Hmm, is it valid to make an inverse conclusion from this in the following way: LLMs have problems with maths that requires multistep processes. Some humans are also bad at maths. In conclusion, these humans can be assumed to also have problems with or are lacking multistep processes?