In a previous piece, I opined that Artificial Intelligence (AI) is a tool to augment, not replace, human capabilities.
I received requests for use cases that may be relatable as the current excitement from the democratisation of access to AI tools continues to grow. Let us look at Github Copilot.
GitHub is a code-repository service that provides tools to empower software development teams to collaborate and work more efficiently.
It covers the entire software building cycle from access control, version control, bug tracking, task management, continuous integration, and documentation.
In June 2018, Microsoft purchased GitHub for $7.5 billion in stock. At that point, GitHub supported just over 28 million developers.
Fast forward to today, it hosts over 94 million developers.
In 2019, Microsoft took a $1 billion stake in OpenAI, an artificial intelligence research outfit.
Over the next few years, it would quietly put in an additional $ 2billion, a clear indicator of the potential of the technology built by OpenAI.
In May 2020, OpenAI released its third-generation autoregressive language model dubbed Generative Pre-trained Transformer 3 or GTP-3 in short, whose free research preview has gone viral.
Microsoft’s investments in GitHub and OpenAI intersected beautifully with the creation of GitHub Copilot, that as its name infers, empowers software developers to output cleaner and higher-quality code faster.
It gets context from what a developer has set up and written, drawing also from training data found in publicly accessible source code.
Based on the code a developer has written, it anticipates what they will likely write next.
This ‘wingman’ is available to millions of developers as an extension of Microsoft’s suite of integrated development environments.
With an always-available companion to offer suggestions, the productivity of a developer or an entire team can go up a notch.
GitHub Copilot has additional features such as error checking, code formatting, and refactoring. Extrapolated for time saved when attending to code, the place of this AI assistant becomes obvious.
For software houses, it may translate to more projects. For in-house teams, it may mean working through roadmaps faster, leading to increased customer delight.
That said, at best GitHub Copilot remains an assistive tool. The developer or teams are in charge of the final output and must go through the rigour of review, testing, and other processes for quality checks.
Closer home, you may use a keyboard tool that borrows from the same concept helping you type that email or text response faster and with better grammar, but ultimately you decide what goes out.
Njihia is the head of business at Safiri Express | www.mbuguanjihia.com | Twitter: @mbuguanjihia
PAYE Tax Calculator
Note: The results are not exact but very close to the actual.
