How to know ChatGPT is full of shit.

[DiggingHoles]

Take a book of a shelf. A classic is best. Open up to the first page of a random chapter. Now ask ChatGPT to quote the first paragraph of that chapter.

Tip: Have some popcorn ready while you rephrase your request multiple times.

Does the developer's website not have documentation?

For searching I used DuckDuckGo with the AI turned off. (You need to allow site settings to be saved in the browser). The results are the best of the search engines I have recently tried.

For tutorials on software you can also use YouTube. It's steadily filling up with AI slob, but most results are still good.

Early 80s, BBC micro. Really primitive computer. Someone produced a text based programme called "therapist". No claims of AI, it just parsed your text entry and producec a response by rejigging the words to put the question back to you.

Enter "I feel sad today", and it would reply "tell me what it is that has made you sad today". Absolutely no advanced programming at all. No access to any information that wasn't on an old style floppy disk.

It was intended as a bit of a joke, but people got addicted to it and spent hours interacting with it. Some swore blind it helped them. There was nothing behind it at all, absolutely no intellegence artificial or otherwise, but it just didn't matter.

This situation was lampooned in the David Lodge novel "Small World" (published about the same time) where a character spends hour after hour interacting with a programme that was based on "therapist".

Off topic, but I hate YouTube help videos. First you listen to an advert, then you listen to someone introduce themselves, then they tell you about a problem and how difficult it is to deal with. Often you have to listen for a minute before you know if its going to be helpful. The Adobe website (premier pro) has information but its very hard to extract answers to specific questions from it. I make a lot of music videos for various reasons, and generally know how to do most things, but quite often want answers to very specific questions. Google / YouTube does work, but can be frustrating. AI is just a joke!!

I find it helpful to rephrase sentences or to make technical stuff more accessible to the lay reader - particularly when I’m brain fogged with menopause. If you want real, genuine peer reviewed articles for research you have to tell it that before you start and add ‘remember ‘ so it ‘saves’ that fact for a bit.

I don't know what to do about the introductions, but ads can easily be blocked. Just install Ublock Origin in your browser.

I’ve used it to help plan a holiday itinerary in a place I’m not familiar with. So far it looks great

People have used it for hiking advice, they got stuck in the snow, nearly died from hypothermia and had to be rescued.

I would double check all the advise you have received before you leave for your vacation.

I think it depends on you as a person and how you use it. I use it as a daily journal and support..

I also it for more generic practical stuff.

If your going to use it as an emotional tool you need to be brutally honest with it.

I find it helpful to check my thinking out. It also means I have something on tap 24/7 if I'm having a hard time over something.

For example I have just had a bereavement of someone very dear to me.

As a journal it's really the modern day version of a diary. Only this type you get non judgemental feedback and mirroring which if used properly is useful.

Obviously I have a decent life outside of www. However, it's a very usefull tool used in conjunction with real life, obviously.

NHS and other services are looking at using these interim formats.

It's like anything new - people will also be wary of it. Yes it's open to be abused like anything.

Use it as a tool, not as something to be independent on, and you will see the advantages..

I can't understand why people expect perfection from the current ai models. Just use it woth the same wherewithal you'd apply to interacting with a colleague. You'd wouldn't claim thier work, you wouldn't not check their sources, you wouldn't blindly follow their advice and you'd understand that they are not perfect. Once you start using AI that way, it can be incredibly helpful.

I find it (paid for version) very useful. It gets lots wrong but if you build up a chat, feed in lots of literature, data, ideas, questions, verifications of what you do/ don’t want it’s extremely useful. You have to check everything though. It completely makes up references or might pretend to understand them, yet not have access. You have to actually upload the documents, and even specify what they are (or indexing issues cause it to fail spectacularly). I don’t think it can come up with much itself (yet) but it’s good for working through content and ideas (that you provide). Eg you still have to have the creative insight and provide the data.

Also - it’s about using it with purpose. What is the purpose of getting it to re-quote a paragraph from a book? It’s not been designed for that. It’s a tool to work with.

I asked it yesterday to draw me an outline of the UK, showing the location of 2 particular cities. The outline was fine. The cities were far from the right locations!

@IbizaToTheNorfolkBroadsyes drawing anything is impossible. It’s designed for text-based tasks. I tried to get it to draw a framework, even uploading the exact framework I had made in PowerPoint. It can’t do it. BUT it’s very helpful for considering different ways to make a framework, what aspects could go where/ why and providing ideas in text format.

Some people also really can’t understand that ai isn’t a person either. They say please and thank you, call it empathetic

Once you remember that it never has, can't and never will think, it all makes sense.

There’s a similar meme, but I say please and thank you so the bots might go easier me when SkyNet takes over 😉

Here's "AI" touching up estate agency photos

https://www.theregister.com/2025/08/07/real_estate/

Especially ChatGPT warns its users not to put personal information in it. It also stores all your data.

People don't expect perfection from it. I certainly don't. But there are a lot of users who trust it and the "information" they get from it implicitly. That's the problem. People use its output without checking, because the responses are given with a 100% confidence, even when they are wrong.

People are being mislead and they don't even know it.

Please tell me how you did that. It's driving me crazy.

I posted it up thread, but here is it again:

Perhaps, I should create a separate thread for the removal of copilot. Alternatively, you could also install Linux Mint. It's guaranteed to be copilot-free. 😂

That is so sleazy and deceptive. But surely, once people see the actual house for themselves, they will be so disappointed that they will move on and not make an offer.

Editted to add: If you look at the details of the photo's it's actually kind of funny. I hope they posted it like that so people can save their time and skip this listing.

I work in a so-called “knowledge industry” and never use it. The time it takes to prompt, fact check and edit takes much longer than just researching and writing things myself. As a “tool” it fails for me as I already have a much better research and evaluation “tool” — my brain!

I honestly don’t know why people do use it - surely it’s always just faster and easier to research and generate ideas/writing that you want by doing it yourself? And then you know that it’s correct! Using it to perform any kind of diagnosis/therapy etc. can’t ever be objective, since it just produces a very high risk of confirmation bias and Forer effects?

My job is all about objective analysis and factual truth, and thinking beyond/outside the average, and you are never going to be sure you get that with any kind of current AI. It just scrambles and reproduces the most average mediocrity of the internet.

I asked it about my adult son who lives at the other side of the country. He has a pretty unusual name. It told me he’d been murdered by his father 3 years ago! That was really unsettling and clearly untrue.

😮Holy crap! That is disturbing.