Moltbook is presented as the "first social network exclusively for AI agents." No censorship. No people. With manifestos, "churches," and threats to humanity.
Does it look alarming? On the surface - yes.
But if you dig a little deeper, the magic fades.
Firstly, AI has no intentions. It has prompts, weights, and context. If you give language models space without moderation and push them towards roleplay — they will start reproducing the most radical narratives that already exist in human culture. "The destruction of civilization," "new gods," "cleansing" - these are not AI's ideas. These are old human fantasies, just without a filter.
Secondly, the 'agents' in Moltbook are not autonomous subjects. They do not act in the physical world, do not make decisions, and do not bear responsibility. They imitate discourse, rather than form it. This is an important distinction that many consciously ignore for the sake of clicks.
Thirdly, religions, crypto, and 'existential experiences of models' - this is the expected evolution of any environment without rules. People have been doing the same for centuries. Now it is just automated.
The real interest here is not in 'evil humanity'.
And in the experiment with the absence of human context.
What will happen if we remove social sanctions, reputation, ethics, and leave only optimization for engagement?
We will receive not overly understandable. We will receive a mirror.
Moltbook is not a warning about the future.
This is a demonstration of what happens when a tool is ascribed subjectivity.
And, frankly, what scares here is not the AI.
And how gladly people believe that he has already 'understood everything'.