With the steady progress of Nimony, I wonder, what will AI's make of it?
Will the whole AI thing hinder the introduction and rise to fame of new languages? AI's will need training material, coders "need" AI....
So good specs (BNF?) would help.
ChatGPT5 is a year old, the last Claude a bit less. I asked them and the Chinese one, they know nothing of Nimony / Nim3 for example.
I discovered with AI types matters less. And working with julia became much easier.
Basically AI makes job of compiler. And I assume in near future it will do much better.
I think the most important is to optimse for optimal representation for human. Meaning - simple, compact, flexible code, with no noise and distractions. Implicit and implied concepts, less direct control and micro management.
Every good AI so far relies on the compiler's error messages, including when it points out type errors.
As for your other point, I think this "simple compact flexible pseudo-code" will be outdated within 1 week and you won't have a job as a programmer anymore at all, enjoy.
you won't have a job as a programmer anymore at all
Yes, no more need for machine translators. :)
Every good AI so far relies on the compiler's error messages
No, AI frequently able to figure out problems just by looking at code and analysing it logically, without running it.
Eventually AI will superceed humans and the only way to move forward long term is to integrate with AI via neurolink etc.
Sounds like Elon Musk already has unrestricted access to your brain.
Sounds like Elon Musk already has unrestricted access to your brain.
Ah, so can we now write plug-ins for it?
I do hope modern tech will become obsolete. As will the tech that come after that, and so on.
But how does that motivate not learning it? The are problems all around us, big and small, that need solving, and we have the tech we have. When new and better tools come, they are to be learned. Now I learn what there is now.
Or are you saying that the tech is already obsolete? Then was has superseded it? The current "AI" is very lacking (useful still), and I bet the skills related to it, like convincing the bot that your grandma will die if it doesn't fix the bug right now, will become obsolete way, way faster than what you call the "modern tech".
As for AI superseding humans, I don't see how that is relevant for choosing what to learn in the current moment.
Nim is that pseudocode. Or one dialect of pseudocode. Think algorithms, data, express pseudocode, aka Nim. I hope Nimony and Nim 3 will make it more enjoyable and consistent to pseudocode.
Beyond that, whatever you describe has been discussed in the book, and ongoing research coverage of "AI snake oil".
Feynman has already answered this for our world: "The machine can calculate beyond what humans can, but that doesn't mean it is thinking. Calling a computer conscious is like calling a hammer angry because it nails hard".
At its core, it is still a deep pattern matching + statistical learning trick exposed to training leakage. In many cases, LLMs have already seen the solutions they're supposed to generate.
"Eventually AI will superceed humans and the only way to move forward long term is to integrate with AI via neurolink etc." - You seem to play too many video games ;)
Deep thought: All intelligence is artificial
Human intelligence is also the product of electrical charges, changing thresholds, past experience, genetically determined hallucination abilities (also known as imagination), and special builtin attention mechanisms (i.e. Id, Ego and superEgo). There is also the soul, which is all the high-performance tuned code written ages ago by the mate that left the team, and that no-one at the office can actually understand (agile code).
My point being, yes, it is still pattern matching, but aren't we all?
My point being, yes, it is still pattern matching, but aren't we all?
Nah,
Inspiration Particles sleet through the universe, each heralding a moment of brilliance: a new symphony, a way of getting from A to B quicker than before, lines for a new play, or deeper understanding of something than was previously comprehended.
Most of them are doomed to miss, or to reveal their brilliance to a brick wall or a starling, which is totally unequipped to deal with the revelation.
Some however, hit the right mind at the right time, and a little later you are blinking foolishly in the TV lights and wondering how the hell you thought of sliced bread in the first place...
The thing is, there is an impedance mismatch here with the AI cores.