LLMs have been proved valid coding assistants. We will discuss in the classroom how to use them by sharing experiences and caveats, as this topic varies greatly depending on the experience and use cases

Some general concepts:

1. You will be responsible of the final code. This means it’s advisable to use assistants to improve code you would be able to write yourself in any case
2. Some assistants are more “transparent”: you see the code generated and you are more in control. In this case you will be likely to learn something new if the assistant comes up with a solution you didn’t know
3. Some tools (generically called agents) automate the process to a level that you will not have as much control of the code unless you review it. They can boost the productivity at the expence of transparency
4. Assistant don’t get bored: they can help us writing more comments about the code we wrote, and better documentation, more thorough tests

---