It was a joke. But one of the points of higher-level abstraction languages like C++ is the portability. The point of the joke was that C/C++ and all high-level abstraction languages is to make it closer to natural language to make programming easier for people. And it adds a degree of separation between the programmer and the “real” code. The compiler turns the pidgeon English C++ code into ASM, and then into object code from there. much like AI turns plain language into C++ code. So really vibe coding just adds one more layer to the abstraction cake.Re5etuk wrote: Sun Jul 05, 2026 1:42 pm i dont think people got 'lazy' and moved to c / c++ etc , it was the hardware people writing for .
Anyone else noticed the increase of Vibe coded plugins flooding the market?
- KVRAF
- 7718 posts since 2 Sep, 2019
THIS MUSIC HAS BEEN MIXED TO BE PLAYED LOUD SO TURN IT UP
- KVRian
- 631 posts since 24 Feb, 2008 from Germany
<3Bunny_boy wrote: Sun Jul 05, 2026 2:04 pm1995? YoungunTiles wrote: Sun Jul 05, 2026 1:54 pm I just loved non programming tools back in the days. RPG Maker 95 ftw !
https://en.wikipedia.org/wiki/Graphic_Adventure_Creator
I started relatively late, with my first Windows 95 PC
“The biggest crime of a musician is to play notes instead of making music.”
Isaac Stern
Isaac Stern
-
- KVRian
- 861 posts since 26 Aug, 2005 from Oregon, USA
The problem with blind vibe coding is using code that the originator does not understand. That's where I draw the line. Someone needs to know the dependencies, performance issues, backwards compatibility et rest et rest.
- Beware the Quoth
- 35496 posts since 4 Sep, 2001 from R'lyeh Oceanic Amusement Park and Funfair
No, that's not actually true, as per this example of one of the earliest high-level languages.jamcat wrote: Sun Jul 05, 2026 2:08 pm The point of the joke was that C/C++ and all high-level abstraction languages is to make it closer to natural language to make programming easier for people.
or some APLP1 max3 (V0[:8.0],V1[:8.0],V2[:8.0]) → R0[:8.0]
max(V0[:8.0],V1[:8.0]) → Z1[:8.0]
max(Z1[:8.0],V2[:8.0]) → R0[:8.0]
END
P2 max (V0[:8.0],V1[:8.0]) → R0[:8.0]
V0[:8.0] → Z1[:8.0]
(Z1[:8.0] < V1[:8.0]) → V1[:8.0] → Z1[:8.0]
Z1[:8.0] → R0[:8.0]
END
{(+⌿⍵)÷≢⍵}
The point of all high-level languages was abstraction at a higher level, hence the name. In some programming languages, some resemblance to some parts of some natural language was the result of that abstraction.
However ASM, which you refer to earlier, is also an abstraction, albeit with a 1:1 correspondence between abstraction and object code. 'Higher level' refers to the abstractions have a complex correspondence with the object code, abstracted away from the machine-specific level.
No, that's incredibly out of date.The compiler turns the pidgeon English C++ code into ASM, and then into object code from there.
No, its not much like that at all. A C++ compiler (especially old compilers targetting ASM as an intermediary form) would generate the exact same code from the same input; there's a direct correspondence between source code and resultant object codemuch like AI turns plain language into C++ code.
That's not true of an AI.
The first example is basically transliteration, the other is very much not.
An idiot on Set Theory:
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
- KVRist
- 102 posts since 2 Jul, 2021 from Netherlands
I don't think the point of higher-level languages is to get ever closer to natural language, but to provide useful building blocks that let us describe how software should behave. Over time, as computers have become more powerful and our understanding of software development has improved, these building blocks have also become more complex and powerful. But the goal has never been to write in regular English (or whatever human language), since English doesn't allow us to express computations in a succinct way.jamcat wrote: Sun Jul 05, 2026 2:08 pm The point of the joke was that C/C++ and all high-level abstraction languages is to make it closer to natural language to make programming easier for people.
My audio programming blog: https://audiodev.blog
- Beware the Quoth
- 35496 posts since 4 Sep, 2001 from R'lyeh Oceanic Amusement Park and Funfair
I guarantee that if people werent using typewriter-style keyboards on computers, english words would not be a prevalent form of symbol used in programming languages; if they'd been using keyboards with logical and mathematical symbols on them, that's what the symbols used would be in programming languages.kerfuffle wrote: Sun Jul 05, 2026 2:48 pm
I don't think the point of higher-level languages is to get ever closer to natural language, but to provide useful building blocks that let us describe how software should behave. Over time, as computers have become more powerful and our understanding of software development has improved, these building blocks have also become more complex and powerful. But the goal has never been to write in regular English (or whatever human language), since English doesn't allow us to express computations in a succinct way.
use of 'written language' words is primarily an artefact of the technologies available, not a reflection of purpose.
(edited to complete my point)
Last edited by whyterabbyt on Sun Jul 05, 2026 2:58 pm, edited 2 times in total.
An idiot on Set Theory:
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
- KVRian
- 631 posts since 24 Feb, 2008 from Germany
I see it somewhat differently. The ability to keep every detail of a language or low-level implementation in one's head is becoming less central, as AI can handle a large part of the actual implementation. What matters more is development in the sense of defining goals, architecture, UX, and making sound technical decisions at the system level.ksandvik wrote: Sun Jul 05, 2026 2:16 pm The problem with blind vibe coding is using code that the originator does not understand. That's where I draw the line. Someone needs to know the dependencies, performance issues, backwards compatibility et rest et rest.
Of course, this knowledge still needs to be built up. Traditionally, this happened by writing code, making mistakes, and understanding systems in detail. That remains a very direct and effective way to learn.
At the same time, working with more abstract development tools is nothing new. Long before AI, there were no-code and visual development environments like RPG Maker 95, Multimedia Fusion, and Adobe Director. Many developers started with tools like these, focusing on logic and systems rather than traditional programming. More modern examples include Unreal Engine Blueprints.
AI-assisted or no-code approaches continue this trajectory and can provide alternative entry points into thinking in systems, architecture, and product decisions. Deeper technical understanding often develops later through more concrete implementation.
Ultimately, any language like C++ is just a tool. What matters is understanding what needs to be built, the trade-offs involved, and how the system behaves. Implementation itself can increasingly be supported by tools and AI, depending on what is most effective for the task.
“The biggest crime of a musician is to play notes instead of making music.”
Isaac Stern
Isaac Stern
-
- KVRAF
- 16792 posts since 13 Oct, 2009
^^^yes. There's a balance here. How many people understand anything, let alone the basics, of the assembler code that the compiler generates? I knew a guy when I was a kid attending computer club meetings. He worked at a bank and his job was writing internal applications for the bank. His language of choice was IBM assembler for mainframes. It used to be somewhat common to insert small bits of assembler inside of C code, that's far less common today. We have seen the abstraction level get raised multiple times over decades. Even embedded code isn't much written in assembler anymore.
That said, I do think that education has to change to accommodate this. Systems thinking and abstraction are more important than ever now and most curricula are still built around human coding as dominant. Additionally, I'm not convinced that the current SDLC approaches are ultimately going to be sustainable. We give lip-service to human in the loop and overburden developers with managing the comprehension debt accumulated through LLM generated code, and, moreover, those LLMs are still trained on dominantly low level patterns that are viewed as problematic in larger code bases. This is less of a problem in industry where the tooling and workflows can be managed, but is a real problem for vibe-coders who don't have the budget and/or skills to implement more durable processes.
Saying C++ is just a tool is going further than I would go, but, I think that we are seeing a shift. Ask people who code and who aren't threatened and I think that you'll start to realize that the earth is shifting, regardless of what the (AI) industry itself is doing, and I don't really care about that one way or the other, LLM driven coding isn't going anywhere.
That said, I do think that education has to change to accommodate this. Systems thinking and abstraction are more important than ever now and most curricula are still built around human coding as dominant. Additionally, I'm not convinced that the current SDLC approaches are ultimately going to be sustainable. We give lip-service to human in the loop and overburden developers with managing the comprehension debt accumulated through LLM generated code, and, moreover, those LLMs are still trained on dominantly low level patterns that are viewed as problematic in larger code bases. This is less of a problem in industry where the tooling and workflows can be managed, but is a real problem for vibe-coders who don't have the budget and/or skills to implement more durable processes.
Saying C++ is just a tool is going further than I would go, but, I think that we are seeing a shift. Ask people who code and who aren't threatened and I think that you'll start to realize that the earth is shifting, regardless of what the (AI) industry itself is doing, and I don't really care about that one way or the other, LLM driven coding isn't going anywhere.
- KVRAF
- 7718 posts since 2 Sep, 2019
Right. It used to be common to tweak the ASM by hand. But as compilers got better, that practice became obsolete, because the code put out by the compilers became more optimized than people could do.
The same thing is happening with C++ code as LLMs continue to get better at coding, and understanding what prompters are asking for.
The same thing is happening with C++ code as LLMs continue to get better at coding, and understanding what prompters are asking for.
THIS MUSIC HAS BEEN MIXED TO BE PLAYED LOUD SO TURN IT UP
-
- KVRian
- 705 posts since 11 Apr, 2006
This is still not true. Compilers remain worse than skilled programmers at this. The "I don't need to worry about this because the compiler is better anyway" thing is self-propagating copium among people who don't want to spend two afternoons learning it.jamcat wrote: Sun Jul 05, 2026 3:26 pm Right. It used to be common to tweak the ASM by hand. But as compilers got better, that practice became obsolete, because the code put out by the compilers became more optimized than people could do.
(Edit: just to be clear, I used to be one of those people, too.)
Last edited by tumface on Sun Jul 05, 2026 6:14 pm, edited 1 time in total.
-
- KVRian
- 861 posts since 26 Aug, 2005 from Oregon, USA
The problem with LLM code is that it's based on a huge depot of code from github, and we all know it's a hit-and-miss concerning quality. LLM is good at giving ideas, but then the good programmer does the actual coding understanding how it all works.
-
- KVRer
- 4 posts since 5 Jul, 2026
Tiles wrote: Sun Jul 05, 2026 9:14 amImho the market was already oversaturated when AI emerged. It is speeding things up a bit. But at the same time th emarket is not longer 100% DAW, but x% DAW and x% AI like Suno or Tunee. So in the sum it might even be that the music market has grown bigger since now people "makes" music who would never have done it with a DAW.tidalwaves wrote: Sun Jul 05, 2026 8:37 amNot due to AI, but the audio software industry doesn’t seem to be doing too well. I can only imagine a flood of sloppy clone plugins could be the nail in the coffin of companies that were already struggling to sell their plugins at 90% off…Tiles wrote: Sat Jul 04, 2026 6:11 am Well, AI has been widely deployed for close to four years now, and the world has not collapsed or seen broad mass unemployment. That makes me somewhat optimistic.
That said, there are clearly areas under pressure. Especially in programming and graphics department where work is highly standardised. The impact so far looks more like uneven restructuring and shifting demand than a general collapse of employment. I'm curious how this turns out in let's say ten years.
Good point, could be. The total amount of released music is constantly increasing, and AI plays a big part in that, that’s for sure.
On the other hand, audio software companies are constantly doing sales, many of them are being sold left and right, and many artists have been complaining about the streaming business model for years.
