Add Artificial Intelligence features into the Object Pascal Language

[Andrew1020]

I am just casually asking that AI features to be build into the Object Pascal language when the developer team deems fit.

[MarkMLl]

I see absolutely no reason for this to be part of the language.

It can easily be put into a library/package, "well volunteered that man" :-)

MarkMLl

[Khrys]

What do you mean by "AI features", exactly? Can you be more specific?
What is it that would require support at the language level of all things?

[zeljko]

You have aiassistant package in lazarus (currently Jan is supported).

[cdbc]

Hi

I am just casually asking that AI features to be build into the Object Pascal language when the developer team deems fit.

When I see the sad excuses for 'source-code', that most AI *Gulps up*, I'm thinking NO! To put it bluntly Hell No!
Regards Benny

[MarkMLl]

I am just casually asking that AI features to be build into the Object Pascal language when the developer team deems fit.

When I see the sad excuses for 'source-code', that most AI *Gulps up*, I'm thinking NO! To put it bluntly Hell No!

Again: there is absolutely no reason why something like that should be in the core language. If OP doesn't understand the distinction then he needs to go away and do some reading.

"Code suggestions" could potentially be available as an optional package for the Lazarus IDE, or if not then use something different (Eclipse?).

"Adding an 'AI' interface to a program being written" would obviously be an external library or package: nothing whatsoever to do with the core language EXCEPT that multiline strings would obviously be useful to hold the prompt used to seed the request.

MarkMLl

[TRon]

I am just casually asking that AI features to be build into the Object Pascal language when the developer team deems fit.

Inside the language, even. tsk. Not wanting to discourage you from expressing your opinion but do you have any idea what you are talking about ?

Which reminds me, just came across got this in my feeds about what using CoPilot seem to be doing to your code-base and the implications that could have for maintainability of the code. Stuff that managers should have a look at instead of FOMO.

[Joanna from IRC]

I am just casually asking that AI features to be build into the Object Pascal language when the developer team deems fit.

Another helpful account has spawned to Remind us that we urgently need to replace ALL pascal programmers with AI 

[MarkMLl]

Another helpful account has spawned to Remind us that we urgently need to replace ALL pascal programmers with AI 

In fairness Joanna, if you checked OP's other posting you'd see that it was a fairly coherent comment about moving away from Delphi when his licence expired and was made in 2023.

Now unfortunately what he's posted today is ill-advised, but it might possibly be because his experience is Turbo Pascal and Delphi rather than FPC/Lazarus and he's never really considered the distinction between language and implementation.

Which is one of your failings as well, so I suggest trying to be just a /little/ bit less confrontational as far as new forum members are concerned. OK?

MarkMLl

[VisualLab]

What do you mean by "AI features", exactly? Can you be more specific?
What is it that would require support at the language level of all things?

I join the above request. I would really like to know what “building AI into a programming language” is and how it is done. Because I don't understand it. I would just like to ask for a simple explanation. Because at this stage, I imagine it this way: all this "data grinding machinery" (i.e. all these neural networks connected to data transformers) is integrated into the compiler. But I'm probably wrong, because this is a behemoth-software. And what about the database that this machinery uses? Should this also be put into the compiler?


This AI diarrhoea is spreading in ever wider circles. But there will be a crash when the bubble finally bursts

[Curt Carpenter]

I agree with MarkMLI that the OP may have used "the language" to mean "the programming environment" -- and if that's the case, I think the question of what "AI" features might be built into the environment is an interesting one.  For example:  when the compiler reports an error, AI could perhaps allow it to offer a suggested fix as well (if asked). 

[440bx]

Maybe there is already some artificial intelligence in FPC.  Those "illegal expression" and "semicolon expected" messages look rather suspicious to me

[LV]

If the topic means built-in AI features, native support for symbolic reasoning, or optimization for AI applications. Programming languages ​​such as Prolog, Lisp, Wolfram ​​are well-suited for these purposes.
While other languages do not have AI-specific features at the language level but rather rely on libraries and frameworks (e.g., TensorFlow, PyTorch).

This AI diarrhoea is spreading in ever wider circles. But there will be a crash when the bubble finally bursts

I shared this phrase with ChatGPT, and it was not offensive; in fact, ChatGPT even agreed with you in some aspects. 

[Curt Carpenter]

I don't think so.  Machine learning is demonstrating some real (and commercially viable) value now.  I was there in the first wave of AI euphoria in the late 80's (and helped kill chasing it at the company I was with in those days) -- but something is different this time, thirty+ years later.  My doctor says the impact in medicine, for example, is real.  Early days still.  Even the "chat" LLM systems are reportedly having a material impact on education processes.

[MarkMLl]

I don't think so.  Machine learning is demonstrating some real (and commercially viable) value now.  I was there in the first wave of AI euphoria in the late 80's (and helped kill chasing it at the company I was with in those days) -- but something is different this time, thirty+ years later.  My doctor says the impact in medicine, for example, is real.  Early days still.  Even the "chat" LLM systems are reportedly having a material impact on education processes.

Yes, it /is/ fundamentally different (and I had a university job during the last bubble).

But I believe it is fair to ask ourselves whether the current "spicy pattern matching" really is an adequate substitute for reasoning based on organised ontologies etc. (and if it is, what it tells us about ourselves and potential other forms of intelligent life).

MarkMLl