A question about Open Banking

Murmansk

I have a Starling Business account and use Quickbooks for my accounting. There is a connection between the two using Open Banking.

What puzzles me is the delay between a transaction becoming visible in my online banking and it making its way into Quickbooks via the Open Banking connection.

Even if I go into Quickbooks and click Refresh it's never up to date and often takes a day or two to reflect recent transactions.

Is there a deliberate delay built into the way Open Banking works?

A few minutes later

I pasted my text above into Google Gemini and got a very good answer. I'd paste it in here but I suspect it might be against the rules on here so try it for yourselves! It's a good example of how good AI can be - assuming it's correct!!!

GingerTim

Never assume that any AI answer is correct (safer to assume the opposite!) - it doesn't actually know things, just presents plausible text based on your prompt.

Murmansk

GingerTim said:

Never assume that any AI answer is correct (safer to assume the opposite!) - it doesn't actually know things, just presents plausible text based on your prompt.

OK but that text is ultimately obtained from a distillation of stuff on the internet rather than just being randomly created to sound plausible?

Nasqueron

Murmansk said:

GingerTim said:

Never assume that any AI answer is correct (safer to assume the opposite!) - it doesn't actually know things, just presents plausible text based on your prompt.

OK but that text is ultimately obtained from a distillation of stuff on the internet rather than just being randomly created to sound plausible?

I saw a post on Reddit today which suggests AI gets about 40% of their data from Reddit lol

The google AI summary when it links directly to a bank page is probably ok to skim and confirm directly on the source, if it's an open ended question, I wouldn't trust it

GingerTim

Murmansk said:

GingerTim said:

Never assume that any AI answer is correct (safer to assume the opposite!) - it doesn't actually know things, just presents plausible text based on your prompt.

OK but that text is ultimately obtained from a distillation of stuff on the internet rather than just being randomly created to sound plausible?

Which is exactly why you shouldn't trust it, because the internet is full of rubbish and doesn't distinguish between rubbish and truth.

Here's a nice example of it failing a simple question.

 

Murmansk

All fair comment, I have to say my limited experience of AI and what it can do has been quite positive but one needs to be cautious. 

Nasqueron

Could be worse, could be using Grok aka Mecha Hitler

GingerTim

Murmansk said:

All fair comment, I have to say my limited experience of AI and what it can do has been quite positive but one needs to be cautious. 

I confess I'm jaundiced by having marked a number of essays over the summer in which students, relying on AI, have been confidently and hilariously wrong about a lot of things...

jimjames

Murmansk said:

GingerTim said:

Never assume that any AI answer is correct (safer to assume the opposite!) - it doesn't actually know things, just presents plausible text based on your prompt.

OK but that text is ultimately obtained from a distillation of stuff on the internet rather than just being randomly created to sound plausible?

This is a perfect example of distillation of stuff that is complete garbage

https://www.google.com/search?q=austria+hungary+in+space

AI can't tell what is spoof, story or truth