AI Adoption for Software Engineers: The Research and Why Delaying is a Mistake
Get started now.
TLDR
Learning a new tool can be hard because you are less productive at first.
Data shows that you get more productive with AI the longer you stick with it.
AI is coming. Make the time to start playing with it at work so you don’t fall behind.
The S-Curve: Why New Tools Are Scary
With any new tool there is always an initial decrease in productivity. You have to learn the new tool. This causes you to go slower initially than if you just kept doing it how you’ve always done things.
This short term setback is offset by the long-term gains of adopting a good tool.
But in the beginning it can be difficult to manage. You have to take a step back, slow down, and struggle a little bit to benefit in the future. This is hard for most folks and can make getting started with something new prohibitively intimidating.
Research on GitHub Copilot Shows Accelerating Productivity
In a study done by Harvard Business School, Github, and Keystone.AI on a sample of 900k+ CoPilot users they found that users on average accepted 30% of suggested code.
More interesting was that after a few months, the acceptance rate began to accelerate.
By the Numbers:
CoPilot Suggestion Acceptance Rate
First 3 months: 28.9%
Next 3 months: 32.1%
The growth in productivity follows a typical S-curve we see when adopting new tools.
How to Start Adapting to AI Today
Whether you’re a fan or not, AI is only getting better. It will be a required part of your job in the near future. The logical step is to start adapting to it today rather than wait until you are forced to.
There are two things you can do:
Invest in timeless skills that won’t be replaced by AI.
Begin incorporating tools like CoPilot into your workflow.
Join the The Augmented Engineer for weekly exercises to adapt your skills as a software engineer. The Augmented Engineer was created to serve as a place for engineers to chart this new course together.