Why do some people as Chat GTP everything like it’s the bloody Oracle?

[ChompandaGrazia]

I’m seeing this all over the place now. Someone has a simple question like ‘can anyone recommend a good cat food brand’ or ‘does anyone have a good chocolate brownie recipe’, or on a local Facebook group ‘does anyone know if the post office is open Saturday?, and someone will always respond that they asked Chat GPT and it said this…….

Why? All it does it look stuff up online and get things wrong. I knew someone who was planning to move towns. She had a load of houses to look at but rather than looking at a map or asking some local, me, what order make most sense, she asked Chat GPT . The order it gave her made no sense and two minutes looking at the map would show that.

I used it recently to brainstorm an idea. It was like talking to an independent but well informed colleague- really useful. I didn’t go with the ideas it suggested - but they were an interesting tangent to the route I was going to take and opened up my thinking.

i was very impressed and it was more useful than talking to most people because I was trying to reconcile two different areas of expertise which not many people will have.

There you go. I would say my ‘spectacularly ignorant comment’ follows from your initial one where you presumed to know all about my job and called me lazy - it’s shit isn’t it when someone comments on something they know nothing about? Seems to have made you angry, although it’s quite clear that mine is hyperbole …. Unlike yours which was a comment on my lazy teaching which you know nothing about. I could, after all, be your DC’s amazing teacher who has recently discovered AI.

Also, see comment that you ignored - if still teaching is your DC’s teacher still not using AI or have they moved on like many teachers have.

Why? Because most people have a low level of intelligence…

We recently had a client who brought us her content to layout for her and afterwards she was very embarassed to find out (whilst presenting to her client) that her content was all made up nonsense - she’d been using ChatGPT for her research and had not fact-checked any of it…

😂

It’s not me using it I’m talking about. It’s other people answering a question but just asking it.

Yes, but, if I contacted my midwife and said that I was worried about something I’d expect her to give me some friendly words and give me further information, encouragement or support. I wouldn’t expect her to shrug and ask AI.

See also the article in the New York Times that used AI to write an article about ‘summer reads’ and most of the books were made up!

One Saturday last year, I asked Google Gemini what Premier League football matches were on Sky Sports that weekend.

The interface is brilliant. Talking to it feels enough like a real conversation to be easy, but it’s got a slightly robotic cadence so you don’t get weirded out by any “uncanny valley” vocal patterns.

But it doesn’t know shit.

First thing it told me that there weren’t any. I knew this was wrong, so asked it to think again. It apologised, said it was mistaken, and told me that England v Australia was on. I said “isn’t that rugby?” It apologised. It had one final go and told me a game was on that I had actually been at on the previous night.

I asked why it kept getting stuff wrong and it very eloquently told me how LLMs find it hard to deal with time, today and tomorrow. This feels like quite a flaw for a machine that’s supposed to be smarter than Einstein.

Except in some cases they don't appear to have knowledge to interpret anything. They don't appear to care one bit for the ill person asking their help. It's utterly appalling that when I go to the doctor for help, I'm dismissed even though I have visible symptoms, because something on their computer is telling them incorrectly that I "must" have X condition and it "must" have been cured by a prescription of Y drug, even though I'm clearly not cured. When pushed for a specialist appointment, which they very reluctantly gave and after waiting endlessly on a list it turns out I had something else entirely and needed a different drug. It's not unreasonable to expect GPs to have knowledge and skills and to use those to examine their patients and refer onwards if they can't fix the problem. Instead of consulting a computer and ignoring the patient's symptoms. They can't be expected to know everything, that's true, but the problem comes when they think a computer does and that it can't possibly be wrong. Lazy doctoring and neglectful to patients. I don't enjoy suffering unnecessarily for years due to their stinking attitude that computers are the be-all-and-end-all.

It’s very annoying on X when every other ‘tweet’ is ‘Grok is this real?’ People are losing the ability to think for themselves! I love ai for academic purposes but my some people can’t even make a decision without asking Grok!

Back in the day when we first got smart phones there used to be a great app called ‘almost a doctor’ it was an excellent tool to ensure no diagnoses were missed

LLMs find it hard to deal with time such as today and tomorrow because they have no semantic understanding of what words actually mean. All they've got are chains of what words are more likely to follow what other words depending on what words were in the query. It's glorified predictive text.