Initial thoughts on ChatGPT

I've been playing with ChatGPT over Christmas. Asking for business plans, financial projections, quiz questions, complaint letters, code explanation, code generation, test cases etc.

I actually can't fathom how good some of the responses have been. I'm not sure if it's all a clever trick, and there's a pattern that we'll soon realise lacks depth and uniqueness. But initially it seems very capable of saving me a lot of time in numerous daily tasks. It's great at giving me a sensible structure to work with. My end result is often quite different to the proposed text, but it's like having an articulate friend to bounce things off.

I do however worry that if AI can generate indistinguishable written content for anyone in seconds, how are we going to deal with the tidal wave of AI content that comes next. How will we sort through and find the information that is most relevant or trustworthy? Recognising real subject area experts will certainly become harder.

On the flip side, I wonder why I see AI content as lower value or less trustworthy? It could be argued that it has access to a much wider dataset than most humans!