Salesforce CodeGen is a large-scale language model that enables conversational AI programming – in other words, speak to the machine and “let AI write code for you”. Sounds futuristic? Well, the future is here.
Also known as “program synthesis” or an “AI pair programmer”, these models work alongside you to convert conversation (the ‘natural language’ we speak) into programs. This goes far beyond autocomplete (which solely completes statements, and reduces typos and errors), which has been mainstream for some time. Instead, CodeGen will try to understand what you are looking to ultimately achieve.
How Does CodeGen Work?
For a full breakdown of the technology evolution towards conversational AI programming, it’s worth reading “Terms, Definitions, Concepts” in the Salesforce AI Research report.
With CodeGen, you don’t write any code yourself, you simply describe to the machine what the code should do. Borrowing Salesforce’s analogy, using CodeGen is comparable to ordering dinner at a restaurant; you tell the waiter what you’d like, and they serve it to you. You didn’t need to purchase the ingredients or cook the meal as part of the process.
It goes without saying that this functionality will make programming faster and less resource-heavy, but it will also give a leg-up for ‘low-code’ Salesforce professionals.
Teaching programming languages to “speak the computer’s language” is no longer a barrier – whether that’s mastering any (or additional) languages to work with specific programs. Salesforce’s ethos is to democratize access to technology, working to eliminate the cost, time investment, and pre-requisite skills that have traditionally stood in the way. Being language agnostic (i.e., not tied to a specific programming language) suddenly makes professionals highly transferable from one project to the next.
What Does This Mean for Developers?
Of course, CodeGen isn’t a silver bullet that will put developers out of work. As it stands, CodeGen can handle simple coding tasks on behalf of low-code professionals. When it comes to more complex problems, programming knowledge is advantageous to guide the system as it searches for a solution (or the human to suggest a more efficient solution!). As you saw in the video above, the user is instructing the AI to “solve the problem using a hash map” and “run the function on the numbers…”
Look at the whole development lifecycle, and you will see how the demand for developers will continue to grow. Writing the code is surrounded by absorbing business requirements, designing solutions to align to the specific tech stack, running tests, deployment, and more – all of which rely on the human touch.
It’s also important to note that the model has been trained by Salesforce AI Research and is open-source – the main reason is to accelerate research with inputs from contributors.
When it comes to the wider market, you’ll find GitHub Copilot (which uses OpenAI Codex) has had similar achievements.
Final Thoughts
The Salesforce AI Research report takes a highly altruistic view of the impact that CodeGen could deliver:
“The democratization of coding should reap society-wide rewards… We expect this disruption to be on the scale of other tech-based breakthroughs such as autonomous vehicles, or even the printing press… will benefit the entire economy”.
It will, therefore, be interesting to observe how Salesforce professionals adopt CodeGen – or conversational AI programming – more broadly; who will it be, pro-code or low-code professionals? And how fast will this concept take off? Stay tuned!