I think LLMs have considerably more probability to make the world worse overall than cryptography does (the sheer level of information bullshit they can develop

I think LLMs have considerably more probability to make the world worse overall than cryptography does (the sheer level of information bullshit they can develop for pennies is going to transform our society and I doubt it is going to be for the better). Still I don't see the point of banning open weight models and LLMs that don't have guardrails. And I'm not sure you can realistically construct laws that would do it accurately. The genie is out of the bottle, pandora's box is opened, etc, etc. And locking down models with guardrails is only something that corporations have to do in order to avoid having a public racist chatbot problem and the associated headlines.