Can English Dethrone Python as Top Programming Language?
The hottest programming language of 2024 was the English language.
That’s right, plain old English, or let’s say natural language, emerged as the top language last year and continues to dominate as the lingua franca of generative AI (GenAI), some experts say.
Indeed, Andrej Karpathy, a co-founder of OpenAI who has now started a new AI+Education company called Eureka Labs, called this out in 2023.
The hottest new programming language is English
— Andrej Karpathy (@karpathy) January 24, 2023
Brad Shimmin, an analyst at Omdia, told The New Stack Karpathy is “spot on” with his claim.
“The biggest programming language of the year has to be natural, spoken human language with GenAI code completion and even full-stack development tools like Aider and Cline enabling developers to use, let’s say English, as a declarative programming language in its own right,” Shimmin said.
He added that he believes the entire IT industry has been building toward this idea for decades, creating higher and higher levels of abstraction, enabling developers to focus more on “what” kind of task they want a program to execute and less on “how” that program should go about executing that task.
In fact, Shimmin told The New Stack he believes the shift to natural language programming interfaces represents a significant evolution in software development, comparable to the introduction of compilers.
Meanwhile, the use of English as a programming language has enabled Karpathy to launch a new type of software development that he has coined “vibe coding” (see related article). Vibe coding is a high-level approach to coding using AI where users describe requirements from an end-user perspective rather than technical specifications.
Some of the major players in the GenAI space enabling English for programming include Microsoft, OpenAI, Anthropic, Google, IBM and AWS among others, Shimmin noted. They are developing models with improved tool use and structured outputs. And some of the key development platforms mentioned include GitHub Copilot with VS Code, Replit (which was an early adopter of AI integration), Aider, Cline, Cursor and Zed.
Natural Language (English) as the Universal Programming Language
“I certainly don’t think English was the most important programming language of 2024, but it is gaining ascendancy,” Arnal Dayaratna, an analyst at IDC, told The New Stack.
IDC predicts that by 2028, natural language will become the most widely used programming language, with developers using it to create 70% of net-new digital solutions. (Source: IDC FutureScape: Worldwide Developer and DevOps 2025 Predictions)
“I actually think that the best phrasing of this prediction would be to replace ‘natural language’ with ‘English’ because of the dominance of English as a spoken and written language worldwide,” Dayaratna said.
Moreover, he said he believes that in four to five years, developers will increasingly go to a chatbot-like interface and use natural language to produce digital solutions. Meanwhile, code will be used to innovate on the technology substrate that enables this kind of technology.
“In other words, we’re not far from a world that witnesses the demise of commercial off-the-shelf software simply because it will be so easy to create such software, in a custom way, for an organization’s business processes,” Dayaratna said.
Hence, he explained that we are seeing the emergence of what Amjad Masad, CEO of Replit, called the era of “personal software.”
“Just as the Mac inaugurated personal computing in 1984, generative AI has initiated the era of ‘personal software’ that recognizes the specificity of individual and organizational preferences,” Dayaratna said.
For his part, Masad told The New Stack that it is “totally true” that English is currently the top programming language.
“We have more customers now that build with Replit Agent using English than we have customers coding in JavaScript or Python,” he said.
Microsoft’s Mike Hulme, who is general manager of Azure Digital Apps and Innovation, weighed in with the software giant’s view, noting that “AI is giving us the ability to program completely in natural language, helping every developer code faster and more accurately while bringing together new streams of developer talent globally. By using natural language as the common model for coding, we can overcome programming skills barriers, understand and maintain existing applications more easily and build new AI apps in a way that is more accessible and efficient for everyone.”
Developers Will Still (and Always) Write Code
Programming languages remain necessary for precise operations, Sriram Devanathan, general manager of Amazon Q Apps and AWS App Studio, told The New Stack. “New programming languages may emerge at higher abstraction levels. Programming languages won’t disappear, but learning methods will evolve,” he said.
For his part, Ameya Deshmukh, head of marketing programs at Tabnine, told The New Stack that it’s not surprising that OpenAI’s founders would see English as the “biggest programming language” due to the volume of code-related prompts being inputted into ChatGPT.
“However, in our experience, contrasting enterprise-grade AI code assistants against standalone LLMs reveals a stark difference: enterprise AI code assistants can accomplish in seconds — with a few clicks and concise prompts — what standalone LLMs often require hundreds of words to achieve,” he said.
Yet, mature engineering teams are still writing significantly more lines of code than natural language prompts, he said. “AI code assistants designed for enterprise use enhance workflows by integrating seamlessly into existing processes, making code creation faster and more efficient while keeping engineering teams in control,” Deshmukh explained.
Enterprise Use of Natural Language and GenAI
Enterprise software vendors like Pegasystems are embracing GenAI and agentic technology.
Don Schuerman, CTO at Pegasystems, told The New Stack that GenAI has made natural language a powerful starting point for developing enterprise applications. This approach dramatically accelerates the speed at which organizations can go from an idea to a running application, reducing what once took weeks into minutes.
“For example, tens of thousands of users engaged with Pega GenAI Blueprint to design workflow apps using just natural language, demonstrating how plain English has evolved from just describing requirements to actively shaping application design,” he said.
But app dev success requires more than just turning English into code. Multiple stakeholders can interpret the same language differently, and more importantly, enterprises need their apps to be maintainable and scalable over time, Schuerman said.
Is Low-Code Dead?
“That’s why I believe the intersection of natural language prompts and visual low-code approaches creates such powerful results,” Schuerman added. “When business users can express their needs in natural language and immediately see that translated into visual business metaphors — workflow diagrams, case lifecycles, and sample user screens — they not only get started quickly but can also ensure every stakeholder sees and understands the same solution.”
This approach preserves the speed and accessibility of natural language while providing the structure and governance that enterprises need for long-term success, he added.
However, Omdia’s Shimmin disagrees with part of that.
Traditional low-code/no-code tools may be becoming less relevant, he said
“I feel like they’ve sort of run their course. I mean, I think there’s always… tools you can use, like a rules engine… that would be very useful. But you know this idea that low code as a market on its own? I feel like that’s not what we’re really looking at anymore. Gone are the days where low code has a market on its own…”
Democratizing Development
Like low-code/no-code, GenAI enables people with little or no technical training or experience to build applications, thus expanding the number of people capable of creating applications.
Amazon’s Devanathan said Amazon’s GenAI offerings enable non-traditional developers to create applications, junior developers to ramp up more quickly, and expert developers to focus on complex problems, as AI handles the routine, boilerplate work of programming. And different tools will emerge to support various user types, he said.
“What changes is that engineers spend less time on the undifferentiated heavy lifting on the boilerplate, and they spend more time on things that are why engineers usually signed up to become developers in the first place,” Devanathan said. “So, they get to spend more time on the really interesting, hard things.”
English Is Still Not a Programming Language! It’s Not the Same Thing
Generative AI is amazing at understanding natural language queries and returning amazing responses, said Eric Newcomer, an analyst at Intellyx.
“And I understand that some are saying GenAI means you can use English — or other human language — as a programming language,” he told The New Stack. “I can see why people say it, because GenAI can produce computer code in response to English language prompts.
“But saying this makes English a programming language is not correct. Programming languages exist because English isn’t precise enough to be a programming language. Programming languages compile down to executable binary code. An LLM transforms English into vectors that GenAI evaluates using statistical probabilities.”
However, “It’s not the same thing at all. English might help you create a program, but it isn’t executable.”
Given its role in GenAI tools, it’s fair to consider English as a de facto language of sorts, said David Mytton, CEO and founder of Arcjet.
“However, I think the lack of precision is a key issue. Unlike actual programming languages, which are explicit and deterministic, English is open to interpretation,” he said. “One of the pitfalls of using plain English as a programming language. You’ll never be 100% sure what you’ll get. Whether that matters really depends on the app you’re building.”
Admittedly, English as a programming language faces some challenges, including language ambiguity and interpretation issues, a need for ontology management in enterprise settings, and a lack of standardization across vendors, Shimmin said.
“Each vendor — Salesforce, SAP, etc. — has their own approach to these challenges,” he said.
YOUTUBE.COM/THENEWSTACK
Tech moves fast, don’t miss an episode. Subscribe to our YouTube
channel to stream all our podcasts, interviews, demos, and more.
link
