Nvidia Says AI Will Kill Coding, but Here Are 5 Reasons It’s Wrong

The CEO of Nvidia has said that we shouldn’t educate children to code since artificial intelligence will do it all. True, AI may be a coder’s best friend, but there are a lot of reasons why individuals should still learn to programme, even if they wind up utilising a lot of AI-generated code.

Who Speaks for Nvidia When Discussing AI and Programming?

Nvidia CEO Jensen Huang said at the 2024 World Government Summit in Dubai that he doesn’t see the need in educating people to code since artificial intelligence would soon be able to perform all the programming:

Our mission is to develop computer systems that eliminate the need for programmers. Additionally, the programming language is based on human characteristics. These days, it seems like everyone is a coder. A.I. has performed a miracle with this.

Even though AI can produce code snippets, others claim that learning the fundamentals of programming is unnecessary for anybody interested in making their own software.

Is AI Skilled at Programming?

We need to examine what led Jensen to make this remark before we can analyse it.

You may find online AI models that can assist you with your programming needs. While general-purpose models like ChatGPT are capable of answering programming questions, you also have the option to employ a specialised model like CodeGPT.

An explanation of the CodeGPT code in the chat windowPictures by Denis Kuria/MakeUseOf
These AI models are quite good at translating commands into code. The fact that AI code generation does not need programming knowledge is its primary strength, as Jensen mentioned before. If everything goes according to plan, all you have to do is type in your instructions in English and copy and paste the results.

The Argument Against Nvidia’s Policy of Not Encouraging Young People to Learn Computer Programming on LaptopsShutterstock user Gorodenkoff

The argument put out by Huang is well-reasoned and supported by facts. But there are a lot of good reasons to keep educating people to code.

One, AI Code Still Needs Human Expertise to Verify Its Correctness

While AI code may not always fail to execute, it does so in most cases. Acquiring coding skills equips you to improve and polish the AI’s output code. In our previous article on CodeGPT, we mentioned:

Therefore, CodeGPT can create code; but, it is not foolproof and may provide subpar results at times. The end result may not adhere to recommended practices and may include errors. Before you start writing code with CodeGPT or any other AI-powered tool, be sure you fully grasp what it does. Paste is not the same as copying.

Problems aside, there’s also the matter of making the code human-readable and checking that it adheres to fundamental programming concepts that all programmers should be familiar with. Blindly adding AI-generated code may lead to a tangled web of problems or even security holes in your programme, website, or other project.

2. Problem-solving by programmers with experience outperforms AI.

At their core, programmers are problem solvers. They need to figure out the best approach to use code to make their ideas a reality, whether it’s for themselves or a customer.

As they build more and more applications, programmers have a clearer idea of the process from concept to finished product. Chatbots powered by artificial intelligence lack the kind of training that would allow them to improve their performance based on lessons gained from prior tasks.

3. Code Can Be Better Adapted to User Needs by Programmers

Making a programme for another person almost always means they will have suggestions and criticisms. As long as the customer receives what they want, it’s an inevitable aspect of the feedback loop.

It would be far easier for a person to build the app since they can take the client’s suggestions into account while creating new versions and keep iterating until it meets their needs. But if you’re stuck with AI-generated code, your only option is to ask the AI to fix it and cross your fingers that it understands what you’re asking (and doesn’t mess up the code, rewrite parts that aren’t needed, introduce vulnerabilities, etc.).

4. Programmers Have the Abilities to Adjust to Evolving Technology Artificial intelligence

from existing data at a faster rate than humans. That makes them excellent at duplicating content from the internet but terrible at keeping up with current trends and technology.

It is always possible for programmers to stay abreast of the newest advancements in their language of choice and incorporate them into their code. They are also adept at switching languages when one is no longer preferred and know which ones perform best for certain jobs.

5. Compared to AI, programmers can respond much faster to urgent patch requests.

By the end of the day, fixing bugs in the code demands expertise in computer programming. When dealing with a zero-day exploit, finding the issue and fixing it as fast as feasible is of the utmost importance.

Code produced by humans or by artificial intelligence (AI) is equally prone to errors. Someone with coding knowledge, nevertheless, will have a simpler time tracking down the issue and comprehending it. No one can possibly instruct an AI to do anything without first understanding how code works.

The remark made by Jensen Huang was not incorrect. With the help of AI, even those without any programming experience may write and run their own programmes. However, fundamental programming expertise is required to guarantee the efficient and effective operation of whatever one wishes to construct, whether for oneself or another.