Have been watching A LOT of content on "AI" and LLM. Developer content, Computer Science PhDs, AI enthusiasts, CEOs, those that think their life is over because "AI" will take over their jobs.

We use LLM all the time but we make sure we are in control of the final product quality.

The first thing anyone should do is to understand how these models work. Once we do, we realise that they can never be right all the time, there is no true knowledge, and they are constrained by the number of parameters that can be added before they no longer work. Indeed, there is a theory whereby the information needed to solve a problem will exceed the information needed to solve the original problem.

There are countless studies that test models, give problems to them. Please read some.

Now, LLM is transformative, but please understand it is a tool that will get things wrong and you need to be able to recognise these mistakes and correct them.

Am totally for non-experts using tools to create great products but do not get carried away with the hype.

https://youtu.be/ShusuVq32hc