Getting started with AI Coding assistents

Ionut Ciuntuc

Nov '25 • General discussion

If you are a begginer at using coding tools, like Visual Studio Code or PyCharm, Jupyter Notebook, this one is for you.

I will soon have 1 year of using AI coding assistents with different tools.

I did the tests, so you don't have to.

I will start with Visual Code and GitHub Copilot.

This is an extension of Visual Code that gives you access to different AI models.

It has a 30 day trial period, a free version and a payd one (10 dollars basic - that I use).

What you get with it:

you enter a prompt - describe what automation you want to do and in what programming language;
it makes a plan;
it configures the programing medium - installs libraries, virtual enviroments and so on;
uses agents too write the code - different modules.
it writes tests and documentations;

How to install it - in the next link.

https://docs.github.com/en/copilot/how-tos/set-up/install-copilot-extension

Or just go to the extensions button in VS code and follow the pipeline.

You will probably be asked to make a github account. It is free and usefull later.

By default you get access to AI models to use.

The best in my experience is Claude Sonnet. And you get a fair amount of usage for 10 bucks.

After that, you use the other free models - ChatGpt, Grok and so on.

The second tools, a fork of Visual Studio Code is Cursor AI.

This one is 20 bucks but you get a 15 days trial.

Basicaly the same principles above aplies.

Possible to be more powerfull than GhitHub copilot.

The third one is another extension of Visual Code called Augment Code.

It handles verry whell big code contexts - works the same like the ones above.

Costs - as far as I remember 30 bucks.

I just used it in the free trial.

Forth - Google Colab.

This one is online. Uses Gemini model.

Free to use with some limitations.

It is similar to a jupyter notebook.

No special set ups to get access to the AI models.

They are built in.

While without advanced knoledge in Software engineering, these tools can't make a full complex software with 100.000 lines of code, one can develop quite nice and usefull apps with them even as a begginer.

Aplying some strategies, fragmentations and having an extension mindset, building a system, not a closed complex tool, you can create your own "Computer" from Star Treck to help you do your job and be awesome at the same time.

It can save tons of hours.

And this is what we are going to build and document here.

7 comments

Getting started with AI Coding assistents