Stop learning only programming.

Recently an article from Nvidia's CEO Jensen Huang has attracted a lot of heat and attention generating passionate debates on social media platforms.

The following line came bolded in several articles: stop learning programming. Why? Because AI will do all the programming in the future.

You just tell it what to do. You don't need to know how to code anymore. With the help of AI everybody can create applications. True? False? As always: details, points of view and nuances count mostly.

Let me take a step back and visualize the bigger picture.

From a company known for decades for creating a niche product (GPUs) by a lucky ricochet it has become somehow overnight a super star,

The Key Player in the AI industry. Without producing AI. It produces the hardware support needed for a niche within the whole AI ecosystem. These are the facts.

Of course Nvidia is massively benefiting from AI. Similarly true is also the fact that "classical" software development will not help Nvidia, on the contrary.

Therefore the reason Nvidia is advocating the massive use of Gen AI tools for the future is quite obvious, hence the aggressive stop learning programming call out.

Now, I would like to give my personal feedback and suggest another slogan: stop learning only programming.

I strongly believe that focusing solely on programming while ignoring the domain knowledge will become obsolete. Sooner than later.

However not knowing the fundamentals of software development and architecture will not get you too far creating valuable software neither. Not in short or mid future.

Sounds complicated and challenging? It is. We, at Elysian Software love complicated challenges :)

Software development has had multiple evolutionary stages or periods. It started with assembly, where developers had to focus on bits and bytes. Important progress has been made with Fortran, Cobol and luckily C. Less bits and bytes, more human readable code. Languages, frameworks have evolved and so the "helping" tools around them. Welcome the era of IDEs in the early 2000s. Frameworks and libraries got also better, richer in functionalities more and more ready-to-use code was provided by default. Development has shifted more to integrating APIs and libraries than to writing algorithms.

All this with 2 purposes: create apps quicker/cheaper and give people more time for creating real value products by engineering genuine ideas instead of "translating" these ideas in machine readable formats.

But to create valuable products, you need to thoroughly understand the domain knowledge.

I'd like to shortly give as example an industry where we, Elysian Software are present since > 10 years: the intralogistics.

To be an efficient software developer in intralogistics you need to master advanced and deep concepts like : the importance of a temporary location buffer between a shuttle and a lifter in a cubical storage system aimed to increase the performance of retrieval in a sequenced manner.

Or when to distinguish between the need of announcing the hardware with a few seconds in advance a or just in time millisecond communication.

Or how to create a storage algorithm that is in the same time balancing the stock distribution, balancing the workload, minimizing the moves and increasing the throughput.

Sounds complicated and challenging?

It is.

We don't just do software development. We create solutions based on understanding your domain challenges and bringing that expertise right from the start to avoid subtle mistakes that cost a lot.

That's why it is important to stop learning only programming.