Raw LLM Responses

Please don't fall for the hype around LLMs! LLMs don't strategize. There is no self-preservation involved. Its all a mirage and a result of biased training data. Feed a model a good load of training material where self-preservation is in causal relation to described actions. A lot of social interactions evolve around self-preservation. We DO know how LLMs work. What we DONT know is the internal state of the trained LLM and what decision trees look like for each prompt. The state is the black box, not the algorithms and the basic technology. When Geoffrey Hint is talking about Artificial Super Intelligence (ASI), he is still talking about some hypthetical technology, about science fiction. LLMs are not it. Recently he is talking loads of BS, like Musk and Altman. Musk and Altman want to make us believe that the technology will lead to ASI, because they profit off that hype and mythology. ASI is like fusion: Since twenty years, it will happen soon.