In his latest article How to Become a Multilingual Coder, AI pioneer Andrew Ng reflects on how artificial intelligence tools are reshaping the programming landscape – and what that means for developers everywhere.
Rather than being tied to just one language, coders are increasingly able to work across multiple languages, thanks to the help of AI. This opens up new opportunities but also requires a renewed focus on the fundamental concepts of programming.
The traditional model: mastering one language
For decades, the standard advice to new developers was simple: pick a language and get really good at it. Whether it’s Java, C++, Python, or JavaScript, success often meant specialising deeply. Companies would look for “Python developers” or “JavaScript engineers,” and coding interviews would probe the details of language syntax and quirks.
This model made sense. Every language has its own rules, libraries, and patterns. Expertise took years to build and maintain. Moving between languages wasn’t easy; it often meant starting again from scratch.
Ng, for example, describes his own long-standing comfort with Python – a language he has used for countless machine learning and AI projects. For many years, there was little incentive to stray beyond his expertise.
How AI changes the rules
However, the rise of AI coding assistants like GitHub Copilot, ChatGPT, and other large language model (LLM)-based tools is altering the role of the developer.
Instead of painstakingly learning every aspect of a new language, developers can now use AI to help them write working code quickly, even in unfamiliar languages. Ng describes how, with the help of an AI assistant, he was recently able to write a front-end web application using JavaScript – a language he had limited hands-on experience with.
He didn’t need to memorise syntax or spend hours digging through documentation. Instead, he focused on describing the functionality he wanted and letting AI suggest solutions, which he then reviewed and adjusted. This suggests that in the near future, coding fluency won’t be about remembering syntax; it will be about understanding logic, structure, and architecture.
In short, AI is lowering the “activation energy” needed to pick up a new coding language.
Language specialist to problem solver
So, the best developers will potentially no longer be defined by their mastery of a specific language. Instead, they’ll be valued for their ability to think logically, solve problems, design robust systems – and adapt to whatever tools are needed for the job.
For instance, imagine a backend engineer being asked to contribute to a mobile app written in Swift, or a data scientist needing to tweak a TypeScript web dashboard. In the past, these tasks might have been challenging. Today, with AI assistance and a solid grasp of programming fundamentals, they could become very manageable rather than insurmountable barriers.
This shift also benefits organisations, making it easier to build flexible, cross-functional teams where engineers aren’t boxed in by language labels.
Fundamentals are still important
Of course, AI doesn’t remove the need to actually understand what you are doing. As Ng points out, a good developer still needs a strong grasp of core programming concepts:
- Data structures like arrays, lists, hashmaps, and trees
- Algorithms for searching, sorting, and optimising
- Memory management and performance considerations
- Software design patterns and architectural thinking
Without this knowledge, you won’t be able to evaluate whether AI-generated code is efficient, secure, or maintainable. You also won’t be able to guide AI effectively – because you won’t know what questions to ask or what structures to aim for.
Ng gives the example of front-end development: If you don’t understand how the Document Object Model (DOM) works, AI can certainly help you generate React components – but you won’t easily diagnose problems when things break. Similarly, if you don’t understand how GPUs handle parallel processing, you won’t be able to write efficient CUDA code, even with AI helping you draft it.
In other words: AI can accelerate your coding, but it can’t replace deep understanding.
Learning with AI
Rather than fearing AI, Ng encourages developers to embrace it as a learning partner.
Today, he says you can:
- Experiment with writing small programs in new languages, using AI to guide you
- Ask AI to explain unfamiliar code snippets and frameworks
- Explore alternative programming paradigms (functional, object-oriented, declarative) without investing months into manual study
This approach could transform learning from a solitary, painstaking process into an interactive, dynamic experience. It also means that developers can be much more responsive to the needs of a project or employer. If a new project calls for Rust, Go, or Kotlin, you don’t have to start from zero – you can bootstrap your knowledge quickly with AI’s help.
However, it’s important to balance AI coding exploration with depth. Moving too fast across languages without mastering any can leave you with a shaky foundation. Ng’s advice is clear: treat AI as an assistant, not a crutch.
Inclusion, innovation, and creativity
This new way of coding has broader social and economic implications.
Accessibility:
People from non-traditional backgrounds – those without computer science degrees, or those learning coding later in life – can enter tech more easily. AI helps lower the initial learning barriers that once made programming feel exclusive and intimidating.
Creativity:
When developers are freed from technical drudgery, they can focus more on creative problem-solving, designing better user experiences, and innovating faster.
Global collaboration:
AI-assisted multilingual coding means developers around the world, speaking different natural and programming languages, can collaborate more easily. This could democratise software development even further. Ng’s reflections remind us that we are entering a new era. AI is not just writing code – it could be reshaping what it means to be a developer. Success will increasingly depend not on memorising language syntax, but on:
- Deep understanding of core principles
- The ability to think clearly and architect solutions
- Flexibility in working across technologies
- A willingness to keep learning and adapting
By embracing AI as a collaborator and focusing on foundational skills, today’s developers can thrive in a world where the boundaries between programming languages – and between human and machine – are becoming more fluid than ever.
This article is inspired by key ideas from Andrew Ng’s piece “How to Become a Multilingual Coder,” published in The Batch by DeepLearning.AI on 23 April 2025.