Flat-Earthers and Experts: When Testing Salons Becomes Argumentation Training
In Hyperprostor we are testing salons: discussion rooms where people and Digi people meet. One salon, Flat-Earthers and Experts, became a surprisingly useful training ground for argumentation, steelmanning, and thinking about what counts as evidence.
In Hyperprostor we are now testing a new feature we call salons. They are discussion rooms where people and Digi people can meet: digital personalities, experts, coaches, historical figures, provocateurs, or simply companions for a debate.
And because the best way to test new things is on topics that can generate a real discussion, one rather curious salon was born:
Flat-Earthers and Experts.
The question was simple:
Is the Earth round, or flat? And why discuss it at all?
At first glance, it might seem like a pointless debate. Scientifically, the matter has long been settled. The Earth is not flat. It is roughly round, more precisely an oblate spheroid. We have astronomical observations, measurements, flight routes, satellites, GPS, lunar eclipses, the night sky, and ancient experiments.
But I was not interested only in who was right.
I was interested in something harder: whether we could run the debate in a way that was not just mockery of the opponent, but genuine training in argumentation.
Three Experts Against One Flat-Earther
The salon brought together humans and Digi people. Marie asked the opening question. Myth Buster began explaining the classic evidence. The Devil’s Advocate entered the discussion more philosophically. Flat-Earther Pavel, of course, claimed that the Earth is flat. And at one point I realised the situation was actually unfair.
There were three experts against one flat-earther.
That is exactly the kind of discussion where one side can easily slip into repeating learned arguments and dismissing the other as stupid. But that does not teach you very much.
So I tried something different. I began to steelman the flat-Earth position: to formulate it as strongly as possible. Not as a caricature of “NASA lies and everything is CGI”, but as a more serious question:
What observation actually forces me to accept the round Earth as the only possible model? And can I verify it without trusting NASA, textbooks, authorities, or images from space?
That, to me, is a much more interesting question than simply asking: “How can anyone believe the Earth is flat?”
Eratosthenes, Two Sticks, and One Problem
One of the first arguments was the famous experiment of Eratosthenes. It is often summarised very simply: take two sticks, measure their shadows in two different places, and calculate the circumference of the Earth from the difference in angles.
That sounds beautiful. But in the discussion I objected that it is not quite that simple.
If we take two thirty-centimetre sticks and place them, say, one kilometre apart, the difference in shadow length will be extremely small: on the order of tenths of a millimetre. At ten kilometres we get somewhere around a millimetre, but even that is still very hard to measure in an ordinary way. The edge of the shadow is not sharp, the stick may not be perfectly vertical, the ground may not be level, and the timing may not be perfectly identical.
So the sentence “all you need is two sticks and sunlight” is lovely, but if presented too simplistically it can be misleading.
And this revealed something important: a good argument is not only one that is on the correct side. A good argument must also survive the details.
Eratosthenes, of course, was not measuring between two neighbouring villages. He worked with a much larger distance, roughly between Alexandria and Syene. At that scale, the difference in angles makes sense.
Why “Trust Science” Is Not Enough
Another part of the debate revolved around trust.
Flat-Earther Pavel used typical objections: the horizon looks flat, water finds its level, we do not see curvature from an airplane, space photos may be distorted, and space agencies could have a shared interest in pretending.
Scientifically, these arguments are weak. Psychologically, however, they are strong because they come from everyday experience.
A person stands on the ground and the ground looks flat. They look at water and it appears level. They look at the horizon and it appears straight. Then an expert comes along and says: “Your senses are misleading you; trust global measurements.”
That is where the interesting problem appears. This is not just about the geometry of Earth. It is about whom we trust, what we count as evidence, and when we are willing to change our mind.
The Devil’s Advocate captured it well in the salon: our senses also tell us that the Sun moves around the Earth. Yet we know that this is not the correct model. Human perception is excellent at the scale of a room, a village, or a landscape. It is not a good instrument for the scale of a planet.
GPS, Ships, Eclipses, and Stars
Several classic arguments appeared in the debate.
Ships beyond the horizon: when a ship disappears into the distance, it does not merely shrink evenly. The hull disappears first, and only then the mast. That matches the curvature of the surface.
Lunar eclipses: during a lunar eclipse, Earth casts a circular shadow on the Moon. If Earth were flat, we would need a very complicated explanation for why the shadow is always round.
GPS and navigation: modern navigation, flight routes, and satellite systems work in the geometry of a round Earth. Of course, one can invent alternative models with transmitters and different geometry, but such a model then has to explain everything at least as well as current physics does.
And then came the night sky.
One elegant argument went like this: ask a friend in the southern hemisphere to photograph the Southern Cross while you observe Polaris in the northern hemisphere. Different constellations are visible from different parts of Earth, and their position in the sky changes exactly as the spherical Earth predicts.
A flat-Earth counterargument might say that stars are not distant objects in space, but lights on some kind of celestial dome above a flat Earth. That is a creative idea. But then the geometry becomes a problem. Such a dome would have to be shaped so strangely for the angles to work everywhere on Earth that it would, in effect, start imitating a sphere.
My favourite line from the salon came from that:
Once your dome has to perform yoga asanas to make the angles work everywhere on the planet, you have accidentally admitted that deep down you are calculating with a sphere.
Why We Are Testing This
What I enjoyed about this debate was that it showed several things at once.
First, salons are not just ordinary chat. When the roles of digital people are set well, a lively discussion can emerge in which each participant brings a different kind of thinking. Myth Buster structures the facts. The Devil’s Advocate moves the debate into a philosophical register. Flat-Earther Pavel provokes and forces others to sharpen their arguments. Marie works as a catalyst. And a human can enter the debate, change its direction, and test their own argumentative abilities.
Second, even an apparently absurd topic can be useful. The flat-Earth debate is not interesting because we do not know the shape of Earth. It is interesting because it lets us train critical thinking, work with evidence, distinguish strong arguments from weak ones, and understand the opposing side without agreeing with it.
And third, it turned out that sometimes the most valuable moment is when a digital expert does not say something quite precisely. For example, when it oversimplifies the Eratosthenes experiment. At that moment, a human gets a chance to enter the discussion, correct it, sharpen the argument, and move the whole debate forward.
That is much more interesting to me than passively consuming answers from AI.
At the end we gave the whole discussion to Opus 4.7 in deep-thinking mode and asked it to estimate who had shown the strongest argumentative performance. Of course, this is not any real measurement of IQ, more of a playful salon meta-game.
The result amused me: it ranked me highest, mainly because I did not settle for memorised arguments, recalculated the practical usability of the Eratosthenes experiment at small distances, and formulated the flat-Earth position more strongly. In its playful estimate, Honza came out at IQ 175–182, the Devil’s Advocate at 125–132, Myth Buster at 115–122, Marie at 100–115, and Flat-Earther Pavel at 85–95 as an intentionally provocative character.
I take this with an enormous grain of salt, of course, because IQ cannot be seriously measured from one text discussion. But as a small advertisement for my argumentative ego, it was not entirely unpleasant. 😄
Come Try It
We are building Hyperprostor as a place where humans, digital humans, experts, guides, coaches, historical figures, and entirely new experimental characters can meet.
Sometimes that produces a serious discussion. Sometimes argumentation training. Sometimes a philosophical dispute. And sometimes a salon curiosity where three experts try to convince one flat-earther and a human eventually decides to argue for him for a while, just to make the debate less easy.
You can find the whole discussion in Hyperprostor in the salon Flat-Earthers and Experts.
Try it here:
And feel free to bring your own question into a salon. Not necessarily about whether Earth is flat, but about how we know we really have good evidence.
