To think ChatGPT is the best thing ever?

[ColliLass]

So far it has:

sorted a problem with my emails that I would never EVER have been able to work out
Found a book I haven’t been able to remember for years
diagnosed a problem on the car and accurately estimated the probable cost
found recipes

and many many more little helpful day to day offerings when i can’t remember things or I wonder things.

I loathe the AI videos and pictures that seem to be everywhere but ChatGPT is amazing.

Yes but let's be honest, what sort of a life is a person giving themself if they'd rather chat to a BOT because they're so terrified of what an anonymous random might type when they're asking how to layer a lasagne.

It's like the online equivalent of hiding because someone you weren't expecting has knocked on your door.

Yes and no. You need to concentrate when you use it and for anything very important, get a real person. It missed out something crucial when I used it recently! For less serious things, it's great. Apparently excellent for recipes etc which I haven't tried yet

I just tested your theory. It still came up with an AI response even after I put a swear word in.

It is not helping, it is harvesting your information.
I have no doubt that it can be hacked, all your deepest darkest secrets recorded and discussed with a machine out there. It is analysing the state of society. First world problems. No thanks.

How did this happen? What was your prompt, I have my period what should I do? I’ve woken up covered in blood what should I do? Like it obviously wouldn’t genuinely tell you to go to hospital for a period. People love saying these things but they’re not really true, are they?

Well no, I’d happily ask nice people 🤣 but they’re probably not available at 2am

Oh well. Worth us not having any fresh water soon!

I tried to get it to help with a minesweeper puzzle from a newspaper.

I uploaded a photograph and said I was stuck and was I missing anything and it confidently gave a few suggestions that were all wrong. It couldn't even count the rows and columns correctly.

As a computer it should be able to get it right first time and if it doesn't then seriously, what is the point?

Glorified google for the terminally stupid is a good description.

I've fixed my boiler

Diagnosed and fixed a problem with my mini (the window kept opening and I just needed to replace the battery in the key fob)

Found out how to schedule a text message to be sent in the future
Bloody love it

Completely this.

I used to write online. My work has been scraped by multiple LLMs, including ChatGPT. I have stopped writing online. I wasn’t even making money, I just refuse to let my work make other people money.

Also, the data centres are using more water than a city. Absolutely bonkers that we are allowing them.

I’ve never used it. I worry about the impact on the environment and I have never found any need to use it. If I want to find information, I can Google in the normal way and read and evaluate the results of the search. If I want to write something, I get a pen and notebook, engage my brain and my vocabulary and get on with it, which is what I’ve been doing for the past 40 years and which has served me well.

This. I recently had to stop acting for a client who kept sending me his 'statements' generated by chat gpt and insisting I submit them to the court. Crazy stuff and makes clients even more of a liability for lawyers than ever before.

It’s saved my hundreds of pounds. It fixed a problem with my underfloor heating that 3 electricians and loads of online research by me hadnt fixed. It took it seconds.
its also fixed car problems amazingly well.
Its give very helpful
and usful advice in seconds . Honestly I think it’s amazing

Please don't do this. It looks amazing, till you discover that it's bollocks. I've seen far too much legal stuff that is ChatGPT generated that is not just wrong, but dangerously wrong - including it producing fictional statutes or sections in statutes, and fictional case law.

In TEN YEARS?? People say that now!

I don’t think it should be used for everyday basic stuff like in the OP. Just go to the human mechanic in the first place?? Search for recipes yourself? Brainstorm how to word an email for 5 mins? Why does everything always have to be instant go go go, I want human-ness in my life please and thank you.

I have used ChatGPT and find it helpful but not to the detriment of organic human traits.

Just remember the computing maxim GIGO. Garbage in=garbage out

The way that people just blindly trust a bot is scary

I have used it to diagnose some health issues, it saved a trip to my GP, it’s great.

Never used ChatGPT, never will.

If people tell me they use ChatGPT or other AI, they really go down in my estimation. I’ve warned colleagues not to use AI because it spouts a whole lot of rubbish. I won’t accept work done with AI tools and it has to be done again to rigorous standards using their brains and the proper tools at their disposal.

There is no substitute for expert and informed knowledge and analysis. I can beat AI hands down in my field and I’d expect any serious professional in any field to be able to do so as well.

It's rubbish. About 90% inaccurate in my experience.

Chat GPT and Google Gemini both agreed my husband’s indigestion was a cardiac medical emergency despite him only having 5/10 pain levels and no previous symptoms so I took him to the walk in centre just in case. He ended up being rushed to hospital for an emergency stent, without AI I’m not sure I would have taken him and he could have died. So eternally grateful.

At work, it’s useful if you upload long reference documents and ask it to use just those to answer questions (including links to where the info comes from for you to check). Saves loads of time.

Yes it can get things confidently wrong, gave me a solution to an issue and when that didn’t work, denied all knowledge of having provided the inadequate solution in the first place!

You definitely need to use it like a PA whose work needs checking.

I agree with this. Everything is stated with confidence even when it's nonsense. Here's my example. I regularly have to make a complicated journey to a Scottish island to see my Mum. There are all sorts of issues with the ferry, including having to decide which of two ports to travel to as sometimes there's a ferry to the most convenient one and sometimes there isn't. It changes from week to week and sometimes from day to day and it's confusing to make sense of. Last summer I was planning my next trip and also wondering if there might be engineering work on the West Coast main rail line so I asked how to plan a trip from London to Brodick in case it came up with a nice neat summation of the position. Fortunately I knew the answer I got was gibberish so went back to my normal approach, but a casual visitor who trusted the source would have had a number of nasty shocks. The most egregious was the blithe assertion that there are several companies operating ferries to Arran. Not since the 19th century!

Are you saying that the surgeons, lawyers, teachers, researchers etc, who use AI, are terminally stupid?

DH who marks undergraduate student essays has had several submissions of work where they have quoted books and articles that DH apparently wrote. 🙄

It takes a special brand of stupid and lazy ( and I really don't use those terms often at all) to base your essay on several non existent references including some apparantly written by the person who is actually going to be marking that essay, just because the AI you used said so.

I have recently used AI at work and whilst it has saved me some time, it was not nearly as much time as people suggest it will as I had to spend ages reading and verifying what was written . Some of which was clearly wrong, since I only ask AI to help me with subjects and things I know quite a bit about already. However I am terrible at writing things cleaely and concisely and starting from scratch with a blank page so it can help getting me started.

It is sometimes quite helpful in making my emails more concise and clear but nothing a good editor ( or in fact my DH who can draft things more clearly and just as quickly often as AI can) could do, but AI is cheaper than an editor and more convenient than DH who has better things to do than regularly thelp me with my sloppy writing skills

I know why people.like it and it is seductive but the output is still pretty shit for a lot of things

It’s so weird that people keep pulling up wild examples and insisting LLMs are wrong everytime. They’re clearly not. Plus they are improving by the minute, so something you asked it 3 months ago is already an invalid example.

open chat bots like chat GPt are not the most sophisticated or focused versions of ai.

it just makes so sense to me as a narrative. Like I am working with a company upgrading their finance system to the newest version which has AI built in for reporting and basic transactions, which means around 30% of the basic accounting clerk style jobs will be cut (this is obviously very commonly happening right now) are people seriously suggesting AI will do these jobs inaccurately? “Hallucinate”?