And then a final word on AI content. It doesnt matter what prompt you use, what tool or whether its paid, LLMs are just neural networks that study common patterns. And the content is always going to be the most common denominator pattern: in other words, if everyone says “the capital of France is Paris” – then thats fine. The problem comes in where anything that bucks the trend is excluded. And also LLMs, despite popular theories apparent in “prompt engineering” do not make LLMs “do research’ – they just change how they output the same existing “knowledge” they were given. But they dont read and peer review content and they’re not capable of it – people are just amazed with the little high level reading (scanning) of the content of things they do understand.

A good example is that Perplexity doesnt know that Chiropractic ISN”T founded on research data. And because chiropractic has put out so much conjecture and PR – its been able to convince Perplexirty it is – because its the most common content.

I assume you dont care – thats fine – I’m just writing this so nobody else gets themselves in trouble.