Is ChatGPT right about our economy?

Hoenir

IanManc said:

You ignored the demonstrable facts that the UK is the 6th largest economy in the world, and the 9th biggest manufacturer in the world.

Productivity levels continue to fall though. Wealth isn't created by printing money nor selling assets to each other at ever increasing prices. Asia meanwhile is younger, hungrier and has a real opportunity to grow. With people more inclined to work harder. 

david72

IanManc said:

Once again you assumed that the UK stock market accurately reflects the UK economy. You ignored the demonstrable facts that the UK is the 6th largest economy in the world, and the 9th biggest manufacturer in the world.

There is, however, the aspect that many US companies are very large, very successful, and very well known around the world.

I'm sure that vox pop interviews with people in the street in the UK would result in most people being able to quite easily reel off a list of several dozen or more very successful US companies without very much effort at all. Ask them to do the same for successful UK companies, and I expect most people would probably struggle to name more than relatively few, and many of those that they would name would be companies that mostly or only operate in the UK itself.

It's quite telling that US business culture does seem to very strongly focus on growing markets, and beyond their own borders as well, and that when they have successful companies, they do a good job of getting media coverage for themselves. Too many British companies just seem to plod on, and don't promote their successes enough…

friolento

IvanOpinion said:

 

Try an old favourite by asking it to list all US states with an 'o' in their name. 

That's much better than my 5 letter word experience    

MX5huggy

I asked CGPT what ex premier league team finished lowest in all English leagues this season. It thinks Luton 22nd  in the Championship is lower than Oldham 5th in the National League because 5th is higher than 22nd. Not sure I trust it with economics. 

Martico

I've just read the visible part of the OP's "question" - it's as loaded as Mrs. Merton's line of questioning to Debbie McGee. Please don't get a job in research.

SneakySpectator

Martico said:

I've just read the visible part of the OP's "question" - it's about as loaded as Mrs. Merton's line of questioning to Debbie McGee. Please don't get a job in research.

It wasn't meant to be a loaded question, it was a continuation question from a series of previous questions where I was talking to it about ways the UK could increase economic growth and it started giving responses about lack of VC funding for new companies, older more established companies are deeply entrenched and how the UK fund raising scene is more cautious and more risk averse. 

Which then lead to discussions about US tech firms getting massive VC funding which allows them to develop and grow and if we don't get the same kind of investments how are our companies meant to grow and expand. 

Which then lead to discussing our lack of tech companies and high paying jobs etc etc. I didn't deliberately try to get a negative response. 

I asked it a more upbeat positive question with a follow up and here's what it said 

thegentleway

friolento said:

I asked CHATGPT for a 5 letter word ending in car.

It gave me boxcar.

I pointed out that this wasn’t a 5 letter word. CHATGPT apologised and gave me tocar. Which I don’t think is an English word but I concede, I had not asked for an English word. 

It never gave me vicar.

Go figure.

It can’t answer questions like that; it use forms (parts of words/small words) not the alphabet…

prowla

The "AI" systems can likely perform statistical analysis to infer rules, based on their input.

If their input is simply to trawl the internet, then they will inherit all of the misinformation.

If their input is guided then they will inherit the biases of the guider(s).

Some of my colleagues use ChatGPT as a technical resource and it can be helpful as an advanced search service, but oftentimes it can be not quite right and you have to scrutinise the answers to check for errors.

The helpful suggestions for descriptive texts can be quite awful and often adhere to the general rules of constructing sentences, but being totally devoid of meaning.

My worry is that people will start to take the AI output at face-value as fact, or even systems will be put in place which make real-world decisions based on it.

Example of where it has gone wrong:

• eBay: 1. their AI-generated product descriptions which tell you absolutely nothing whatsoever about the item for sale, and 2. their issue reporting capability which insists that counterfeit items are genuine.
• I lost the serial/password sticker off a Dyson product and did a google search "dyson lost sticker off base of fan"; google's "AI" responded with: "If you've lost the sticker with the serial number on the base of your Dyson fan, you can still find the serial number. For most Dyson air treatment machines, including fans, the serial number is located on the base of the machine, according to Dyson UK. You'll need to gently turn the machine over to locate it. If you're having trouble finding it or need further assistance, you can contact Dyson customer support.".

Example of where it's been helpful:

• A case at work where I wrote a description of a product architecture and told the tool I was using to draw a diagram; it did pretty much what I wanted, correctly labelled. (I think my description was quite precise and structured, so it fed in to it quite well.)

GingerTim

Hate AI with a passion for all the reasons rehearsed above, it being built on theft, and because I see it every day having a deleterious effect on students' critical thinking skills.

But a stopped clock and all that...

twopenny

The thing is basically a screen scraper tarted up .

Would you make investments on the judgement of AI?
Would you trust it with your money?