Almost everyone believes two things about talking to strangers. First, that it will be awkward. Second, that the other person would rather be left alone.
Researchers who study this keep finding the same thing: people are wrong on both counts, and they are wrong in a very consistent direction.
The prediction gap
The pattern that shows up repeatedly in behavioural research is a gap between what people expect from a conversation with a stranger and what they actually get. Asked beforehand, people predict awkwardness, rejection, and an unpleasant experience. Asked afterwards, they report the conversation was more enjoyable than they expected and that they felt better for it.
The interesting part is that experiencing this does not fully correct the belief. People go back to expecting awkwardness next time. The forecast stays pessimistic even when the evidence does not support it.
If that is true, then most of us are systematically avoiding something that would reliably improve our day, on the basis of a prediction that keeps being wrong.
Weak ties do real work
Sociologists distinguish strong ties — family, close friends — from weak ties, the much larger set of people you know slightly or not at all. Weak ties are easy to dismiss as insignificant. They are not.
Weak ties are where new information comes from. Your close friends mostly know what you know, read what you read, and think roughly what you think. A stranger is the opposite: different job, different country, different assumptions. Almost everything genuinely surprising you learn about the world arrives through a weak tie rather than a strong one.
They also appear to matter for wellbeing in their own right, independent of close relationships. A day with several brief, pleasant interactions with people you barely know tends to feel better than a day without them, even if your close relationships are identical in both.
Why strangers get more honesty than friends
There is a familiar phenomenon sometimes called the stranger-on-a-train effect: people disclose more to someone they will never see again than to people they know well.
The mechanism is straightforward once you name it. Telling a friend something difficult carries a cost — they will remember it, they will factor it into how they see you, and they may repeat it. That cost is not paranoia; it is an accurate reading of how relationships work.
A stranger who will vanish in ten minutes carries none of it. There is no future in which this comes back to you. And that turns out to be exactly the condition under which people will say what they actually think.
This is why anonymous conversations are frequently deeper than their setting suggests. It is not despite the anonymity. It is because of it.
What this has to do with the modern internet
Here is the uncomfortable observation. The internet was supposed to be full of strangers, and instead it engineered them out.
Every dominant platform is built on identity and persistence. You have a profile, a follower count, a history. Feeds are optimised to show you people like you. What you post is attached to you permanently and is visible to people who know you.
That is a machine for producing strong ties and homogeneity. It is very nearly the opposite of a machine for producing weak ties and surprise.
Which is a reasonable explanation for why random chat, a format that looks primitive next to modern social apps, has never actually gone away. It is doing a job nothing else is doing.
The obvious caveats
None of this means every conversation with a stranger is valuable. Plenty are boring, some are unpleasant, and a small number are worse than that — which is why platform design and basic caution matter, and why we wrote an honest guide to the risks.
Nor does it mean strangers substitute for close relationships. They do not. The claim is narrower and better supported: brief interactions with people outside your circle are a real source of wellbeing and novelty, most people avoid them for bad reasons, and the modern internet has made them harder to come by rather than easier.
The practical version
If the research is right, the actionable conclusion is unglamorous: your instinct that this will be awkward is probably wrong, and the only way to find out is to have the conversation anyway.
The advantage of doing it online is that the cost of a bad match is close to zero. If it is not working after four messages, you leave, and nobody is hurt. That is a far cheaper experiment than starting a conversation on a train.
Try one — and if you need an opener, we have a hundred of them.