My Experience with GitHub copilot
Srinivasa Raju Indukuri
Director of Engineering @ Fresh Prints | Web Technologies
Last week, I got my access to GitHub copilot. I was slightly skeptical about how much will this be useful before I started to use it.
Today I had installed my vs-code extension and logged into the account. Then started working on my personal project. Spent most of my time on refactoring my code today. There were suggestions which I felt not great when I was doing the code refactoring.
Then I had to write a simple utility to get Socket URL and Graphql URL based on the hostname and protocol. If the host is localhost we have to point to a local backend server which is running on different port. This is a very common use-case. When I started to code this utility I was amazed by the amount of code that was generated by copilot. The following code is what we(me and copilot) have come up.
Out of the the 24 lines of code in the file, 11 lines of code was auto generated by copilot. This is close to 50% of the code. I crossed the line numbers which were generated by copilot.
Though at times the code generation is slightly slower, But once the tool matures it would be of great help to programmers. I feel that this will help developers with their productivity and improve the quality of code.
Let us see how these AI for coding will shape the future of application development.