Why are there thousands of programming languages? And why understanding them makes you better at using AI

The newest programming language is plain English.

That sounds like a joke, but it isn't quite one. In early 2023, Andrej Karpathy, one of OpenAI's founding members, wrote that "the hottest new programming language is English." His point: you can now describe what you want in ordinary language, and AI writes the code. For a Vietnamese business owner, the same is true of Vietnamese.

In early 2024, Nvidia CEO Jensen Huang went further: the tech industry's job is to make computing so easy that nobody has to program, and the programming language becomes human language.

Many people concluded that there's no need to learn to code anymore. That is half true.

You don't need to memorize syntax or remember where every semicolon goes. But the better you understand how a computer works, the better you can hand work to AI. Think of building a house: you don't need to mix the concrete yourself, but an owner who knows where the foundation, the frame, the wiring and the plumbing are won't get a contractor "fixing" the wrong thing.

This article draws that map for you. Picture the computer as a 5-floor building. Each floor is one layer of the system, and each layer has its own languages. You won't need to read a single line of code.

The short answer: why so many languages?

People have created thousands of programming languages, although everyday software relies on a few dozen popular ones. There are three main reasons:

1. Each floor has its own language. The closer you are to the hardware, the more control you have, and the more effort everything takes. The higher you go, the faster you can build, and the less control you have. No language is best on every floor.
2. Each job has its own tool. Websites, data analysis, phone apps and machine control each have a language that fits best. An electrician and a carpenter don't share a toolbox.
3. New languages are created to fix the weaknesses of old ones. C was created so people didn't have to write Assembly. Python was created to make code easier to read. Rust was created to prevent the memory bugs that C is prone to.

Now let's go from the ground up.

Floor 1 · The ground: hardware

A modern chip contains billions, sometimes tens of billions, of transistors. Each transistor is a tiny switch: current on or current off, which is to say 1 or 0. Those two states are all a computer understands. Everything you see on screen, from a photo to a chat message, ends up as long strings of 0s and 1s.

There is no programming language on this floor yet. There is only electricity, circuits and physics. The people who work here are chip and hardware engineers.

A familiar example is overclocking a CPU. Gamers who want a faster machine don't write code. They go into the BIOS and adjust voltage and clock speed, which means working directly on the ground floor. You can only fix a problem once you know which floor it lives on. That is the core idea of this whole article.

Floor 2 · The foundation: machine code and Assembly

A CPU can only run machine code: instructions written in 0s and 1s, along the lines of "take the number in this memory slot, add the number in that one, and store the result in a third." Almost nobody can write raw 0s and 1s, so each instruction got a memorable name: ADD to add, MOV to move data. That is Assembly.

Assembly is very fast and controls every step the chip takes. The price is that small tasks take a lot of writing: printing one line of text on screen can take a dozen instructions. Each chip family (Intel and AMD in computers, ARM in phones) also has its own Assembly, so code written for one doesn't run on another.

Today Assembly is mostly used where absolute precision matters: the code that boots a machine, parts of device drivers, and routines that need to be fast down to the microsecond. Almost nobody writes a whole application in it anymore.

Floor 3 · The frame, wiring and plumbing: the operating system, written in C

In the early 1970s, Dennis Ritchie at Bell Labs created the C language, and he and his colleagues used it to rewrite the Unix operating system. That was a turning point: for the first time, an operating system wasn't tied to one kind of chip. You write it once in C, and a compiler translates it into machine code for each chip.

The operating system is the building's frame, wiring and plumbing. It shares out memory, manages storage, networking, the keyboard and the screen, and lets the applications on the floors above share those resources without stepping on each other. The cores of Windows, macOS, Linux and Android are written mostly in C and C++.

C has a weakness too: it gives programmers full control over memory, which makes mistakes easy. Microsoft has said that about 70% of the security vulnerabilities it fixes each year come from memory-safety bugs. That is why Rust was created, and since late 2022 it has been part of the Linux kernel. New languages are born to patch the weaknesses of old ones, exactly as the third reason above says.

When do you touch this floor? When you rent a server for your website: Linux or Windows, how much RAM, how much storage.

Floor 4 · The rooms: applications, one language per room

This is where most of the software you use every day is built, and it's the biggest reason there are so many languages. Each room has a purpose, and each purpose has a language that fits it best:

• Website: HTML builds the page structure, CSS handles colors and layout, and JavaScript makes the page "work": clicking buttons, opening the cart, submitting forms.
• Data: SQL is used to ask a database questions, such as "list customers who bought at least twice in the last three months."
• Analysis, automation and AI: Python, because it's easy to read and has a huge number of ready-made libraries.
• Phone apps: Swift for iPhone, Kotlin for Android.
• Drag-and-drop workflows: tools like n8n let you connect steps as blocks instead of writing code. That's programming too.
• Spreadsheets: Excel formulas. If you've ever written =SUMIF or =VLOOKUP, you've already programmed on this floor.

Languages on this floor are easier to write because the floors below do the hard part. Python actually runs on a program written in C. Your browser runs JavaScript with an engine written in C++. Every floor stands on the shoulders of the one below it.

Floor 5 · The rooftop: where you talk to AI in your own language

This is the newest floor. You don't write code, you say: "Build me an event sign-up page and save each guest's name and phone number to a spreadsheet."

AI translates that sentence into HTML, JavaScript and SQL on floor 4. That code is in turn translated down through floors 3 and 2, and finally becomes electric current flowing through transistors on floor 1.

Think of AI as a skilled general contractor: it listens to the owner in plain language, then directs the crews on every floor below. But even the best contractor needs the owner to be clear about which floor and which room to work on.

Summary table

Floor	In the building	In the computer	Languages	When you meet it
1	Ground	Hardware: chips, transistors	None yet, only electricity (1 and 0)	Choosing a computer or server
2	Foundation	Machine code	Assembly	Almost never
3	Frame, wiring, plumbing	Operating system	C, C++, Rust	Renting a server, installing software
4	Rooms	Websites, data, apps, automation	JavaScript, SQL, Python, Swift, Kotlin, Excel formulas	Building a website, reports, workflows
5	Rooftop	Talking to AI	Vietnamese, English	Every day

How does this map make you better at using AI?

1. Ask on the right floor

You message an AI: "My website is too slow, fix it." The AI might rewrite the whole interface, which takes effort and leaves the site just as slow. Meanwhile, the usual causes are somewhere else: images that are too heavy, slow database queries, or an underpowered server.

Someone who understands the building would ask:

"The home page takes six seconds to load. Check the image sizes, the database query times and the server configuration. Tell me where the bottleneck is before you change anything."

It's the same AI, but the second request points straight at the three floors to check, so the answer is in the right place from the start.

2. Give the right context

However capable an AI is, it can't guess what your system looks like. Add one line such as "The website runs on WordPress, on a Linux server, with data in MySQL," and the AI knows which building it's in, which floor, and which languages to use. The answers get much more accurate.

3. Notice when AI overdoes it or works on the wrong floor

AI sometimes proposes a bigger fix than the job needs: you want to change one button, and it suggests moving the whole platform. With the map in mind, you can see right away that "this job is inside one room; there's no need to knock down the frame."

Also remember: the lower the floor, the bigger the risk. A mistake in one room breaks one room. A mistake on the server or the operating system can affect the whole building. Work on the lower floors should have a backup and be reviewed by someone who knows what they're doing before it goes live.

4. Know which parts you own

Selling only on a marketplace or a rented platform is like renting a room in someone else's building. When they change the rules, raise fees or lock the door, you have to go along with it.

When you have your own website, your own customer data and your own workflows, you hold the keys to your own rooms. AI then works inside your system, the way you want, on data that belongs to you.

So should you learn to code?

You don't need to learn to code to become a developer. But it's worth spending a few hours understanding the map of your own system:

• Which floors your system has: what your website is built with, where it runs, and where your customer data lives.
• Who holds each floor: you, your staff, your web agency, or a rented platform.
• Roughly what AI just changed, and on which floor: you don't need to understand every line, only which room it touched.

A small tip: ask the AI itself. Paste in what you know about your system and ask: "Explain my website's system floor by floor, from the server to the interface, in plain language." AI is a very patient teacher for drawing this map.

Closing thought

AI has removed the hardest part of programming, which is remembering syntax. It hasn't removed the value of understanding the system. If anything, now that anyone can give AI instructions, the people who understand the building are the ones who point it at the right place.

Have you ever asked AI to fix something, only to have it fix the wrong thing? Tell LaPage about it.