The following practical tips are adapted from the tabnine website where they are covered in more detail:

DO

  1. Make sure that you know and understand your company/university policy on the use of AI coding assistants.
  2. Take time to choose the assistant that provides the best value for your context. Consider capabilities that go beyond code completion and include more advanced features such as code explanations and intergration with your repository and other development tools.
  3. Look for an assistant that can be fine-tuned for your specific development technology.
  4. Take advantage of the time saved by using the coding assistant to give more attention to feature development, user experience, etc.
  5. Spend time experimenting with what your assistant can do. This is a fast-moving area and new tools will be released often. Make sure you stay up to date.

DON’T

  1. Try to hide your use of coding assistants. Transparency is vital.
  2. Use the first option that you come across or the most well-known without checking its suitability. Some tools are tuned for software develop while others are not - take the time to understand their capabilities.
  3. Assume that choosing a coding assistant is a one-time decision.
  4. Worry about AI taking your job - instead, concentrate on becoming an expert in how to deploy generative AI tools effectively.
  5. Let the AI do all the work - they will make mistakes and are no substitute for human judgement.