Amid the generative artificial intelligence revolution, the debate about the future of programming and the role of programmers has intensified. While influential voices like Jensen Huang, CEO of NVIDIA, claim that the developer profession will disappear in a few years due to the simplification that AI will bring to the software creation process, other experts disagree.
One of them is Andrew Ng, founder of DeepLearning.ai, who has been adamant: “AI is not going to eliminate programmers. In fact, there has never been a better time to learn to program.” According to Ng, the ease with which we can now create code—thanks to AI-assisted tools—should be seen as an invitation for more people to enter the world of programming, not as a reason to abandon it.
The Natural Evolution of Programming
Throughout history, programming has evolved from punch cards to the intelligent development environments we know today. Each step, from assembly language to C, from desktop to the cloud, or from text editors to IDEs with autocompletion and automatic debugging, has lowered entry barriers and enhanced creative capacity. The arrival of assisted AI is simply the next stage.
Ng points out that those who understand how to speak the “language of software” and coordinate multiple AI tools can become what he calls 10x professionals: individuals capable of having ten times the impact of the average professional. In an environment where more and more tools allow for the automation of routine tasks, the key will be knowing how to guide the AI, not competing against it.
A New Professional Profile: From Traditional Programmer to AI “Orchestrator”
It’s no longer just about writing lines of code; it’s about designing processes and establishing clear instructions for the AI to produce optimal results. This is what some are starting to call vibe coding: letting the system generate code based on guidelines and then reviewing the quality and accuracy.
In professions like medicine, law, or marketing, those who know how to use this “new language” will be able to obtain tailored solutions without relying on third parties. The impact of this capability will be enormous.
The Mistake of Excluding Juniors
However, it’s not all positive. A dangerous idea is starting to spread in various tech forums: that of not hiring junior developers on the grounds that they only know how to program with ChatGPT and do not understand the fundamentals. This mentality is, at the very least, a strategic error.
The demand for programmers remains high, and thinking that the workforce can only be fed by senior profiles is unviable. Juniors not only bring fresh talent but are also key to generational succession and team evolution. However, companies must ensure that these young developers have a solid foundation and do not rely solely on AI to write code without understanding what they are doing.
The Competition Grows: Increasingly Accessible AI Models
The race to improve language models continues. The most recent case is QwQ-32B, a model with 32 billion parameters developed by Alibaba that competes with reasoning capabilities against giants like DeepSeek-R1, which uses 671 billion parameters. The optimization of these models will allow developers, startups, and medium-sized companies to access advanced tools without relying on unreachable infrastructures.
On the other hand, Microsoft has introduced Phi-4 Multimodal, a model capable of processing text, images, and voice simultaneously, opening new possibilities for creating hybrid applications without the need for advanced programming knowledge.
Conclusion: AI Transforms, But Does Not Eliminate
The consensus among the most sensible voices in the industry is clear: AI will not replace programmers, but will transform them into guides, supervisors, and orchestrators of much more complex processes. The ability to define problems, structure solutions, and communicate precisely with machines will be the true differentiating value.
In a world where tools evolve at a breakneck pace, programming remains the universal language of innovation. And this is, without a doubt, the best time to learn it.
Source: Programming News