Github Copilot - AI pair programming power tool

Github Copilot - AI pair programming power tool


Hello Everyone 🙏🏾,

I couldn't wait to publish this one, so I am pushing this out early and I hope you find it useful.

If you are a software engineer and haven't heard about the Github copilot feature then it's time to get an overview of this new shiny tool.

Twenty Years ago, I used the first smart code editor IntelliJ IDEA, which was fast and had intelligence (hence the name intellij) - it was fast, generates code and helps with refactoring the code base.


Now in 2022, Github Copilot is the new next generation code editing feature. Copilot is an AI powered tool trained on publicly available source code (I am guessing largest training data must have come from github itself ... remember microsoft bought github a while back)


I had a chance to try it past few months and I can say, it has definitely improved my productivity. I tested this for Java on IntelliJ IDEA, There is also support for other editors like Visual Studio, Neovim.

Here's my review so far.

The Awesomeness 👍🏽 ,

𝗜𝘁 𝗹𝗲𝗮𝗿𝗻𝘀 𝘄𝗵𝗶𝗹𝗲 𝘆𝗼𝘂 𝗮𝗿𝗲 𝘄𝗿𝗶𝘁𝗲 𝗰𝗼𝗱𝗲 𝗮𝗻𝗱 𝗰𝗮𝗻 𝗽𝗿𝗲𝗱𝗶𝗰𝘁 𝗻𝗲𝘅𝘁 𝗹𝗶𝗻𝗲 𝗼𝗳 𝗰𝗼𝗱𝗲..𝗹𝗶𝗸𝗲 𝘁𝗵𝗶𝗻𝗸𝗶𝗻𝗴 𝗮𝗵𝗲𝗮𝗱.



𝗜𝘁 𝗱𝗼𝗲𝘀 𝘂𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱 𝘀𝘁𝗮𝗻𝗱𝗮𝗿𝗱 𝗽𝗮𝘁𝘁𝗲𝗿𝗻 𝗶𝗻 𝘄𝗿𝗶𝘁𝗶𝗻𝗴 𝗰𝗼𝗱𝗲 𝗹𝗶𝗸𝗲 𝘀𝘂𝗴𝗴𝗲𝘀𝘁𝗶𝗻𝗴 𝗶𝗺𝗽𝗼𝗿𝘁𝘀, 𝗶𝗻𝗶𝘁𝗶𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻, 𝗰𝗼𝗻𝘀𝘁𝗿𝘂𝗰𝘁𝗼𝗿𝘀, 𝗴𝗲𝘁/𝘀𝗲𝘁 𝗺𝗲𝘁𝗵𝗼𝗱𝘀 𝗲𝘁𝗰. 𝘀𝗲𝗲 𝘃𝗶𝗱𝗲𝗼 𝗯𝗲𝗹𝗼𝘄,

𝗧𝗵𝗲 𝗺𝗼𝗿𝗲 𝗽𝗼𝘄𝗲𝗿𝗳𝘂𝗹 𝗳𝗲𝗮𝘁𝘂𝗿𝗲 𝗶𝘀 𝗽𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗹𝗼𝗴𝗶𝗰 𝘁𝗼 𝘄𝗿𝗶𝘁𝗲 ... 𝘁𝗵𝗶𝘀 𝗶𝘀 𝘀𝘁𝗶𝗹𝗹 𝗻𝗼𝘁 𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗲𝗱 𝗯𝘂𝘁 𝗶𝘁 𝗵𝗲𝗹𝗽𝘀 𝘆𝗼𝘂 𝘄𝗿𝗶𝘁𝗲 𝗹𝗲𝘀𝘀 𝗰𝗼𝗱𝗲 𝘆𝗼𝘂𝗿𝘀𝗲𝗹𝗳. 𝗛𝗲𝗿𝗲'𝘀 𝗮 𝘀𝗶𝗺𝗽𝗹𝗲 𝗮𝗻𝗮𝗹𝘆𝘁𝗶𝗰𝗮𝗹 𝗰𝗼𝗱𝗲 𝘁𝗵𝗮𝘁 𝘄𝗮𝘀 𝘀𝘂𝗴𝗴𝗲𝘀𝘁𝗲𝗱 𝗯𝘆 𝘁𝗵𝗲 𝗰𝗼𝗽𝗶𝗹𝗼𝘁, 𝘀𝗲𝗲 𝘃𝗶𝗱𝗲𝗼 𝗯𝗲𝗹𝗼𝘄,

𝗦𝘂𝗿𝗽𝗿𝗶𝘀𝗶𝗻𝗴𝗹𝘆 𝗶𝘁 𝘄𝗼𝗿𝗸𝘀 𝗼𝗻 𝘁𝗲𝘅𝘁, 𝘁𝗿𝘆 𝘁𝘆𝗽𝗶𝗻𝗴 𝗮 𝗰𝗼𝗺𝗺𝗲𝗻𝘁 𝗮𝗻𝗱 𝘆𝗼𝘂 𝘄𝗶𝗹𝗹 𝘀𝗲𝗲 𝗿𝗲𝗰𝗼𝗺𝗺𝗲𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗳𝗼𝗿 𝘁𝗲𝘅𝘁. 𝘀𝗲𝗲 𝘃𝗶𝗱𝗲𝗼 𝗯𝗲𝗹𝗼𝘄,

And the Caveats 🥸,

  • Nuances; It doesn’t recognize the library version you are using so the recommendation does not match. 
  • Be careful: It may introduce destructive bugs and it’s indistinguishable from correct code so review the code before you accept it.
  • Limitation, It will repeat suggestion after some number of times. To get new suggestions you have to write a new line of code.
  • Ugly: Lots of "system.out" recommendation, don’t write production code with system outs.
  • Beware: Before you use it at work, make sure you talk to your manager about Github data collection, because the data snippets are sent to GitHub to generate suggestions.

Read more about it here https://github.com/features/copilot.

Now you have a new tool in your toolbox, so save some time for a coffee and take a stroll.

If you like my newsletter subscribe to get notifications, and feel free to connect with me on LinkedIn if you like to talk to me.

Happy coding 👋🏾!!!

#githubcopilot #github #ai

Helpful info . Keep them coming .

Like
Reply

This is very cool indeed ... however - remember the accountability is still lies with us as developers. While bots may suggest / recommend ... we still own the code ... as in when bugs are discovered and they will be ... we need to address them :) ... not 'blame the bot' 😊

Like
Reply

To view or add a comment, sign in

More articles by Suresh Ramakrishnaiah

  • Enterprise Agentic Architecture - Strategy

    In the age of AI, software architecture needs to evolve rapidly—but how can organizations do this? How do you modernize…

    1 Comment
  • Using LLMs to create a ChatBot Experience

    I spent years architecting and building a Conversational AI (Artificial Intelligence) Chatbot. Building a chatbot…

  • ChatGPT - Conversational AI Chatbot

    ChatGPT from OpenAI has created a buzz in the tech world. This is a brief introduction to ChatGPT.

  • Top 5 Books for Managers

    This is my first article in my newsletter series, I plan to write monthly on Artificial Intelligence, Cloud Computing…

    7 Comments

Others also viewed

Explore content categories