As a lifetime technologist, I frequently get asked by my non-tech friends and family: “What is going to happen with AI? What jobs are going to be impacted?”  When I tell them that you probably shouldn’t be majoring in accounting or radiology, they aren’t too surprised. But when I tell them that software development as an industry will be equally disrupted, they are usually a bit taken aback.

In the relatively near future, accounting as we know it as a job for humans will be gone. Humans won’t be categorizing expenses in Quickbooks and ensuring balance sheets tie out. Of course there very well may be new finance-related positions. Individuals who can work with AI to run a variety of financial scenarios, to look for unexpected opportunities, or to act on concerning predictions. But the job as we understand it today will vanish, or at least transition. 

Of course this belief is not universally shared—there are elements of nitpicking here. If AI is “aiding businesses in systematically addressing reconciling bank statements, categorizing transactions, generating financial reports, and more”, then by definition the role of the bookkeeper has been transformed.

The same thing is true for software development. Today, software engineers are already making use of generative AI as part of their jobs nearly every day. AI has been integrated into products such as Microsoft VS Code to provide AI-assisted development. Contextual’s AI Assistant will write your data schema for you based on a natural language conversation. And lots of developers just ask chatGPT to write code for them and proceed to copy/paste the code.

Developers who regularly use generative AI to write code all say the same thing—that AI is really good at writing little “snippets” of code. Today’s large language models have been trained in literally every programming language ever invented, from COBOL to Mojo, and they are fantastic at writing small programs, methods, or functions.  In fact, you can do that within Contextual’s flows today since the underlying ‘atoms and molecules’ of javascript are well understood by those tools.

So today's software developers are leaning in and using AI to be more efficient and effect. But why then do I think that AI will replace software engineers entirely?

Primarily because we are now only at the beginning of this “AI revolution”. AI in 2024 is comparable to the state of the World Wide Web in 1991 when the first website went live and the first browser was able to display it. Today, Ai is saying "Hello World".

So what’s the next step in this evolution?

While AI assistants for helping developers will continue to evolve and improve, there is also significant energy being put into training AI models and putting in place AI orchestration to replace the developer entirely. Business users can now start to engage with the AI directly to define their needs, create tests, and deploy  new solutions.

Within the next two years, AI will build entire software systems with humans defining requirements, user stories, and business needs in plain—non-code—language. Initially, this will be facilitated by limiting the scope of components that the AI is allowed to use. At Contextual, this is our core premise. By delivering a well-defined and understood building block approach to software development, our platform will allow AIs to focus entirely on solution development versus infrastructure and architecture. This allows systems integrators, business analysts, and developers to focus on delivering business value quickly.