I've found that if you talk to one of these generative AI models (e.g. ChatGPT) and go into a technical topic that you're pretty knowledgeable about, you'll soon realise just how wrong AI is - it can very easily hallucinate completely incorrect information and be very adamant that it is correct.
For example I was asking it about networking protocols and it was getting simple terms completely mixed up and giving me explanations that made zero logical sense.
However, one thing I do like it for is helping me "sound human" - I sometimes use it to draft emails or even script a phone conversation if I need to ring somewhere to book an appointment or whatever. I actually get a little upset knowing that I'm asking a computer to help me interact with humans, but it is something I appreciate and it has helped me on occasions.
It's also great for when you have to Google something super specific - you know you won't find a result if you actually type it into Google, but you can do a long-winded explanation to ChatGPT and it will output something relevant.
Overall, for personal use I do like it - but I don't trust it for anything other than helping me communicate with others, and even then I just use it as a bit of extra guidance. Essentially, I treat it like asking a friend for help - I keep in mind my friend could be completely wrong about something.
In terms of its wider context within the world - it is both awe-inspiring and slightly scary. But... meh, we'll see.
That's helpful for me to know. Thank you. You have a measured view based on experience. Because I have no experience with it, I seek out "measured views" to help me understand what it is.....in reality.
That's helpful for me to know. Thank you. You have a measured view based on experience. Because I have no experience with it, I seek out "measured views" to help me understand what it is.....in reality.
