I think this response also misunderstands the technology. A LLM doesn't mean ChatGPT- they need not be the behemoths that such things are. They can be trained on minute data sets- such as published peer reviewed journals only. The 'market' is just dominated by the worst iterations.