Coding Assistants - Friend or Foe? - by Chris Westerhold
Coding assistants have been around for around two years and have generated more hype than just about anything in the past twenty years. The first mainstream model, ChatGPT, was released in 2022, and many other variants have followed. There have been promises of 30-50% productivity gains and even full replacement of engineers. Let's dig into the details and see if we can get an answer to a very complex question: Do genAI tools drive better productivity?
If you haven't heard of coding assistants in 2024, then you must have been under a rock for the previous couple of years. While most have heard of them, there is still quite a bit of confusion about what they actually are. This is not a deep dive into the technical nature of LLMs so let’s just call them what they are, excellent text predictors.
They have the ability to answer questions, summarize data, and provide code for certain tasks. At first glance, they can seem to be magical with seemingly great answers to lots of different questions across many different languages and tasks. This ability has immense use cases across software engineering and the broader employment landscape.
