This repository is meant to demonstrate the capabilities GitHub Copilot has in regard to solving algorithm-adjacent interview questions.
It's main intent is to demonstrate how Github Copilot breaks bad interviews and encourage employers to stop using non-representative interview questions.
In our testing, Copilot was easily able to solve the following problems using only the function name as reference:
While we currently only have JavaScript code samples, you're more than welcome to add code samples in any language you'd like!
We initially chose to use JavaScript since GitHub claims that Copilot works best on it
Since Copilot is able to use your current file for guidance on what to generate based on context, we've done the following to make sure that our data is not skewed by our own code or more information than intended.
- We start with an empty dev file
- We type the least necessary code and wait no longer than 5 seconds for recommendations
- Make sure to add both a solution and a usage example of the solution
- We'd also like Copilot to generate the usage example whenever possible. If not, please make a note of it in the source (see below)
- Any comments starting with "// Copilot" or "// NOTE: " were added after-the-fact.
- We add any additional context (if we had to check Copilot's other suggestions, if we tried other function names, etc)
- If unspecified, the present code was Copilot's first suggestion
- We add a comment block comment at the top of the file explaining the code problem after-the-fact for the reader's benifit
Have access to GitHub copilot and want to submit other questions similar to this? Please do!
Simply fork this repo. Then, add a question in your fork. Make sure to follow our methodology in order to avoid mistaken inputs.
Afterwards, send us a pull request and we'll review and merge!