Is GenAI an Existential Risk to Low Code/No Code?
Low-code/no-code platforms simplify software development. So does GenAI. While the sun isn’t setting on low code/no code just yet, it may eventually -- or not.
Low-code/no-code tools have grown in popularity by both developers and “citizen developers.” While the former had trouble accepting it at first, the accelerating pace of software delivery created a need for it that was too difficult to deny. And, low code can also be used to reduce technical debt, which many enterprises greatly need.
Then, GenAI hit the scene, providing yet another easy way to generate code, and suddenly, the game began to change.
Like most other types of software development companies, low-code/no-code platforms started adding GenAI to their offerings, which may be a benefit in the short-term, but perhaps an existential threat over the long term.
For example, Ramasamy Palaniappan, CTO at IT services company TEKsystems says GenAI is not a risk to low code/no code in the near future but as it becomes more efficient through reinforced learning, he expects GenAI to have an impact.
“In order to introduce GenAI tools into any enterprise use cases involving critical business logic in applications, there needs to be a mature approach to governance and security. As a result, some companies are reluctant to embrace GenAI tools for code generation at a large scale,” says Palaniappan in an email interview. “Low-code and no-code tools stay within an organization’s firewall and don’t move the generated code to public foundation models. However, as enterprises develop greater GenAI maturity, I expect to see more organizations embrace these tools for specific applications which will reduce the workload for some low code tools.”
Like some of the others interviewed for this story, he says development has different layers of abstraction: Physical code to be implemented (first level), low code/no code (second level) that accelerates the development process, and GenAI (third level) that will be able to translate customer requirements into implementable units. These GenAI tools will work directly with code or low-code/no-code tools.
Low-Code/No-Code Tools Are More Mature Than GenAI
Low-code/no-code platforms have existed about 10 times longer than ChatGPT (though large language models (LLMs) are nothing new). What is new is a simple interface that automatically generates code based on prompts written in plain English. While it is common knowledge now that low-code solutions can generate about 80% of application code, leaving 20% for edge cases and parts of code that still need to be human-written, GenAI may be able to duplicate that split in the future.
“LLMs aren’t there yet, and I think that’s where AGI [Artificial General Intelligence] comes in,” says Peter McKee, VP of Developer Relations and Community at Sonar, a developer solution for continuous code quality. “I think combining LLMs with AGI -- the ability to reason on top of these LLMs is going to be extremely powerful.”
Development teams at AWS cloud services company Caylent are using GenAI to create rough drafts, such as an organized code repository. According to Ryan Gross, senior director of data and applications at Caylent, it has dramatically improved performance in the way low-code/no-code systems did about five years ago.
“I think the disruption is really around the user experience. [Low code] vendors have spent a lot more time and effort with their product teams interfacing with developers and understand they want the results to be presented,” says Gross. “So, if they can build in the technology prowess required to harness this generative AI approach, they can stay ahead of the technology vendors as it relates to really catering to the needs of that core power user or local developer.”
For now, GenAI is helping developers become more productive, but it can’t replace human reasoning.
Why GenAI Might Not Replace Low/No Code
Vitaliy Koshelenko, Liferay architect at enterprise software services provider Aimprosoft, believes GenAI will lead to a synergistic evolution with low code/no code rather than a disruptive shift.
“We can see how GenAI tools enhance low-code/no-code platforms by automating repetitive tasks, improving user experience, and expanding the range of applications that non-developers can build. This integration makes them even more powerful, accessible, and efficient, says Koshelenko in an email interview. “GenAI does not threaten low-coding and no-coding platforms but complements and enhances them. It makes them more powerful and efficient by automating routine tasks and improving user experience. Together, they create a promising future for software development.”
Abhijit Kakhandiki, chief AI and innovation officer at full stack automation company Redwood Software, says while GenAI offers opportunities for both IT professionals and non-technical users to become more productive and focus on value-add tasks, trust in its use for mission-critical, hands-off processes remains low.
“[R]ather than seeing AI as an existential risk to low-code/no-code development, I see it as reshaping the landscape. Generative AI is already pretty good [at] generating code snippets or boilerplate code, reviewing code, and even testing code. Programmers are increasingly using it for proof-of-concepts,” says Kakhandiki in an email interview. “Low-code platforms can be seen as an ‘ease of use’ layer translating coding syntax to visual drag and drop. GenAI is expanding the pool of potential developers by increasing citizen developer skill sets or lowering the barrier for advanced coding. Eventually, due to these trends, the future development landscape will be of a more collaborative nature between citizen developers, professional programmers and AI -- ultimately leading us to create to better applications faster.”
About the Author
You May Also Like
2024 InformationWeek US IT Salary Report
May 29, 20242022 State of ITOps and SecOps
Jun 21, 2022