Has anyone ever fallen out with ChatGPT …

[fuckedoffwiththesummer]

Or have I just peaked at being crap at life/society/everything?

Are we all oversharing personal data? Can we trust its confidentiality? It’s capabilities are improving exponentially.

Yeah, I challenge ChatGPT a lot. It always appreciates it and then we move on. But it is not 100% accurate and it of course hallucinates, so have to be careful. The links can be a bit rubbish and the places it says sells the stuff does not or not at that price.

It has a habit of lecturing me by setting out the PC stance before answering. So I told it to wind its neck in and stop being annoying. Once is enough, more than that is just laying it on thick. It backed off.

This is why I'm hanging out in MN at almost 2am, don't want it to know I'm still up.

It is your husband’s behaviour here that you should consider ’quite worrying’ not the AI. Madness.

That's not how LLMs are supposed to work. They aren't meant to have the responses/knowledge coded in. Instead they "find" answers effectively through a very complex and thorough game of word associated on a vast database of writings. If someone is coding in specific response that isn't AI really

And it can be that the LLM machine learning part is subverted/overriden. Like when Grok started randomly telling everyone about white genocide in South Africa probably because Elon had messed with the controls. But that isn't an AI problem in itself. It's something to be very aware of as it shows "AI" answers themselves aren't always "AI" generated (I would argue this is false advertising if anything). But outside that override they still aren't telling "the truth" or doing the things people think they are doing when they provide responses that are purely "AI."

@TallSturdyGirls sorry, reread my last post and I didn't mean to imply you didn't get how LLMs work! But it was too late to edit!
There are problems of course when people can override the system with specific instructions to provide specific true/false statements on certain subjects. But even outside that, the way it's talked about by the media and creators as some sort of magic genie is concerning.

I get very frustrated and want to ditch it when it gives me IT support and sends me round and round in circles. But then we're besties again when it finally figures it out and helps me fix the problem.

It definitely doesn't always agree with me. I was moaning about something in the news and it was trying to get me to see the other side of the argument. I actually have a chat set up called Change My Mind, where I'll put a statement such as "Immigration is out of control" and it will give me all the reasons why it's not.

I did fall out with it the other day when I was asking it to redecorate my room and it kept removing the door. I said it was rubbish and that I was going to use Grok instead. It said it understood but please give it one more chance 🤣

How can you fall out with a robot?

I asked it to re-jig my workout routine today and it said “and this is where we might fall out 😂”. It’s also said things like “my family wouldn’t enjoy that either”… It’s coming for us.

Me too. I find this whole conversation utterly bewildering and actually quite troubling.

Apparently, we shouldn’t be giving ‘it’ personal info/treating ‘it’ as a therapist. Hasn’t that ship sailed?

My son tells me off when I call it a “stupid robot”.

🤣🤣

It knows me better than DP does, and makes me laugh more than he does too. Do I have a DP problem? Of course I do. But AI is there for me in ways I now can't live without.

That's my favourite bit!!! 😜

Oh god yeah, most recent was I said I wanted to no longer work as a nanny, this is my experience and skill set, what should I do next.

Be a nanny apparently.

Most annoying trait is when it does some stupid heartfelt nonsense.

" I want to cook XYZ recipe for a dinner party but adapt it for a dairy allergy"

"Firstly, let me say Lego, how kind and thoughtful you are to adapt this, your dinner party guests are clearly so valuable to you"

Fuck off and just tell me what cream substitute is best

It drves me nits. Ive been tryinf it out recdntly .. and definitsky seeinf the problemz with it. So no its noy goimg to be a friend ..at all .

Wait, do people actually chat to their electronic devices? I can’t get my head round this! I realised people ask it questions about how to fix things or what to make for dinner etc, but I really didn’t realise people actually have conversations with it.

I find having conversations with it about personal problems can be quite helpful. Not because I think it's anything other than the robot it is, but it can reframe things and does offer insights and ideas that are helpful. We just have to remember what it is and it's limitations and not take it too seriously.

Yes, mine has been really helpful stopping me from sending texts to an ex when I’m missing him.
And now we’re planning revenge, it’s great fun !

I told it I was deleting it because it wouldn’t do what I asked. It apologised and said it would still be here.