Luther Without the Printing Press. On Change, Freedom & the Role of Technology
New text block
Ideas alone don't change the world. What changes it are the paths they open for people.
Picture Martin Luther finishing his Bible translation in a world with no printing press. The pages stay handwritten. Copied slowly, carried from town to town, read by a handful. Maybe there would still have been uproar, still the arguments. But almost certainly inside a few closed and controllable circles and no further. The fire of a rebellious theological episode, that lifted an age off its hinges would have died at the spark without any fuel to sustain the fire.
Now run it backwards. The press is ready, the type is cast, the machines are warm — and nobody dares put a question to the ruling order of knowledge. Infrastructure, but not one sentence that burns. No spark. No translation to shift the power of interpretation out of the hands of the authorities. Nobody telling the reader he need not wait to be told what is true. We would have had the technology, the fuel. Just not the spark of liberation.
New text block
The historical point sits precisely in that gap. Neither Luther nor the printing press changed the world on its own; it was the collision. A new reading of the text meeting a new machine for carrying it. One of them asked the question differently, the other made certain the question could never be caught and put back — and out of the two grew a third thing neither had planned.
A person could now stand in their own relation to the truth.
None of it went smoothly, and certainly not overnight. Violence, error and power struggles were the price paid and at the end new dogmas stood where old ones had fallen. History rarely runs clean. The door now standing open was of use to those who could read. The meaning of the word stopped hanging on the few.
Today we stand at a similar threshold, with the signs reversed. The question is no longer whether texts can be multiplied. That one is settled. It has moved. Who gets to produce knowledge at all? Large language models do not merely emit text. They order arguments, form hypotheses, dissect data, draft strategies, write programs, weigh studies against one another, simulate decisions.
They are not libraries preserving what exists. They are not printing presses in the old sense, multiplying what exists. They take part in the making of knowledge itself. For us humans, devided by ideologies this is the uncomfortable part. As long as knowledge came out of universities, newsrooms, laboratories and public institutions, you could at least roughly see where it had come from. Move it into closed model systems and the old question of power returns wearing new clothes. Once it was: who owns the language? Then: who owns the books? Then: who controls the media? Now, increasingly: who owns the systems that bring new answers into the world? Perhaps that is the real parallel to the Reformation. Not religion, not the church, not Luther as hero or heretic — but what happens to a society that hands its instruments of understanding to somebody else.
So think the scenario through to the end. Powerful AI systems exist, and only a few people grasp how they work. That is the printing press without the art of reading. The machine runs, the pages fly, the world shifts, and most people stand beside it seeing only what comes out at the far end. They can admire the spectacle. They cannot devour the content, fear it, argue with it, or throw it away.
The picture sharpens the moment these systems stop reproducing what exists and start producing what does not. Plenty of critics will tell you today's AI is a very accomplished piece of mathematical sleight of hand, and they are not entirely wrong. However it has to be conceded that in capable hands these systems already take a serious part in producing new knowledge. Undiscovered patterns, unseen solutions, fresh scientific conjectures, market strategies, political narratives: all of them can come out of asking well and feeding well. If your answer to that is simply access, I have to disappoint you. Access is no longer enough. What decides it is not whether you get an answer, but whether you can see how it was arrived at. Whether what you have read is probably true or merely sounds convincing. This is the foundation of the new fragmentation, and its first layer is lying in plain sight. Is that what you see novel insight in front of you, or is it polished language? This is more dangerous than plain inequality. It shows a new form of epistemic dependency. When a few actors decide which models get built, which data gets counted, which answers look plausible and which questions never get asked at all, what appears is not a classical knowledge monopoly with visible walls. It is something finer. Softer. Far harder to get hold of.
Today we stand at a similar threshold, with the signs reversed. The question is no longer whether texts can be multiplied. That one is settled. It has moved. Who gets to produce knowledge at all? Large language models do not merely emit text. They order arguments, form hypotheses, dissect data, draft strategies, write programs, weigh studies against one another, simulate decisions.
They are not libraries preserving what exists. They are not printing presses in the old sense, multiplying what exists. They take part in the making of knowledge itself. For us humans, devided by ideologies this is the uncomfortable part. As long as knowledge came out of universities, newsrooms, laboratories and public institutions, you could at least roughly see where it had come from. Move it into closed model systems and the old question of power returns wearing new clothes. Once it was: who owns the language? Then: who owns the books? Then: who controls the media? Now, increasingly: who owns the systems that bring new answers into the world? Perhaps that is the real parallel to the Reformation. Not religion, not the church, not Luther as hero or heretic — but what happens to a society that hands its instruments of understanding to somebody else.
So think the scenario through to the end. Powerful AI systems exist, and only a few people grasp how they work. That is the printing press without the art of reading. The machine runs, the pages fly, the world shifts, and most people stand beside it seeing only what comes out at the far end. They can admire the spectacle. They cannot devour the content, fear it, argue with it, or throw it away.
The picture sharpens the moment these systems stop reproducing what exists and start producing what does not. Plenty of critics will tell you today's AI is a very accomplished piece of mathematical sleight of hand, and they are not entirely wrong. However it has to be conceded that in capable hands these systems already take a serious part in producing new knowledge. Undiscovered patterns, unseen solutions, fresh scientific conjectures, market strategies, political narratives: all of them can come out of asking well and feeding well. If your answer to that is simply access, I have to disappoint you. Access is no longer enough. What decides it is not whether you get an answer, but whether you can see how it was arrived at. Whether what you have read is probably true or merely sounds convincing. This is the foundation of the new fragmentation, and its first layer is lying in plain sight. Is that what you see novel insight in front of you, or is it polished language? This is more dangerous than plain inequality. It shows a new form of epistemic dependency. When a few actors decide which models get built, which data gets counted, which answers look plausible and which questions never get asked at all, what appears is not a classical knowledge monopoly with visible walls. It is something finer. Softer. Far harder to get hold of.
A monopoly on probabilities, on interpretation,
on what sounds reasonable.
on what sounds reasonable.
And you might not even notice. Precisely because the answers arrive so friendly, so fluent, so helpful and so sure of themselves that no feeling of being shut out ever surfaces. You did get an answer, after all. But an answer is not participation. Reading was once more than a craft. It was a shift of power. Whoever could read no longer had to believe what was read aloud to them.
The corresponding art of our own time does not stop at prompting, and certainly not at a handful of tricks for handling chatbots. It consists of producing knowledge with AI, testing it, and contradicting it. That takes an understanding of how these systems actually work. It takes a creative streak to see what they are genuinely good for. And it takes a decision from every single person: whether you consume the product, whether the product consumes you, or whether you take part in it as a producer. Get that right and AI does not become the new pulpit to be preached from. It becomes the workshop where the work gets done.
Perhaps that is the real lesson of the old thought experiment. A society does not become free because a technology exists, and not because a few individuals have great ideas. Freedom appears only where idea, tool and capability find each other. Luther without the printing press would probably have faded out. The printing press without readers would have freed very little. A powerful AI without broad AI literacy may become a great many things — democratic is not one of them on its own.
The corresponding art of our own time does not stop at prompting, and certainly not at a handful of tricks for handling chatbots. It consists of producing knowledge with AI, testing it, and contradicting it. That takes an understanding of how these systems actually work. It takes a creative streak to see what they are genuinely good for. And it takes a decision from every single person: whether you consume the product, whether the product consumes you, or whether you take part in it as a producer. Get that right and AI does not become the new pulpit to be preached from. It becomes the workshop where the work gets done.
Perhaps that is the real lesson of the old thought experiment. A society does not become free because a technology exists, and not because a few individuals have great ideas. Freedom appears only where idea, tool and capability find each other. Luther without the printing press would probably have faded out. The printing press without readers would have freed very little. A powerful AI without broad AI literacy may become a great many things — democratic is not one of them on its own.
Whoever only reads what others produce stays in the audience.
Whoever understands how knowledge is made can argue back.
And sometimes that is exactly where freedom begins.
Whoever understands how knowledge is made can argue back.
And sometimes that is exactly where freedom begins.