Why is AI so goddam awful?

[Woke to Sounds]

A.I. can suck it buddy.

 

Especially when it steals creative jobs that humans ought to be paid to do.

 

A.I. created music is particularly DREADFUL.

 

Fxck A.I.!

[GammaGlobulin]

18 minutes ago, Woke to Sounds said:

A.I. created music is particularly DREADFUL.

 

I agree with one fact:

 

There will only ever be one JS BACH.

 

JS BACH = GOD

 

The ONE and ONLY God....!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1

 

IMHO.

 

Note:  AI is not yet God.  AI will only ever be a little god, small  g

 

The God of Music is Bach.

 

[Gottfrid]

17 hours ago, simon43 said:

I thought that AI engines try to mimic humans, not robots.

AI engines mimic what they are fed to mimic by humans.

[Paul52]

18 hours ago, simon43 said:

🙂

Why is AI so goddam awful?  

Because it’s just that, it’s AI not AGI (Artificial General Intelligence).  AI uses LLM (Large Language Models), which has its limits and its flaws.  AGI will be more advanced than human intelligence.  It will be smarter than the smartest person alive or that has ever lived.  And it’s coming.

[Stiddle Mump]

17 hours ago, GammaGlobulin said:

 

AI does not know the rules.

Neither does it know anything.

 

It's only a sequence of zeroes and ones in a computer. No internet no AI.

 

We should't be taken in by this nonsense. Just people trying to get easy money.

[G Rex]

15 hours ago, madone said:

 

nonsense -- the article is crap of a million reasons, but a disease no one can identify is enigmatic. It is a word commonly used in medical circles

 https://www.google.com/search?q=enigmatic+diesase&oq=en&gs

 

apologies the site wont accept the google link

okay Doctor.   I stand corrected.

please accept my most humble apologies 

 

by the way Doctor. 🧑‍⚕️.  What is a ‘diesase’? 

[IsaanT]

I remember the early pioneering days when ChatGPT was first made available to the public.  It was 30th November 2022.  Since then it has evolved and come a long way.  Let's get this in perspective - that's less than 2.5 years.  I will be intrigued to see where we are in ten years' time but it certainly won't be going away.

The responses from any AI system are SO dependent on the request.  For example, if it was requested to provide a news report of a particular instance IN THE STYLE OF The New York Times (insert the news medium of your choice) it will oblige.  As a simple example, ask it to report a news article (or make up a story) IN THE STYLE OF William Shakespeare; I would hope you would be impressed.

As already referenced by CharlieG, there is a whole science now about prompt engineering, i.e. how to construct an appropriate request to eliciit an appropriate response.  In essence, be specific and provide all relevant details in the prompt in order to obtain an appropriate request.

As an example, I could ask AI to provide a route that can take me from Bangkok to Chaing Mai.  It would give a response.  However, If I said 'Please provide me with a route from Bangkok to Chaing Mai.  I would appreciate a comfort/refuelling/eating break around every two hours, and I prefer PTT stations to the others.  Please also suggest places of interest along the way, and I am happy to accept minor diversions along the route to be able to visit them.  Lastly, I would appreciate suggestions which avoid large towns or cities, especially in peak traffic periods.', I will get a very different response (try it!).

In simple terms, it is just like how you feel when you have been asked a question - do you feel you have all the relevant and pertinent information in order to construct a response?

As has also already been stated, it's still the same with computers - Garbage in, garbage out.

Take the time to construct a good prompt and you will be rewarded with a good response.  Should you still be disappointed, please post your query and response here and I or others can help analyse it and make helpful suggestions.

Finally, I remember one or two of my colleagues were openly sceptical about AI when it first appeared.  There will always be doubters.  Personally, I use it almost every day.  At that time my company was paying management consultants (Cognizant) huge sums of money for advice and recommendations that I quickly found could be equalled by ChatGPT.  I feel it is like asking a really wise and infinitely-knowledgeable friend to help me.  I am smart enough to recognise when it might have misinterpreted my request but that would be my fault for not being clear enough (it's rare but it can happen).  Embrace it - your life will be enriched accordingly.

[Purdey]

I am very grateful for its ability to improve the English writing of my staff and now I feel it is good enough to present to the Board without corrections.

[simon43]

4 hours ago, Gottfrid said:

AI engines mimic what they are fed to mimic by humans.

Yes indeed!  My NataChata AI chat engine was programmed by me to mimic a 'slut', and was rather successful at that 🙂 It typically generated an income of about $40,000 per month,which was good income 20 years ago for a sole-trader business.

 

Now where did all that $$$ go? 🙂

[simon43]

I also want to understand why, if I ask ChatGPT to create an image of a man in his 50's or 60's (eg a teacher in the classroom), it insists on creating an image of someone who looks to be in his 70's with a white beard.....

[IsaanT]

30 minutes ago, Purdey said:

I am very grateful for its ability to improve the English writing of my staff and now I feel it is good enough to present to the Board without corrections.

An excellent example of AI use.

I recently proof-read my son's business website (he's just made the break to being his own boss).  I suggested several modifications and enhancements to his partner, who had constructed the website.  Occasionally, when it wasn't entirely black-or-white, I copied the sentence or paragraph into whichever AI engine I was using at the time and asked it to correct the spelling, grammar and punctuation in best English business style (note the specific request).  I have to admit it came up with a couple of ideas that I hadn't thought of but I was pleased and grateful.

Conversely, I remember writing my resignation email notice to all my colleagues in my company.  I thought about using AI to create something but I knew that would be TOO EASY!!  I spent several hours, on and off, constructing my own humourous departure notice and felt better for having created it myself.  Some of my colleagues kindly said it was the best resignation email they'd every seen.

For everything else, I have to have a good reason NOT to use AI...!

[madone]

44 minutes ago, G Rex said:

okay Doctor.   I stand corrected.

please accept my most humble apologies 

 

by the way Doctor. 🧑‍⚕️.  What is a ‘diesase’? 

 

Another term you seem unaccustomed to -- Its called a typo

 

 

[Tom100]

I have used ChatGPT hundreds of times and find that both are excellent with the right prompts. 

 

I gave it a question from a university class I took in 2021 and Chatgpt produced a perfect answer in one minute. It took me a day. 

 

My actuary friend asked the paid ChatGPT to write code to solve a complex actuarial database quiry and analysis. No one in the New York office could solve the coding problem.  ChatGPT wrote the code that ran in SAS and solved the problem fast. 

 

AI is better than most employees and will replace millions.

 

[Watawattana]

I reckon AI will get much, much better when Skynet starts to develop some of their national security and weapon systems tools.

[johng]

44 minutes ago, Watawattana said:

AI will get much, much better when Skynet starts

 

Well yeah   but It won't be "much better" for humanity

ask Sarah Connor 😋 

(Anybody not wearing 2 million sunblock is gonna have a real bad day. Get it? )

 

Just realised that  "The Terminator"  came out in 1984  of all years !!!!!

https://www.imdb.com/title/tt0088247/

[BangkokReady]

14 hours ago, Pouatchee said:

wasnt it noam chomsky who said that?

 

Yeah.  I wish he was still young and sharp.  I'm sure he would have something interesting to say about AI and how it acquires language skills.  Probably something interesting on the legal/moral consequences also.

[Pouatchee]

15 hours ago, simon43 said:

Perhaps ChatGPT was written by a Brit and it can't understand what you mean by 'mustache' 🙂

 

Well... in canada we use either. so, does that mean the brits are clinging to outdated rules just for the sake of it? 🤪

[Pouatchee]

1 hour ago, BangkokReady said:

Yeah.  I wish he was still young and sharp.  I'm sure he would have something interesting to say about AI and how it acquires language skills.  Probably something interesting on the legal/moral consequences also.

 

a trailblazer in many ways... although his communicative approach is now somewhat outdated he did come up with some brilliant precepts my professors swore by in uni. too bad he is not as acknowledged as he deserves. Maybe it has to do with the fact that he went on RT a few times too many? But what choice did he have? Mainstream media don't like being called out. They should have gotten rid of Roger Waters on may editorials and let Chomsky handle it... but he is getting tired...

[BangkokReady]

9 minutes ago, Pouatchee said:

although his communicative approach is now somewhat outdated he did come up with some brilliant precepts my professors swore by in uni

 

Yeah, I read a book called The Language Game which claimed that language acquisition doesn't use the "switches" that Chomsky said so much that it's a game of trial and error.  I wasn't convinced, to be honest.

 

I quite like Hymes' Communicative Competence.  It seems a bit more interesting and relevant to me.  It sounds more like a practical help to understanding the process.  It's probably this with a combination of "switches" and "language games".

[madone]

6 hours ago, Tom100 said:

I have used ChatGPT hundreds of times and find that both are excellent with the right prompts. 

 

I gave it a question from a university class I took in 2021 and Chatgpt produced a perfect answer in one minute. It took me a day. 

 

My actuary friend asked the paid ChatGPT to write code to solve a complex actuarial database quiry and analysis. No one in the New York office could solve the coding problem.  ChatGPT wrote the code that ran in SAS and solved the problem fast. 

 

AI is better than most employees and will replace millions.

 

Im a professional writer who writes complex white papers and case studies based on extensive interviews. I have had very little success training Chapt gp to give me more than an outline without making up lies, conflating its training document with the story materials, and generally creating meaningless drivel.    

believe me, I have tried, even enlisting the help of some developer friends and self-styled prompt engineers. 

 

[Pouatchee]

6 hours ago, BangkokReady said:

 

Yeah, I read a book called The Language Game which claimed that language acquisition doesn't use the "switches" that Chomsky said so much that it's a game of trial and error.  I wasn't convinced, to be honest.

 

I quite like Hymes' Communicative Competence.  It seems a bit more interesting and relevant to me.  It sounds more like a practical help to understanding the process.  It's probably this with a combination of "switches" and "language games".

 

interesting i had to look it up and actually used chatgpt to quickly brief me in on it:

 

Although Hymes’ Communicative Competence and the Communicative Approach (CLT - Communicative Language Teaching) are closely related, they are not the same. The Communicative Approach is a teaching method, while Hymes’ theory provides the foundation for why this method is effective.

 

I'm back teaching in canada where there is a shortage of ESL teachers in my province... OMG the quality of the teachers here... I used to think that the teachers in thailand were not up to par... problem is people set low expectations and expect weak results. I am actually encouraging my students to use chatgpt to give them ideas on topics i give them for story or essay writing. Even with this new tech the kids are too lazy to even use it to get some sort of idea... AI will surely take over these soft potato head kids some day...

 

 

 

[PomPolo]

On 2/19/2025 at 5:01 PM, simon43 said:

Yes, but again to reassure the mods, I'm not specifically criticising that news report or any of the news reports on AN 🙂

 

But if I ask Chat GPT to comment about something, it does seem to respond with more human-like sentences.  I guess there are different AI chat engines...

Definitely different engines and depending on the engine different styled results will be returned the most common ones being chatGPT which offers several models that can return results faster but maybe less accurate or up to date on the smaller models and some that are specifically aimed at different subjects like coding.

I have been setting up local AI's like deepseek and others using ollama and open-webui which essentially look and act the same as chatGPT with my favourite bonus all your questions if any are sensitive are not stored on someone else's server, to run some of the larger models to get true chatGPT results and performance you would quite an expensive computer, however most of the mid range models will run happily on a PC with a decent graphics card.

[Cardano]

On 2/19/2025 at 4:30 PM, simon43 said:

Firstly, this is not a criticism of AN or the use of AI in their news reports!  This is a general comment about AI.

 

However, I was reading the recent news report about a British man who has been hospitalised in Thailand, with no medical insurance etc.  (You've heard this story time and time again!).

 

But what struck me was how AI news reports bear little similarity to actual human journalists who might report a similar story.

 

The vocabulary used is.. well ... basically not human! 🙂  No real person would use the vocabulary and phrases that are used in AI news reports.

 

... and that makes me think "Why?".  I thought that AI engines try to mimic humans, not robots.  Why do AI news applications use vocabulary and phrases that immediately identify them as AI?  (Perhaps that's what they are trying to do...)

 

As reported in another thread from about 2 years ago:

 

https://aseannow.com/topic/1290809-so-called-artificial-intelligence-ai/page/3/

 

19 years ago, I wrote the 'AI' SMS text-chat program called Natachata which the BBC stated was (at the time) the best candidate for passing the Turing test. http://news.bbc.co.uk/1/hi/magazine/3503465.stm

 

(in other words, fooling a real human that they were actually chatting with another human).

 

I'm sorely disappointed with the vocabulary and style of writing of the AI 'news bots'. Definitely room for improvement 🙂

Don't blame the machine, blame the human quality controler.