Photo supplied.Coding academies and robotics clubs flourished on the promise that technical skills would define the next generation’s careers. Now, AI tools can generate software in seconds. Headlines about mass layoffs at tech giants have parents wondering if that promise still holds.
There is still a strong case for learning to code. It teaches young people to break a problem into steps, test an approach and work out why something has gone wrong. These skills are equally important when AI generates the code, because someone still needs to assess the result.
The global employment picture offers few simple answers. As at 2 November 2026, TrueUp had tracked close to 200,000 people affected by global tech company layoffs this year, while listing more than 280,000 open roles across the tech companies and startups it monitors. Current vacancies and job cuts over a year measure different things. Neither figure, on its own, tells a parent what their child should learn.
A more useful starting point is to ask what young people need to understand when increasingly powerful tools can do some of the work for them.
South Africa's skills gap
In South Africa, the challenge extends beyond choosing which technology to learn. Young people also need opportunities to apply their knowledge.
Collective X’s 2025 research identified about 118,500 ICT vacancies nationally, including more than 40,000 junior opportunities. Yet graduates still struggle to secure their first role. Its analysis points to a gap between qualifications and workplace readiness, highlighting practical experience and the confidence to apply technical knowledge.
For parents and educators, that strengthens the case for learning through projects: giving young people a problem to solve, room to experiment and feedback that helps them improve.
Why learning to code still matters
A poll of Investec Developer community members offers a practical starting point. Asked what a Grade 10 learner should focus on first for a career in tech, 65% of the 204 respondents chose problem-solving and critical thinking. Coding fundamentals followed at 26%, ahead of AI literacy at 8%.
Investec Developer brings together more than 3,500 developers across over 850 companies, including software engineers, founders and technology leaders building with Investec’s APIs and programmable banking capabilities.
One respondent captured the thinking: “Software developers are problem solvers first and foremost and just happen to use code as their tool for solving problems.”
The poll reflects participating members’ views on where to start. It does not make coding fundamentals optional. Learning to code can put critical thinking into practice: define a problem, break it into manageable parts, test a solution and explain why it works.
Consider a learner building a simple budgeting app. AI might generate the code, but the learner still needs to decide what the app should do, check its calculations and consider how it handles personal information. A working screen is only the beginning.
The message for parents and educators is to teach young people to think, then give them increasingly powerful tools to act. Coding can help them develop both capabilities.
What recruiters and specialists watch for
Sanisha Naidoo, who leads Investec’s early careers and graduate programme, still sees a degree as a strong foundation for a first technical role, although results alone no longer decide who gets hired.
“One skill students often overlook is the ability to slow down and think critically,” she says.
Investec’s cybersecurity specialists show why that thinking needs to be accompanied by communication. Awareness specialist Ndivhuwo Mulaudzi’s work involves “engaging with highly technical specialists while also supporting people who may not be comfortable with technology”. Information security architect Javan Mnjama similarly lists “presentation and writing skills alongside technical ability” as central to his work.
Penetration tester Nsovo Baloyi highlights “curiosity, analytical thinking and problem-solving”. Together, their perspectives suggest that preparation for a technical career should include opportunities to investigate problems, explain findings and work with others.
From Promaths to machine learning
Dineo Ramokhothoane, a machine learning engineer in financial services, offers an example of where strong technical foundations can lead.
At school, she joined Promaths, a partnership between Investec and the Kutlwanong Centre for Maths, Science and Technology that supports high-potential learners in mathematics and physical science.
For Ramokhothoane, AI’s ability to generate code makes understanding the underlying ideas more valuable.
“Right now, we are in the dawn of the AI era and it can do a lot of coding for you. But what sets you apart is understanding the fundamentals and the basics of where the algorithms and systems come from.”
Her perspective brings the argument back to understanding. Access to a tool is useful. Knowing enough to question its output gives a young person a stronger basis for using it well.
What this means for parents, learners and educators
For parents, the practical question is what a coding course asks children to do. Do they simply follow instructions, or do they solve problems, test ideas and explain their decisions? When they use AI, can they describe what it produced and how they checked it?
Encourage small projects built around a real need, with room to make mistakes and improve. Ask learners to explain their approach and what they would change next time. This brings technical knowledge, curiosity and communication together.
No course can guarantee a future job. A sound technical foundation, practical experience and a willingness to keep learning can, however, give young people more ways to respond as work changes.
Invest-ED, designed for Grade 8 to 12 learners and their parents, offers career perspectives and practical guidance to help families explore those choices.
Should young people still learn to code? Yes. Learning to turn an idea into something that works, understand why it works and recognise when it fails remains a valuable education, even when AI helps write it. DM