GITHUB COPILOT : Your AI pair programmer

GITHUB COPILOT : Your AI pair programmer

In mid-2021, Github shook programmers with its technology, Github Copilot. GitHub CEO Nat Friedman, Copilot works with a broad set of languages and frameworks such as Python, Ruby, TypeScript, JavaScript, and Golang. But it is not limited to these languages. This matter is designed to dominate others as well.

 in this article us will focus on discussing Github Copilot which can work like autocomplete when coding. All you need is to type in the requirements in the Editor and Github Copilot will suggest or suggest the next syntax. How cool is it?

What is Github Copilot?

Copilot is an AI programming tool released and developed by Github that aims to help programmers to write code faster automatically.

Github Copilot is based on Codex, an AI system created by OpenAI and trained on the source code found on github itself.

How does Github Copilot work?

Visual Studio Code is currently providing Github Copilot as an Extension in beta. You can use it on your system that has VSCode installed.

                                                   Sumber : https://copilot.github.com/

In its note, Codex has collected training datasets since May 2020 from 54 million source code repositories downloaded to Github, with a capacity of 179GB of unique Python programming language files less than 1 Mega Byte in size.

The criteria filtered by Codex are as follows:
  1. The code can at least be auto-generated
  2. Have a line length of approximately 100 or more
  3. The maximum line length is approximately more than 1000
  4. Have a small percentage of alphanumeric characters

After Codex does the filtering, the dataset that will be used by Codex is 159 GB to develop Github Copilot. From here, Github Copilot can see the patterns or flows that you write when writing the program. This Copilot will understand what you're writing and suggest solutions for the next code write.

I looked at and observed the article https://medium.com/@cecepbudiman/ about Github Copilot here I am disappointed, in the process Github Copilot always writes the wrong code like there are some libraries that are also stuck in the suggestions even though I don't use the library (because it is still in the development stage). But I think the ability to provide automatic code suggestions is good and will definitely be a perfect tool in the future.

Github Copilot features

1. Convert comments to code

Copilots can understand comments and function names. Write a comment explaining the logic or flow and Copilot will set up the code.

2. Test without the toil

The test in question is Import unit test package on this feature, let Github suggest tests that match your code.

3. Autofill for repetitive code

Copilot helps autofill (write directly) to loop code. You just write the initial code example after which the Github Copilot will automatically pass it on.

4. Show alternatives

Every programmer must use different methods in implementing their code. Here Copilot can make adjustments to the code you implement (code suggested by copilot) or you can change it according to your needs.

Will Copilot replace Programmers?

Behind the many benefits of Github Copilot, it turns out that many people are also worried that with this tool, the role of programmers becomes very small. And many rumors and news are circulating "Will Github Copilot replace programmers in the future?"

I can't be sure, but from the Github side itself states that Copilot is expected to be useful as a programming assistant to help programmers be more productive and cut time to focus more on work.

Berita Rekomendasi

Malware Baru Menargetkan Sistem Linux dan Perangkat IoT!

12/11/2024

New Malware Targets Linux Systems and IoT Devices!

A new part of a hidden Linux malware called Shikitega has been found to adopt a multi-stage infection chain to compromise endpoint points, IoT devices and deposit additional payloads. The findings add to the list of malware...

View
Perisalah AI: Meeting Jelas, Eksekusi Tepat

30/12/2025

AI Issue: Clear Meetings, Precise Execution

Have you ever felt nervous after a meeting because you were worried about forgetting important points? Or have you ever made a mistake in execution because you didn't clearly remember the agreed-upon action items? You are not alone. Meetings are…

View
Strategi Tepat Memilih Metode Deployment

19/02/2025

Strategi Tepat Memilih Metode Deployment

Pemilihan strategi tepat dalam memilih Metode deployment dalam Continuous Integration/Continuous Deployment (CI/CD) berperan penting dalam memastikan bahwa perubahan kode dapat diterapkan dengan cepat, aman, dan tanpa gangguan pada pengguna akhir.…

View