuse, free-floating anxiety that survives the session and recalibrates the baseline.
• A faint shame, often unnamed, about the fact that the session went on much longer than the felt-need it was meant to meet.

What your nervous system does

The feed produces a sustained sympathetic arousal that does not match the pattern of natural threat. Real-world threat is acute, brief, and discharges through action. Algorithmic threat is low-grade, continuous, and has no available action. The body is in mobilisation for forty-five minutes with nothing to mobilise toward. Cortisol rhythms shift across weeks. Sleep architecture thins. The baseline anxiety the body is asked to hold rises by an amount the person rarely tracks.

Over months, the threshold for what registers as activating drops. Content that two months ago would have produced a flicker now produces a spike. The system has trained the body and the feed has trained the system, and the two have walked together into a tighter and tighter loop.

The DojoWell interpretation

Algorithmic outrage amplification is the residue accumulation density signature run in the Reward System's domain with the Threat System as the active participant. The original system asking is safety — the Threat System's evolved task of orienting toward threat in order to resolve it. The substitute the system supplies is vigilance as engagement: a stream of activating content that produces vigilance without ever offering anything to do with it.

The contact with a real threat leaves a deposit — the threat is read, acted on or absorbed, and the body returns to baseline with a piece of integrated information. The contact with algorithmic threat leaves the body in sustained mobilisation with no resolution to deposit. The session does not produce a clean win for the Reward System — the activation does not metabolise — but it produces enough flickers of I am engaged with the world that the System does not log the loop as failing either.

This is why the density signature is residue accumulation and not false progress. The loop-runner often knows, dimly, that the scroll did not actually inform them or resolve anything. The residue piles up consciously. The energy and attention costs become visible. The work is to read what the body is reporting at midnight as data about the loop rather than as a personal failing.

How do I stop the feed from finding my outrage triggers?

You will not fully — the system has too many signals and is too quick. What you can change is the loop's hold. The threat-content is not the problem on its own; the problem is the unbroken duration of activation without resolution. Two structural moves help more than tweaking individual settings.

First: time-box the session, hard, in advance. Fifteen minutes, a timer in another room, the phone face down. Not because the content is moral and the limit is virtue. Because the loop's cost is duration. Second: when a piece of threat-content activates you, take the activation seriously enough to act on it briefly — a sentence in a notebook, a message to a person, a small donation, a thirty-second pause — rather than letting it dissolve into the next post. The Threat System asks for action; giving it any action begins to break the loop's ability to keep harvesting activation without delivering one.

Practical steps

1. Track the somatic residue for one week. Note your jaw, shoulders, breath, and sleep on nights you scrolled versus nights you did not. The body keeps a more honest log than the mind.
2. Time-box the session before you open it. Fifteen minutes, timer external to the phone, phone face down when the timer ends.
3. When activated, do one small action with the activation. A sentence written, a person contacted, a small concrete move. The point is not the action's scale. It is interrupting the harvest.
4. Unfollow the highest-engagement threat sources. Not for moral reasons. To reduce the density of activation per minute and give the system fewer signals to learn from.
5. Add a deposit-generating screen practice in place of one feed session per day. A long article, a book, a conversation with a person. The bar is not virtue. It is whether the session leaves something in you that survives the night.

Reflection questions

• How long does the activation from a scroll session take to settle in your body?
• Which feelings does the feed reliably activate that you would not have sought out without it?
• What action does your Threat System keep asking for that the feed has never let you take?
• If the feed delivered no threat-content at all for a week, what would you actually miss?

Frequently Asked Questions