ChatGPT is becoming increasingly powerful
A year ago, I was only using ChatGPT to search for some simple example codes and to add comments line by line to my own code, with comments so precise that I was amazed.
Now, I can simply describe the desired effect to ChatGPT, or even just send a picture, and ChatGPT can directly generate MATLAB code.
For example, if I see a beautifully styled plot, I can just take a screenshot and send it to ChatGPT, which will generate the code to perfectly reproduce it.
If I see a Simulink control logic diagram in a document and want to build a similar one, I can just send a screenshot to ChatGPT, and it will directly generate the code to create it.
ChatGPT is truly astonishing, making me start to doubt whether MATLAB will be eliminated. This has also been one of the most frequently asked questions I have received in recent years, as more and more people begin to question the value of traditional engineering tools.
For this question, it mainly depends on what you are using MATLAB for.
If it is for data processing with low complexity, data visualization, or parameter fitting, AI might solve it in just a few sentences. You could say MATLAB will be eliminated, and similarly, you could say Python will be eliminated, or Excel will be eliminated.
However, MATLAB has never been positioned as a tool for writing code.MATLAB is aninteractive platform for computation, modeling, and simulation aimed at engineering problems, which is irreplaceable in fields such as Simulink, control systems, and communications.
Moreover,MATLAB’s advantages in reliability are something that AI tools like ChatGPT cannot match.
Imagine, as an OEM customer, if a supplier tells you that the delivered controller code was entirely designed and verified using ChatGPT, would your first reaction be to think it’s cool or to be speechless? A tool that occasionally makes mistakes even in simple comparisons like 0.11 and 0.8, how could you trust it for model verification and code generation tasks that require high reliability?
In fields like automotive and aerospace, where “errors in engineering = loss of life”, no matter how smart ChatGPT is, we still rely onMATLAB/Simulink. This is why enterprise-level certification and process standards still require the use of MATLAB/Simulink.
When it comes to the question of using AI versus using MATLAB, why must we think in black and white? We are not children.
AI can serve as a supportive tool for MATLAB. We no longer need to spend two or three hours debugging a trivial error, nor do we have to repeatedly flip through documentation to find parameter spellings. We can even convert a segment of natural language requirements directly into scripts, treating AI as part of our toolchain to enhance efficiency and lower barriers, allowing us to focus more on the engineering problems themselves.
In fact, MathWorks is doing just that, introducing “code suggestions and auto-completion” in MATLAB, starting to support more AI toolboxes (Deep Learning Toolbox, Reinforcement Learning Toolbox), implementing “AI process modeling” in Simulink, and soon integratingCopilot into MATLAB.
If we are to talk about elimination, then in the future, MATLAB users who do not use AI will definitely be eliminated. If you do not want to be one of them, I recommend attending MathWorks’ official seminar “AI-Driven Innovation: The Integration of MATLAB and Large Language Models“. From practical cases to technical integration, this seminar will showcase what the future of MATLAB + AI should look like.
Interested friends are welcome to scan the QR code below to register.
