Recently, Emad Mostaque, the CEO of Stability AI, stirred up the internet when he boldly predicted that human programmers would be obsolete within five years. He went even further with an even more aggressive and dire forecast, stating that outsourced coders in India would lose their jobs within two years! Naturally, this sparked a lot of skepticism from online critics, questioning the "stability" of his claims. However, given that his company is named “Stability AI,” many did not expect such an unstable spread of information from him. But as Steve Jobs once encouraged, we should be more understanding toward those who believe they can transform the world, like these so-called “square pegs in round holes.”
In reality, long before GenAI entered the scene, engineers and scientists had already been diligently working on the creation of low-code/no-code platforms, particularly in the realms of Data Science and Machine Learning. One only needs to look at the rapid growth of these platforms in Gartner’s Magic Quadrant for Data Science, from 2017 onwards, to witness the increasing dominance of these low-code platforms. In fact, platforms like Dataiku, Alteryx, KNIME, and Rapid Miner have consistently dominated the top quadrant over the past five years.
As Simon Sinek wisely pointed out, to inspire action, we must begin with “why.” Similarly, just as great leaders spark action by starting with their "why," some innovations that have fundamentally transformed human existence also start with this question. Let’s dive into this.
Why are Low-Code/No-Code Platforms Essential?
The answer to "why" can be broken down into three components: 1) Democratization 2) Industrialization 3) Efficiency
#1. Democratization – We can frame this as “AI/ML and data science for the masses”. While data science offers organizations unique competitive advantages, there is still a significant shortage of readily available talent in the field. As a result, many forward-thinking companies have expanded the field of data science and AI/ML to nontraditional roles, such as the “Citizen Data Scientist”. Gartner coined the term Citizen Data Scientist to describe individuals whose primary roles lie outside of statistics and analytics but who are trained to perform predictive and prescriptive analytics within organizations. Low-code/no-code platforms serve as excellent tools to foster the growth, application, and scalability of AI/ML and data science within businesses.
#2. Industrialization – This can be framed as the “Rapid scaling of data science and AI/ML” throughout organizations. The drag-and-drop features and plug-and-play capabilities of these platforms, utilizing reusable components (or “accelerators”), allow for the creation of new solutions at roughly half the time compared to using traditional coding tech stacks (SAS/R/Python + Analytics Database). For instance, in the “coding world,” building a handful of models might be possible, whereas in the “low-coding world,” thousands of models can be created in the same timeframe.
#3. Efficiency – This can be defined as the development of data science solutions “With half the workforce, in half the time”. In the traditional “coding world,” you might need 20 people to complete 20 data science tasks (e.g., descriptive, prescriptive, predictive analytics) over 20 days. In contrast, in the “low-coding world,” only 10 people are needed to accomplish the same tasks in just 10 days.
How are Low-Coding Platforms Transforming Our Work? How Does a Typical Data Science Project Differ?
Teams leveraging low-coding platforms are now able to perform real-time analytics on the fly, even during meetings. This is a groundbreaking shift. The old practice of waiting weeks or months for business questions to be answered is being replaced by the ability to respond to ad-hoc requests almost immediately, as the questions arise.
In the traditional “coding world,” analysts must query databases (requiring knowledge of SQL) to retrieve data, then perform analyses using SAS/R/Python before presenting results via a Business Intelligence tool (like PowerBI, Tableau, Qlik, MSTR, etc.). In the “low-coding world,” everything can be done within the same platform, eliminating the need for multiple team members with varying skill sets. Additionally, in the “low-coding world,” there’s the flexibility to write code in R/Python and integrate it directly into the platform's native workflows. This makes it an ideal environment for “coders,” allowing them to focus on analysis, insight generation, and model accuracy, rather than grappling with syntax and switching between different platforms. This represents a revolutionary new approach to working on Data Science projects.
What Are the Advantages of Using a “Low-Coding Platform”?
The benefits are twofold – “top-line impact” and “bottom-line impact.” More data-driven decisions can be made, and AI/ML-driven intelligent decision-making systems can be seamlessly integrated into the business, driving democratization and industrialization. This directly boosts the business’s top line. Furthermore, companies can reduce their data science teams by half and still achieve the same results with these platforms, which leads to significant savings and improved efficiency, impacting the bottom line.
Finally, returning to the original question of “whether GenAI will signal the end of low-coding/no-coding platforms”, my answer is a resounding “no.” Even if GenAI is employed to generate code, the understanding of how to build analytics and AI/ML solutions through coding will still be crucial for utilizing code generated by Large Language Models (LLMs). The code generated by LLMs will likely remain cryptic and unreadable to most non-coders, preventing them from understanding and deploying data science solutions. This is where low-coding platforms excel. Their visual coding framework offers simplicity and transparency that is unmatched by anything currently available. While LLMs might eventually generate workflows similar to the way they generate code today, the outcome will still lead to the universal adoption of “low-coding platforms.”