How readily did your devs adopt AI into their inner development loop? Was it a seamless addition or did you have to encourage them to embrace it?
Sort by:
We use GenAI in our core product, so it's important for my devs to be up-to-date on AI, but it's still a bit of a push from me to get everyone to experiment.
Honestly, the older, more experienced crew members have been the most hesitant to try using it, and I get it. However, tools like Claude Code are picking up a ton of grunt work and those using it are 'more productive' in many ways.
Getting AI tools to summarize code, offer plans on refactoring, writing small focused methods/functions, crafting or updating tests, and more specific spot checks, such as going through compilation or lint warnings, has been great.
It's able to do a lot of the work that people, including myself, typically just put off.
AI doesn't get tired, nor does it argue that these small tasks are a waste of its time.
On the other side I am also warning everyone that AI code tools can not become our primary coding tech. The tools are still extremely limited and can bring in all kinds of logical errors, old code, libraries, etc...
They are -tools- but often those who are masters of many tools end up being able to navigate more complex work.
We've been using all the GitHub Copilot available models. Developers have to opt-in and we do not force anyone to use AI ... yet. The early adopters have been sticky and have found great value in it. The feedback is consistent that it is useful and this is in a highly customized codebase with odd use cases on the core platform systems.