Coding is Dead!

The computer was one of the most useless inventions of humanity. They do nothing on their own unless you tell them exactly what to do. They can only interpret and act on a set of instructions at a time, composed of strings of zeros and ones.

Now, the issue is how to make the computer understand what it needs to do; well, no one speaks to their computers directly through zeros and ones; that's why programming languages were invented.

Programming languages are essentially a set of instructions and structures capable of forming algorithms based on logic within a paradigm or platform that allows us humans to create strings of zeros and ones that the computer can interpret and execute.

There are countless programming languages; the closer they are to the machine (with the ability to generate optimal code), the more complex they are. Modern high-level languages have been simplified to such an extent that they have gained worldwide popularity and adoption.

At the end of the day, we use code to translate business requirements into a set of flows, rules, screens, and data that will change states to control something. I see no need to develop any system if the ultimate goal isn't to control and organize something (from a business perspective).

The act of programming becomes a dialogue between the mad. At the very least, it involves two people whose approaches are generally diametrically opposed: a user or business owner and a programmer. The former knows the business, the flows, the processes, and needs a solution to organize and control it. On the other hand, we have a developer, familiar with one or several development platforms, who knows how to create instructions and programs but is completely unfamiliar with the business. This is where the problems begin.

The developer expects the business owner to make ONE structured and logical request that can simply be translated into code, comprehensively and definitively on the first go. And the business owner expects the developer to understand the business dynamics from a brief chat and some rough sketches on a napkin.

The ordeal starts with this three-way dialogue among the business owner, programmer, and the computer; inevitably, this triangle will turn into a nightmare.

Now imagine this scenario amplified to the 'n'th power when we're talking about organizations with hundreds of users and programmers.

Wouldn't it be easier to tell the computer what is required, in a clear and concise manner, and have it respond with a proposal that can be used and tested immediately? If there's any deviation between what was requested and what was 'generated', could we clarify or specify it in more detail, perfecting the 'solution' interactively and iteratively?

Let's take a few steps back, actually several, many, two or three decades back.

Do you really think your business is so different from the one next door? The way you account for your business, manage your accounts receivable/payable, manage your prospects or execute a marketing campaign, or manage your supply chain are business processes and daily challenges that most businesses face. Do you think that the systems and business processes (cycles) that a bank has or manages are, functionally speaking, different from those of the bank across the street?

More than three decades ago, configurable systems that do everything, known as ERP, were born. SAP is the king of the hill, and while it does involve some coding, roughly 80% of a 'custom' implementation is configuration. In these systems, all common business cycles have been pre-developed into configurable modules that can be interconnected. In banking, there are platforms like the legendary Altamira, which was essentially the same but on a mainframe system.

If you've had the great privilege of being part of an ERP system implementation team, you'll know the not-so-small nightmare this represents. And the nightmare originates from the same love triangle we discussed earlier, between the system, the user, and the configurators/implementers.

Wouldn't it be better if a business expert could tell SAP about their business processes, show them the data, and SAP would begin to configure itself for the user to start testing the system and dictate adjustments?

I believe you're already catching on to where we're heading in the world of systems.

GenAI is a singularity in itself, but it's just a small taste of the potential it holds.

The intermediary that exists between the person who needs a system and the system responding with a solution is about to disappear. In that sense, I say that Programming is Dead.

Where do programmers fit in now?

Ultimately, for this magic to happen, AI models or LLMs (Large Language Models) need to learn from something to generate things similar to what they've learned beforehand. Imagine if we fed the model only with horror novels; it would generate horror novels. To keep current systems running, they will need to continue learning from the source of all genius: human creativity.

In this aspect, innovative software components developed by humans are needed, which these self-configurable systems can 'use' to learn. All the prefabricated pieces capable of building a skyscraper from scratch must exist somewhere, created by some genius.

Think about it: the GitHub phenomenon, where everyone's code is used for copy-paste to quickly get that tedious university project done, has now vanished overnight; the reason? It was used to train several GenAIs which now basically write the code for you. If this trend continues, we might reach a drought of innovation in coding because everything will be a new representation of the same thing (thinking negatively and in the long term).

How do I prepare for this future today?

The future does not belong to those who have the answers; after all, we have Google for that, but to those who know how to ask the right questions to get the answers they need.

The same applies to this upcoming future; the key role will be for those who have deep knowledge of the business, its strategies, and differentiators, and who have the ability to abstract that complexity and express it in words, in questions, in prompts that will be entered into these self-configurable systems.

1. Become an expert in prompt engineering: Play with all available GenAIs and do the same thing you do now, but through prompts, even planning your day, learning something, or generating your gym routine.

2. Program without programming: Create an app for your phone, build a website, or perhaps a game using only prompts from a GenAI. It will be torturous and/or painful, but achieving it will open your mind to what you now consider impossible. If you get stuck on something, ask the GenAI.

3. Dive into Low Code/No Code tools: SharePoint is the ideal toy that comes free in the cereal box. It's the Fisher-Price platform for LC/NC. Become an expert in creating data-driven solutions using SharePoint (and without using Excel). You can get a personal SharePoint for less than a couple of tens of dollars a month.

4. Learn to diagram: To create diagrams that model your business and capture its complexity in diagrams that an 8-year-old child could understand. Use free tools like Miro. Map everything that happens around your work: processes, systems, data, people, clients, products, etc.