For the moment, AI is helping developers, not replacing them
GitHub Copilot, a code generation and autocomplete wizard developed by Microsoft and OpenAI in 2021, has been widely adopted by the developer community. A number of studies now reveal how it is impacting on them and the job market; and the conclusions are, to say the least, interesting.
In many ways, GitHub Copilot has led to a certain level of automation in code development, allowing developers to increase their productivity. It is estimated that up to 92% of them use some kind of AI tool for their work, with Copilot the clear leader with 54.8%, followed by Tabnine (12.9%), AWS Codewhisperer (5.1%) and others, which implies not only a very high percentage of adoption, but also the evidence that, despite this, the professional category of developer has not suffered as a result, nor has it resulted in mass layoffs.
This type of tool is clearly for assistance, not replacement: auto-completion of lines of code, tedious debugging and other boring and repetitive tasks that, when carried out by the wizard, allow the developer to focus on more interesting and productive ones. Like all generative assistants, it reduces creation time, but significantly increases monitoring. What’s more, as it is trained with public and open repositories, it often suggests outdated code or ones with security problems, which requires care and…