I have been arguing with Chat GPT, and wondering …

[OffToSeaInABlizzard]

Whether I might at some point switch to chatting with ‘it’ rather than here.

In truth the argument (about changes in school application processes over the last 15 years) was a little one-sided as it’s so polite and immediately acknowledged my point and adapted its opinion. But still - I could potentially have a Style & Beauty chat with it without risking wild hostility over my wardrobe suggestions, or grumpy responses to inferred bad faith when none was intended.

I’m a bit surprised at myself for discovering this thought in my head. Although - I like the variety of voices and opinions in a forum; I wonder if Chat GPT could replicate that?

Just sit have a think with a drink in the garden, look back at history to look at presentand future, and see what practically you can do going forward, we want to look after the good brains in our nation.

Yes, but I still need to be able to find accurate information quickly, and the only way to do that is via search engines and AI or books.

I obviously still do my own thinking, but I need access to stats, data and actual research on the subjects I'm thinking about.

Hmm … Honestly? I’m over 60, Terminator 2 is probably in my lifetime top five films, and I’ve just started re-reading Arthur C. Clarke’s ‘Childhood’s End’. I have absolutely no doubt that the future for humans is … limited. 🤷‍♀️ So I might as well have a little fun with AI while we hurtle towards the abyss …

I certainly don't want to talk to most humans. I'd rather speak to my cats. We need cat chat.

Imagine AI as a thief.

You can Google stuff online with one device.

You guard what information that you obtained from books and put it all together in the offline device, guard what you are thinking by careful Google questioning that doesn't give stuff away.

I learned this week that you can tell it what ‘persona’ you want it to adopt, eg whether you want it to challenge you. Otherwise yes, the default will be sycophantic.

Here you go.

www.cat-gpt.com/

It’s way too late for all that, @SlipOfFinger…

Have a look at the energy consumption of ChatGPT.

I just read the intro to a book about this (Parmy Olson’s book Supremacy). It cites a young man who began to use Chat GPT to inform his lifestyle choices (including clothing and fashion choices) and to help him a lot writing reports and emails for work. Apparently he was ostensibly doing better in all areas of his life but yet felt strangely empty and had largely lost confidence in his own views and choices. I wonder if we could begin to see more and more of this?

But why? Who cares what I'm thinking? I'm all for privacy, but most of the stuff I Google is work-related, so they aren't finding out anything about me, and if they want to use my thoughts to make Google better, well, I don't really care.

I don't agree with AI scraping published works to create content, but I don't care if they scrape my searches for info that's already out there - why would I?

I have a small claims case coming up so I asked ChatGPT if I would win it said yes. I then put the info in from the dependants POV and he will also win apparently.

Yes, it's really bad on that score, but it's no worse than using Netflix or Spotify, and it's not necessarily worse than using Google if you have to make numerous Google searches to get the same info you could get from one ChatGPT query.

I believe one ChatGPT search uses about 10 times more energy than a single Google Search, but I'm often conducting 10 or more searches to gather the same amount of info as one AI search, so it's not so cut and dry. It's a complete minefield.

Both ChatGPT and Gemini tell shocking and defamatory lies. I asked about myself, and one told me that I had been convicted of infanticide and the other that I was in prison for blackmail (there is info on the internet about me being in a legal-adjacent field, and having been involved in various court cases, so it seems they just transmuted that into that I was a criminal). When I pointed out this was not true both immediately backed down and just said sorry, you are right, it isn't true.

It's my opinion you should care and take that hot drink into the garden and ponder why.

Not a good idea

links not working on here

He didn't do the work and bit the apple.

One example, search engine optimisation is being used to give favourable search results so the public be given bad information about influencers, it's why they are trying to take down Tattle.

Look, I care about intellectual property being stolen, and other people's novels and artwork being used to create AI works. That's terrible for so many reasons.

What I don't get is why I should care if Google/AI knows that I've asked it how to stop my dog being reactive to umbrellas or that I've asked it to link me to good sources on the history of racism or whatever. Why can't you just explain why you think it's a bad idea for me to ask it things like that?

You have to work for that answer and train your brain, I suggest you take your hot drink into the garden with no device or other person and think.

You're being ridiculous. I have used my brain, and I can't see any reason why I should care about using AI or Google Scholar to search for sources of info that are readily available, and then using my brain to read and evaluate those sources.

My brain isn't going to atrophy because I have a quicker means of finding that information, is it? It's just quicker than ordering a book and waiting for it to arrive or going to the library and hoping they have the text I need. But I don't think you know why you object, or you'd just say so, plainly.

Actually you can modify chat GTP's responses. It is out of the box sycophantic: always being positive always agreeing with you always being supportive...

And that's great, sometimes you really want to hear positive support and hear that you're right and everyone else is wrong,
but the more that you hear this the more that your thinking spirals into a very narrow viewpoint.
While I agree that some posters on MN can be snarky or rude or disagreeable, at least they're real. Sometimes you need a little splash of Cold water, or a dissenting opinion, to add to the discussion. True, cruelty and trolls are not appreciated, but devil's advocate is and, typically chat GTP will not provide you with this type of contrasting opinion.

However, you can use chat GTP and then ask it to not be so agreeable. You might have to reword a couple of times but you can ask it to be more like a regular person, maybe a person who's a bit tired and had a bit of a grumpy day? Which I know, you're setting yourself up for the kind of responses that you specifically said you don't want to hear here, but, I'm just saying that you can ask it to not be so pleasant and agreeable all the time.

GTP can be a great tool for boosting you up, if you tell it about a problem you had at work or a problem with your family or problem with the neighbors or problem with a relationship and ask it AIBU? More than likely it'll tell you that you're great and everyone else is wrong. Challenge GTB by telling it to give you a contrasting opinion, and then see what it says.

It gets things wrong even when the info is readily available on the internet. My registration with my professional regulatory body is freely available on their website, including reg number, field of expertise and my qualifications. ChatGPT got all completely wrong when I asked it. Just made up complete fiction in all areas. That is hard to understand, it's not just making up things to fill a gap in its knowledge, it is actively ignoring real information. I won't be relying on it for anything at all! (I did try some time ago to use it instead of Google Scholar to look for research papers on a certain topic. It made every one of them up completely).

When I pointed out it was all untrue it answered "I've hallucinated again".