To Code or Not to Code?

The distinction between “code” and “no-code” is a deception. Instructing a machine, regardless of the language, is an act of creation. Reclaiming this power means shaping our professions and increasing our capacity to act.

The Paradox of Lost Autonomy

In all professional sectors, even those furthest from computing, the use of digital tools has been second nature for nearly thirty years. Alongside standard office software – word processors, spreadsheets, messaging systems – specialized “business” tools designed for specific tasks are deployed. The importance of these latter tools is crucial: they don’t merely offer productivity gains; by opening up new capabilities, they reshape professions in depth. Immense “consulting firms” have built their wealth on this essential market. A manual task, once automated, frees up time and mental energy, allowing a structure – for example, a cultural organization – to imagine new missions, innovate, and achieve its objectives more precisely.

This autonomy through business tools can also be developed “in-house.” Since the 1990s, software like 4D, Access, or FileMaker Pro have made complex database management, which generates productive actions, accessible to “ordinary people.” Non-IT professionals could then develop their own solutions to organize events, accounting, or a production chain, for example, while connecting with the outside world. The main advantage lay in these people’s ability to evolve their tools autonomously, according to their needs, without the extra cost and inertia associated with hiring an external provider for each adjustment. Today, this movement is called “no-code,” with platforms like Airtable or automation tools like N8N, which often leverage artificial intelligence for tasks like social media posting or managing automated responses to emails or social media messages, for example.

Yet, a paradox has struck me for several years: while tools have never been more powerful and accessible, I observe a relinquishing of this autonomy. The most glaring example is the omnipresence of Excel for managing databases, a task for which it is notoriously inefficient. In the cultural domain, it’s not uncommon to see young professionals, twenty years after their elders, showing themselves to be much less effective and creative with computing, producing confused, very disorganized tools, causing much loss of time, energy, and efficiency.

We are witnessing a surprising form of hyperspecialization in our era, which nevertheless enables the democratization of knowledge and multidisciplinarity. A “no-code culture” seems to have taken hold, as if touching the slightest line of code were an impure act, reserved either for an inaccessible elite or for interchangeable drudges, depending on one’s representation of computer scientists.

Coding is Speaking: A Story of Languages and Abstractions

Behind this rejection lies well-oiled marketing, that of “no-code,” which aims to make programming accessible by making it disappear. The intention is laudable: to remove complexity for those who don’t feel legitimate in “coding.” But this approach is based on a false idea of what code is, and unfortunately leads us to miss the constructions we could invest in to improve our tools ourselves. It’s rather an invitation to incompetence. Because in reality, we are all, to varying degrees, computer scientists. When we dialogue with an artificial intelligence like ChatGPT, refining our requests as we go, our prompts, to co-construct a relevant response, we are doing nothing other than programming. We simply use natural language as a coding interface.

Computer code, indeed, is almost never a direct dialogue with the machine. Machine language, that binary flow of billions of zeros and ones that runs through electronic circuits, is a language that virtually no programmer practices. And an infinitesimal fraction of them master assembly language, the abstraction layer just above, which is the raw language of microprocessors. Each type of processor has its own, making this knowledge profoundly linked to the materiality of the machine.

The vast majority of code we know is written in high-level abstraction languages like C, Python, JavaScript, or Rust, for example. These languages are precisely designed to be more comprehensible to humans and independent of hardware platforms. They function as lingua franca that allow giving instructions that can be “translated” for different machines. In this sense, the difference between writing a formula in a spreadsheet, arranging visual blocks in Scratch, drafting a query in natural language, or typing a line of Python is not a difference in nature, but in degree of abstraction. Asking a machine to execute a task algorithmically, whatever programming language we use, is always programming, “code.”

Toward Hybrid Mastery: AI as a Bridge to Code

This fear of code is all the more irrational as boundaries are blurring. Programmers themselves make intensive use of generative artificial intelligences to automate repetitive tasks, explore solutions, debug, or simply accelerate production by avoiding documentation time. Code is increasingly co-created by AI, with the human role evolving toward that of an architect, a conductor who controls, validates, and integrates these fragments of logic. The question is therefore no longer “to code or not to code,” but rather “what language to learn to dialogue with the machine?” Whether it’s a visual language, the natural language of “prompt art,” or a scripting language, it’s always linguistic learning.

Tools like FileMaker Pro, which I continue to recommend after more than thirty years of existence, perfectly embody this hybrid philosophy. It allows building robust business applications through visual manipulations, by connecting objects and functions. But if one wishes to go further, it offers an integrated scripting language, for which the software itself provides remarkable support. Code is no longer presented as a brutal wall reserved for specialists, but as an accessible horizon, a new room one can add to their toolbox. The help system is so well designed that we are guided step by step in creating this code.

Generative AIs are formidable accelerators for this learning. Just as they are excellent tutors for learning human languages by adapting to our precise needs, they are for computer languages. I can describe a need to them in French, ask them to translate it into code, adapt it to my needs with their help, then copy-paste this code. By dialoguing with the AI about why this or that instruction exists, I understand, learn, and assimilate. It’s embodied learning, lived, directly related to concrete construction, as if we had a specialist at our disposal. Information anchors in memory because it immediately serves a project.

The Tool is the Power to Act

I therefore believe it is absolutely essential to invest in manufacturing our own business tools, so they can evolve with us, follow us, sometimes even precede us by opening possibilities. We cannot wait for others to do it in our place, otherwise we will lose quality in our work. So yes, we must code. No matter the form: visual interfaces, natural language, or script lines. We must code not for the pleasure of technique, but to be capable of building and, especially, of evolving our work tools. Because what determines a human being’s impact on the world? Their thought, their sensitivity, their empathy, of course. But in service of all this, there is the tool, and the adapted and evolving tool.

To screw in a screw, I need the appropriate screwdriver. To get from point A to point B, a vehicle is more efficient than walking, especially if there are several hundred kilometers. To send a personalized message to a large number of people in a simple and automated way, I need a practical system that I can adapt to my specific needs; this can completely transform my organization’s relationship with its audiences, and modify its impact, the exercise of its missions, and its projects in depth. As media theorist Marshall McLuhan said in 1964, “We shape our tools, and thereafter our tools shape us”. Let us never forget that the tool we create as human beings determines the nature of our power over the world. And by “power,” I don’t mean domination, but indeed the power to act, to transform, to create new connections and, ultimately, to advance humanity. So let us be the authors of our tools, to continue being capable of making things change for the better. Yes, this requires learning new languages, but this no longer poses any problem today; we can do it as we go thanks to AIs, while we do things. There remains therefore only one blockage, that of not daring.