AI usefulness discussion

[affonso]

I believe that artificial intelligence is here to stay and I don't understand why many people are prejudiced against it. Many people are prejudiced against the visual programming of the Lazarus IDE, but as we know, this does not limit the IDE; on the contrary, it is just another resource to help with rapid development. I see artificial intelligence as just another resource.

I am not a programmer, I don't know how to code, but using ChaTGPT I was able to simply create a carousel component to display images with the help of ChaTGPT who suggested me how to implement it and showed me the code. This is amazing! It helps you understand and code, which is fantastic.

Another case, when I was following a tutorial that taught how to create an API in Golang, using the VsCode IDE with the Codeium extension, an opensource alternative to Copilot, at a certain point the project did not compile and the Codeium chat simply analyzed my entire project and humanly guided me what the project error was. I was speechless, so I thought, I want this in Lazarus.

I know that the Lazarus and Delphi community is very active, so I soon found this fantastic project, Delphi AI Developer (Copilot), which is open source and integrated with Gemini AI, which is free, but also supports others such as OpenIA and Grop. The most interesting thing is that the latest version supports database integration.

The author of the project asks for contributions with code and dissemination of the project. I found it interesting to export this project to Lazarus IDE, but unfortunately I am not a programmer nor do I have that ability, but here is a tip for anyone who is interested.

Adding a feature to turn the AI ​​on or off would also be interesting. I will continue to love Lazarus as it is and I am very grateful for the community.

Thanks avra for the links, I will try them out!

https://www.youtube.com/watch?v=HJVSrTWWchU

https://github.com/Code4Delphi/Delphi-AI-Developer

[Joanna from IRC]

I am not a programmer, I don't know how to code, but using ChaTGPT I was able to simply create a carousel component to display images with the help of ChaTGPT who suggested me how to implement it and showed me the code.

What great news. Now all programmers can be replaced with chat gpt!

You’ve been coming here since 2012 , is 12 years not enough time for you to learn how to code? I want to be like you, you’re amazing. 

[Curt Carpenter]

You’ve been coming here since 2012 , is 12 years not enough time for you to learn how to code? I want to be like you, you’re amazing. 

Snark for snark's sake?

I don't consider myself a "programmer" either, although I've certainly written a lot of stuff I call "programs."   I can play a few tunes on the piano too, but don't call myself a pianist.

Etc..

[Joanna from IRC]

@curt carpenter
I know I might be a little unreasonable...
Who am I to expect that people in a forum for pascal compiler would contain pascal programmers wanting to talk about programming?
Anyone who wants to advertise anything should be allowed to post here

[Bogen85]

I know I might be a little unreasonable...

A little?
Once again, this a software tool we are talking about.
I know programmers at all experience levels who benefit from it.
And use it as a learning aid.

Just like some find no benefit (for themselves personally) in RAD tools like Lazarus for what they do when programming, it is fine that some don't like the idea of AI tools.

So you go to write a complex configuration file for something (that has no IP (intellectual property in it)) that you use to know how to do 10 years ago, and it took you days to learn the first time around.

But you don't want to spend hours on a refresh, you remember the high level concepts, have a quick AI chat, and in a few minutes have a working configuration file.
You get your application up and running in under 10 minutes, rather then spending a half a day.

Same goes for some non IP code snippet someone does not remember how to do off the top of their head. Why spend hours instead of minutes?

Will AI tools replace sysadmins and programmers? I don't believe that will happen in the foreseeable future. Anyone who says the contrary I do not believe have used AI programming aids extensively.

I see AI as an amplifier of existing skills, just like any other tool.

But to deride anyone who uses tools you don't like, that is not showing common courteously.

Yes, users of different OSes can get very hostile towards those not using their OS.

But they can also be very helpful to users of OSes when it comes to a common programming language passion.

Yes, many users in the forum don't like AI. I understand that.

But their are also users that think very lowly of Linux, others that think very lowly of MacOS, and others that think very lowly of Windows.

Should all users that think very lowly of the other OSes deride users posting free pascal or Lazarus questions concerning their use on an OS other than their favorite?

[Curt Carpenter]

I know I might be a little unreasonable...

I'm sure your beloved by your colleagues and a joy to work with. 

[Joanna from IRC]

I know I might be a little unreasonable...

I'm sure your beloved by your colleagues and a joy to work with.

The objective is to get things done not be a joy 

But you don't want to spend hours on a refresh, you remember the high level concepts, have a quick AI chat, and in a few minutes have a working configuration file.
You get your application up and running in under 10 minutes, rather then spending a half a day.

I want to talk to live people about pascal not to a glorified search engine ai.
Why would you want to replace friends with AI? That’s really disturbing. Do you not have any friends to chat with about code? I don’t know everything but I’ll be your friend if you have no one else.

I don’t care if people use ai instead of writing their own code because they want to fake it instead of learning to program. What I do take issue with is persistent advertising of ai here no matter many people tell them no thanks they never give up and keep making threads  about it over and over. It’s annoying.

[TRon]

Yes, many users in the forum don't like AI. I understand that.

I share the sentiment of what you wrote but picked this sentence because even though sharing the sentiment it is the buzz that is killing it for me.

It seems not a matter of choice anymore but rather a we shove it in your face until you like it kind of attitude that is bothering me. That results in all kind of software integrating such features (for which we did not ask) and forcing you to use these features (on by default, always there in your face etc).

So while I am not opposed to have something like a plugin I am certainly against forcing it upon users.

For work I am not allowed to use any kind of AI for obvious reasons that any software engineer has to deal with. That 'most' developers seem to ignore common sense is their own prerogative.

imho the whole AI buzz is an enourmous bubble that is never going to match expectations (practice has already proven that). so while you and me (and maybe other readers as well) might perhaps have reasonable expectations, the majority of zealots certainly do not. And the most alarming part about these zealots is that they consist of governments, law-makers, banks, hospitals etc.

The basics of AI is the same as a lottery ticket, everything happens by chance unless manipulated to do something else... and who actually does the manipulating ... and in what way are things manipulated ? Nobody seem to ask common questions anymore these days. It is never about the how, but the why.

2 cents

[Martin_fr]

Well, AI (some AI engines) is at the very leat a great search engine. And can be way better at finding the result (or way worse / some engines are, IMHO).

And lets face it, who here hasn't googled before? And gone through many a page listed, until finding one that hold an understandable answer. AI can (sometimes) shorten that.
And for anyone looking for an example, it does not matter if it came from a website or an AI (both can have good and/or terrible results).

As for, how much copy and paste, and how much to still work out and learn from... That question is the same as before AI. Though the temptation to do more copy and paste may be bigger now.

Its not the tool, its how it is being used.


For me, any playing around (just to see what would happen) has resulted in no useful code. I.e. any request for code that I can't write myself in the time that it takes to ask an AI for it, has resulted in erroneous code.

Asking an AI, to check some piece of code, has (for me) 99.9% resulted in a list of errors that were NOT present in the code.

But other's experiences may differ.

And AI can and does IMHO deliver reasonable templates for documentation. Tiny bit of editing may be needed, but overall faster than drafting it all by hand. Well, I have only tested it on smaller fragment of code. But given that some of them even where out of context, the result were really good.

[Curt Carpenter]

imho the whole AI buzz is an enourmous bubble that is never going to match expectations (practice has already proven that).

I was working when the first "AI" wave struck.  The company I worked for got very involved and it took two years to put out the fire and get people to take a realistic look at the technology.   But this wave is genuinely different I think.  Consider the Nobel prize this year for chemistry:

The Nobel Prize in Chemistry 2024 is about pro­teins, life’s ingenious chemical tools. David Baker has succeeded with the almost impossible feat of building entirely new kinds of proteins. Demis Hassabis and John Jumper have developed an AI model to solve a 50-year-old problem: predicting proteins’ complex structures. These discoveries hold enormous potential.

It's a mistake to dismiss results like that this time.  It's not the LLM stuff, but the machine learning stuff behind the LLM stuff.

That said:  I just had some work done by an attorney.  It involved a) understanding an aspect of "the law"; b) generating a few pages of legal boilerplate with a place for some signatures  at the bottom; and c) knowing where to send the signed papers.  I have no doubt that a current AI could have done all three steps given five minutes of the attorney's keying in a description of my legal issue.  Maybe that explains why the legal community is so interested in "regulating" the technology!

[LV]

Today, while reading the news, I came across an opinion piece. The Linux creator is interested in AI, but the hype means he "basically ignores" it.
https://www.tomshardware.com/tech-industry/artificial-intelligence/linus-torvalds-reckons-ai-is-90-percent-marketing-and-10-percent-reality.

The Nobel Prize in Physics 2024 was awarded jointly to John J. Hopfield and Geoffrey E. Hinton "for foundational discoveries and inventions that enable machine learning with artificial neural networks".
https://www.nobelprize.org/prizes/physics/2024/summary/

[Curt Carpenter]

Today, while reading the news, I came across an opinion piece. The Linux creator is interested in AI, but the hype means he "basically ignores" it.
https://www.tomshardware.com/tech-industry/artificial-intelligence/linus-torvalds-reckons-ai-is-90-percent-marketing-and-10-percent-reality.

The Nobel Prize in Physics 2024 was awarded jointly to John J. Hopfield and Geoffrey E. Hinton "for foundational discoveries and inventions that enable machine learning with artificial neural networks".
https://www.nobelprize.org/prizes/physics/2024/summary/

There is a lot of hype, just as there was in the 80s -- but this time we have two 2024 Nobels crediting  the underlying technology and that has to be taken seriously in my view.  Hard to say who is ahead in the hype race:  AI or quantum computing. 

[kapibara]

I like Affonso's topic. Some years ago AI for programmers was discussed in this forum and for the first time I became aware that it can even write code. At the time I didn't try that, not until this year. O'boy how useful it is. If there is a complicated SQL query, or some very complicated programming logic, or some ellusive typo in the code or what ever. Its really neat to have AI support. I use it almost every day now, not only for programming but for carpentry, cooking, baby nursing, car repair etc.. My effectiveness is doubled or trippled, even things that was impossible before or at least very timeconsuming is sometimes just a walk in the park. So thats fantastic.

But another question, and not hard to predict, is that when AI grows in popularity (which it is bound to do, and at least in parity with the popularity of facebook!) and enough people discover it's incredible usefulness, they will want it. ALL THE TIME, 24/7. Much further on, I predict employers not accepting employees without AI support since they are less effective. Just like you are required to have a cellphone at work today. And then.. why not just think your question? AI replies, with the answer carried to you by WiFi or 5G and channels into your mind through... a CHIP of course. Deep waters, isn't it? https://www.youtube.com/watch?v=63bT5pn5CT4

In the meantime I enjoy all the support I can squeeze out of any AI available.

[RayoGlauco]

When people talk about AI, I always remember an old story.

It was 1983 and I was reading a book by a chess grandmaster. The grandmaster analyzed in the book a game in which the machine lost miserably against a good human player. Then, he explained why a machine could never play chess better than the world champion. Obviously, he was wrong.

My opinion is that, whatever a computer can do today, it may do it inaccurately, slowly, with errors, etc. But one day it will do it much better than any human. Over time, techniques and algorithms will be improved and refined, and we will have much greater computing power.

I do not dare to say that a computer will never program better than a human programmer.

[Joanna from IRC]

I can’t say much for veracity of the Nobel prize considering some of the people they awarded the peace prize to. It seems like it’s influenced by politics. The thing that bothers me about ai is it encourages people to do even less thinking than they do now.

Nobody ever thinks about who owns and controls all this ai stuff. I doubt that ai was invented for altruistic purposes. At best it is a way to plagiarize other people’s work and sell it at a profit. At worst it will lead to a very dystopian future. I don’t like either option.

I don’t think it will solve problem of bad web search results. What you get in the search will be determined by whomever owns the search engine just as it always has. Before that, knowledge was limited to whatever you could think up yourself or asking people you knew or the approved books from library. Knowledge has always been very limited and biased and always will be. AI will not improve this situation.