All video posts

Learning to code · AI · Career change

Why Now Is the Best Time to Learn Coding, Because of AI

A few years ago I was an accounting and finance graduate scratching my head about where I belonged. Today I build startups and software with AI tools. If you are wondering whether AI means you have missed the train, my answer is the opposite: now is the best time to learn to code, not despite AI but because of it.

Deric YeeDeric Yee 16 November 2026 8 min read 19 min video

Prefer to read? The full article is below. Watch on YouTube.

The short version

  1. 1The problem in tech is a skills mismatch, not too few jobs. Many roles want coding and AI skills that too few people have.
  2. 2Coding is simply telling a computer what to do step by step, and it is becoming a new kind of literacy.
  3. 3Computational thinking, breaking problems into steps, is useful even if you never write code full-time.
  4. 4AI makes skilled programmers far more valuable. It does not understand your users, goals or system design.
  5. 5Learning by building, one project at a time with people holding you accountable, beats cramming theory.

Things feel shaky, and that is the point

Maybe you are 25 to 35 and early in your career, or you no longer enjoy what you studied at university and feel stuck. The stable jobs you signed up for are changing. Businesses are automating whatever they can, the AI wave keeps growing, and headlines talk about tech layoffs and economic uncertainty. It is natural to ask: is my job safe? Should I be learning something else?

Automation and AI are reshaping industries. Repetitive tasks and basic data crunching can now be handled by machines, and many roles are shifting towards higher-skill, human-centred work. That sounds intimidating, but it points to a key truth: the skills you build now matter more than ever.

Do not fall for the myth that you have missed the train. I hear it all the time: "AI will write all the code, so why learn it?" Plenty of non-technical commentators claim developers are automating themselves out of a job. The people actually building in this space are thinking the opposite. As AI tools improve, the return on knowing how to code grows, and AI helps you learn much faster too.

A talent gap, not a talent glut

Yes, the last few years brought layoffs and cooling salaries. But remember the world got used to the extraordinary growth of 2021 and 2022, when engineers were getting 30%, 40%, even 50% raises and salaries ran way above market. We are no longer in hyperdrive. Economies go up and down, and smart people look for the opportunity.

Companies in Singapore and across Southeast Asia are still innovating, adopting AI and expanding in areas like fintech, cybersecurity and digital services. They all need skilled people. The real problem is a mismatch: many jobs want digital and coding skills, and too few people have them. The World Economic Forum's Future of Jobs Report 2025 lists technology roles, from AI specialists to software developers, among the fastest-growing jobs in the world.

And people no longer learn to code only to become software developers. It is useful across a broad range of jobs. Learning to code moves you from passenger to driver of your own career. If you want to see what that looks like locally, have a look at the tech jobs market in Malaysia.

Coding from first principles

At its core, coding is telling a computer exactly what to do, step by step. Think of it as learning the language of machines. Once you speak it, you can build apps, automate tasks and analyse data. The world runs on software: food delivery, factories, financial systems, payroll, dashboards. If you know how to code, you hold the keys to that engine.

Humans bring the ideas and the problems; computers do the work based on instructions. Coding bridges that gap. You break an idea into logical steps, click this, then this happens, then send an email, and then you write the code. Without programming skills you are stuck outsourcing and hoping someone delivers. With them, you do not have to wait. All you need is a laptop and an internet connection.

That is why many people now describe coding as a new kind of literacy. Reading and writing went from an elite skill to a global standard that raised everyone's quality of life. Basic coding fluency is heading the same way if we want to understand, maintain and control the AI that will shape our future. If you are curious what code actually is under the hood, start with what coding actually is.

“If you want skills that remain valuable when machines are doing more, learn the skills to talk to those machines.”

Deric Yee, in the video

The real skill is computational thinking

At a granular level, coding teaches you computational thinking: breaking problems into chunks and solving them step by step. That applies even if you never write thousands of lines of code. Frontline workers and managers benefit from automating and optimising parts of their jobs, and any tech or data role needs these fundamentals.

It is no longer about memorising syntax or frameworks. It is about solving problems in a logical, systematic way. Coding is the ultimate multipurpose skill: durable, flexible and compounding. Each concept makes the next one easier, and soon you are not just coding one project but designing systems and working with data.

How I made the pivot

My journey began far from Silicon Valley. At university I studied accounting and finance, and my first taste of entrepreneurship was flipping graphing calculators: buying them cheaply and reselling them to students. It taught me supply and demand, margins and market fit, and that you can create value from scratch as a student.

After graduating I worked in corporate finance but felt out of place. As a student venture capitalist I had seen that tech sat at the heart of the biggest companies, and I wanted to build products, not crunch spreadsheets. So in 2019 I quit to teach myself to code full-time. I could not afford a bootcamp or a second degree, so I used free resources and built lots of projects, often from 9am to midnight. It was one of the toughest periods of my life.

About six to eight months in, something interesting happened. Friends with IT degrees were coming to me for help with technical interviews and back-end projects, and I was freelancing on software projects. A degree does not guarantee skill. What matters is what you do with what you learn.

What I learned from learning in public

In 2020 I started the Hacker Collective, weekly meetups where we learned by doing. Lawyers, accountants and engineers turned up. Many dropped off, because coding is tough and hard to sustain alone, but those who stuck it out built real things. Those years taught me a few lessons I still live by.

01

One project a month

I challenged myself to build one project every month. It forced me to turn theory into something real, whether a small web app or a script.

02

Teach to find your gaps

Teaching some sessions forced me to master the gaps in my own knowledge. When I am stuck, I still imagine explaining the concept to a friend who knows nothing about it.

03

Accountability keeps you going

Holding each other accountable was what kept people learning when it got hard. Alone, most people quit.

04

Proof of work beats credentials

People from finance, marketing and even construction could talk about projects they had built. You need real skills and examples of using them, not a degree or ten years of experience.

The elephant in the room: AI

With ChatGPT, Copilot, Claude and Cursor everywhere, it is tempting to think you do not need to code any more. Here is what people miss. AI may mean companies need fewer programmers for some work, which is a fair point, but it makes skilled programmers far more valuable. It also helps a forward-thinking junior close the gap with a senior engineer who is not using AI.

AI pioneers like Andrew Ng and Bill Gates have argued that now is the best time to learn to code. Andrew Ng's point is that as these tools make coding easier, more people can and should learn the language of software and make computers do exactly what they want. AI can generate code, but it does not understand your users, your project's goals or how to design a system from scratch. It is like a calculator with no numbers or context. Its output needs review and direction.

A Microsoft engineer I follow called replacing engineers with AI one of the most overhyped trends of the year: you can automate a toll booth, but not the creative problem solving of software engineering, and new roles open up to maintain AI systems. I unpack this further in will AI replace developers.

Pick one thing today

I will not sugarcoat it: learning to code takes effort. It is not a get-rich-quick scheme. But if you are reading this, you are curious, so the mindset is there. Treat it as an investment in yourself that compounds. Spend an hour a day and a year from now you will have hundreds of hours behind you, enough to be comfortably proficient. Even in marketing, operations or strategy, coding lets you automate your work or launch a side project.

I understand real life gets in the way: jobs, bills, families. That is why learning efficiently matters. Focus on hands-on practice and building projects, which give the biggest return. A few focused hours a week is enough to make progress. If you are juggling work, our guide to learning to code while working full-time can help.

So pick one thing right now. Write hello world in Python or JavaScript. Watch a five-minute video. Tomorrow, do a little more. Stay consistent, stack the skills and build something real. The easiest first step is our free trial, where you can start building with guidance. The best time to plant a tree was yesterday. The next best time is today.

Questions people ask

  • Is it too late to learn coding now that AI can write code?

    No. AI lowers the barrier to learning and multiplies what someone who understands code can do. It can generate code, but it does not understand your users, your goals or how to design a system, and its output needs review. People who know how software works and can direct AI tools are in a stronger position than before, not a weaker one.

  • Do I need a computer science degree to learn to code?

    No. I studied accounting and finance and taught myself using free resources and lots of projects. Within months, friends with IT degrees were asking me for help. What employers and clients care about is what you can build and explain. A structured path and accountability help a lot, but a degree is not a prerequisite for learning to code well.

  • What is computational thinking?

    Computational thinking is the habit of breaking a problem into small, logical steps and solving them in order, the way a computer needs to be instructed. Coding trains it, but it is useful far beyond programming: it helps in operations, marketing, finance and management whenever you need to automate a process, analyse data or design a clear solution.

  • How much time do I need to learn to code while working?

    A few focused hours a week is enough to make real progress, and an hour a day adds up to hundreds of hours within a year. The key is spending that time on hands-on practice and building projects rather than passively watching lessons. Consistency matters more than intensity, so pick a schedule you can actually sustain.

Your next step

Our 12-week AI-Native Software Development Programme is a live, mentor-reviewed cohort where you build and ship real products with AI, and learn to explain every line. Start with the free trial to see how we teach, then decide.

Watch next

All videos

Learn the fundamentals that pay off in any direction.
Coding plus AI, taught properly.

Whatever you end up doing, the people who understand how software and AI actually work get to direct them. Try the first lessons free, no card needed, or see the full 12-week AI-Native Software Development Programme.