I found this video, where I disagree violently with the notion that AI will eventually replace programmers. While AI can certainly assist in coding tasks, as demonstrated by tools like GitHub's CoPilot, CodeLLaMA or other similar local offline capable code AI models, I believe it's unlikely to fully replace human programmers. Programming involves not just writing code, but also problem-solving, understanding complex systems, keeping the overview of the whole codebase, and making nuanced decisions. These are areas where AI currently falls short and may always struggle due to the inherent complexity and variability of these tasks. And AIs aren't free from errors, in other words, one still has to check everything oneself anyway. Therefore, I see AI as a tool that can aid programmers, making their work more efficient and accurate, rather than a threat that could replace them. Therefore, in my opinion, they will never replace human programmers, they are and will always remain just another tool for software development. At least for the foreseeable future.
So you can't give an example and are evading. So I just don't take you seriously anymore until you give an example here. Because this is a typical defensive attitude of those who have no real arguments in the end.
Sorry, but Object Pascal has been providing most of this for everything. So this shows that you haven't really informed yourself.
Even smart-pointers can be implemented in Object Pascal quite fast with a Generics class together with a Reference Counting interface, because, admittedly, such a thing is >not yet< implemented in the Object Pascal Standard Library itself.
Exceptions have ObjectPascal (try finally end; try except an; TException etc.), namespaces too (unit Foo.Bar.Thing.Stuff; TClass.TSubClass.TSubSubClass etc.), generics too (type TClass<T>=class ... end; function Func<A, B>(Foo: A; Bar: B):Boolean; etc.), lambdas (anonymous functions, BlaVar := function:integer begin result := 123; end; ) also, collection literals also (BlaVar := [1, 2, 3];
BlaVar := BlaVar + [42, 23]; FooTValueArray := ['v', 2, 3.14159]), destructured bindings also via RTTI magic, sum types also in the form of Variant or TValue, and so on.
Only type inference a la C++'s auto is not, but in my opinion this is a bad anti-pattern anyway, because Object Pascal code should always be readable, where something like "auto Foo := Bar;" would only interrupt the reading flow, where you would have to think about what data type it is in the end.
Here one sees again what I wrote in my blogpost, that people like you don't really know Object Pascal or because of their arrogant attitude don't want to know what it includes and what not.
But the importance of type inference (which does not, BTW, originate in C++) is that it enables declarations in circumstances where the type cannot be named, or even just not conveniently.
