The Step Change in Programming at the Intersection of Large Language Models and Low-Code-No-Code Platforms
Low-Code-No-Code + Large Language Models = New Programming Paradigm
A new paradigm in programming is being unleashed at the intersection of Large Language Models (aka Generative AI, e.g. @OpenAI @openaigpt) and No-Code-Low-Code platforms (e.g. @MSPowerAutomate & @bubble). The first step change in programming since Object-Oriented Programming.
Large Language Models such as GPT-3 (and ChatGPT) unlocked a new wave of innovation in how we write code. While simple auto-complete of a few words existed for a long time, the new capabilities to auto-complete a full block of code or even to generate an entire code template from a single prompt is revolutionary. LLMs will cut development time significantly. They will lower the barrier to entry for software developers. And they will increase productivity for businesses. Additionally, No-code low-code solutions have existed for a while and been growing with big players such as Microsoft’s Power Platform, SAP, and Bubble IO. No-code low-code solutions limit what people can develop to an existing library of pre-programmed components and integrations.
New LLM-enabled code generation tools can overcome the no-code low-code limitations and will significantly reduce the cost of developing and launching a new software product. Like how cloud computing enabled companies to outsource their infrastructure and platforms, code generation with AI has the potential to alter how companies develop software, lowering the cost, cutting time-to-market, and unleashing a new generation of lean startups that can create and test a software prototype faster than ever before with minimal resources.
Vision
While the copilot model–an advanced auto complete model–is the most common today, the vision is for AI to generate code from a natural language description of what the code needs to do, without requiring the human to write any code.
Companies
Replit: the company started primarily as a cloud-based programming environment (IDE) and education tool. It launched Ghostwriter, an AI-based copilot that can auto-complete entire blocks of code. And with the vision to make it generate the entire code from a natural language description. The company offers development environments in the cloud for the most common languages and frameworks. This lowers the barrier to entry for people to start using it. The company’s focus is on education and helping people learn how to code from anywhere–mobile or desktop–positions it as the go to tool for future programmers who learn how to code with Replit and continue using it as they enter the job market.
Tabnine: the company offers an AI assistant for software developers. The product is a plugin that can be used with existing IDEs such as VS Code. Tabnine trains the model on a developer’s existing code so that it can recognize unique coding patterns for an organization. Tabnine can already go beyond autocomplete. The product can generate blocks of code from one-line natural language prompts (Exhibit 1). Other companies in the space: Mutable AI and Github Copilot.
Here is the Twitter version of this post:
You can also find me on LinkedIn:
