GitHub in danger because of GitHub Copilot?

A not-for-profit organisation that provides support and legal services for open source software projects has urged the community to ditch GitHub.

GitHub in danger because of GitHub Copilot?

Software Freedom Conservancy is a not-for-profit organisation that provides support and legal services for open-source software projects. It has called on the open-source community to ditch GitHub after they themselves quit its internal company usage.

But why is this happening?

We need to go back in time a little to understand it truly.

Microsoft launches GitHub Copilot

On June 29, 2021, GitHub launches a tool for writing code better & quicker. Basically, an AI-powered programming buddy (thus the name Copilot), scrapes your code, understands the context and provides you with machine-learning generated code snippet which is a developer's wish to use.

It came in with a lot of hype and everyone was interested to use it but it had limited access and slowly it started getting accessible by all the developers.

Backlash

GitHub Copilot was received with lots of enthusiasm but there were many criticising it as well. The open-source community criticised GitHub copilot because it went against the basic ethics of the community, it is a paid service!

The open-source community was already not happy when Microsoft bought GitHub for a whopping $7.5 billion back in 2018. The reasons for backlash were simple, the developer community couldn't trust Microsoft. Since Microsoft is a corporation and the community worried that they might start generating some kind of revenue from GitHub and kill the sense and power of open-source software. Because the highlighting point of open-source software is its free usability (more about it later in the article).

Cut to 2019 Microsoft partnered with OpenAI, an AI research organisation, and poured $1 billion dollars to develop GitHub Copilot. This tool leans hugely on OpenAI Codex, and it was trained on a gigantic amount of open source code available. The community raised the question of copyright infringement. But in front of the hype of this new amazing tool, this backlash never gained much of light.

What has changed now?

So now, Microsoft announced various Enterprise versions of GitHub copilot and Software Freedom Conservancy said they had enough! The organisation said, "We were already considering the action (leaving GitHub as their internal tool for open-source projects) for some time, but the latest announcement (GitHub Copilot Enterprise version) showed that this action is overdue." in this blog post.

Although GitHub copilot is offering free service to many developers around the globe, but there are few very important questions that are being raised.

  1. Open-source doesn't mean a complete free-for-all, and there are still licence requirements, attributions, and certain requirements to fulfil.

  2. It is trained on data of publicly available codebases and natural language models, does it mean that providing code snippets of some author to the author of another project can lead to copyright infringement lawsuits?

  3. Microsoft and GitHub claim Copilot to be trained on billions of lines of code publicly available but have completely denied the requests to release a list of repositories included in the training model.

The secrecy from a corporation like Microsoft has only left the community under speculations.

A long-time foe of Microsoft, Amazon recently launched their own AI-powered programming assistant CodeWhisperer, and it is clear from the functioning of the tool that they are addressing major issues and backlash faced by GitHub Copilot. CodeWhisperer will suggest code snippets and if it matches any existing code snippet used during the training of the model, it will display the licensing and the original code snippet/function/codebase and after that, it is up to the developer how to use it.

That's why the problem still remains with GitHub copilot as there is no clearance that code suggested by GitHub copilot then is completely owned by the end-user and can have copyright over it? And if it is so, the AI-powered tool is always learning and can suggest that code snippet to someone else and can it lead to major legal issues? These questions have been time and again raise but never answered by GitHub.

Is GitHub in Danger?

Honest opinion, I don't think so. GitHub has maintained that special kind of stickiness with the developers that although there are alternatives like GitLab and Bitbucket developers choose Github over any other platform. It is still the most popular collaborative platform for developers. Also, open-source communities never get the limelight against these large corporations because the community works on ethics and not on monetary benefits. Thus, GitHub is hardly going to get affected by the decision taken by Software Freedom Conservancy, but they are pushing this agenda and want people to know the truth.

I'm personally a long-time user of GitHub and it is really hard for me to leave it right now and move to some other platform. But there are concerns to be raised about the hard work and time and efforts put by developers around the globe to power up the developer community. And we shouldn't let these big corporations turn those hardworking projects into a stream of revenue for them.

My gadgets

Laptop (Macbook Air 2020 M1)

Keyboard and Mouse: amzn.to/3eTvFTP

External Display Monitor: amzn.to/3doZI5z

External SSD: amzn.to/3BrrSoD

USB Hub for Macbook Air: amzn.to/3ScC3E1

Headphones: amzn.to/3qR5yzP

(some of the above are affiliate links ie if you purchase any item using the links above I'll get a very small percentage of it and you'll get the product at the same price)

Did you find this article valuable?

Support Lavish Goyal by becoming a sponsor. Any amount is appreciated!