Rina who is a journalist and former software engineer based out of New Zealand, shared 4 tips for programmers to stay ahead of generative AI, in a recent IEEE Spectrum article.
In the following Video I try to summarise the key takeaways from the article. You can find link to the article in the description.
Being a Coder these days is creating both excitement and job insecurity. With the advent of ChatGPT and other generative AI models there is bound to be mixed feelings.
But there is some good news. “Some experts argue that AI won’t replace human programmers — not immediately, at least. But yes you will have to worry about people who are using AI replacing you”.
Let’s now look at the 4 key ideas
1. Focus on the fundamental principles and best practices
This is the place where you can shine. While AI has the capability to generate code and complete certain tasks, coders can play a larger role. Humans are still needed for activities like understanding user requirements, debugging, and testing. Problem-solving skills continue to be crucial. To make the most of AI-based tools, it is important to adhere to good software engineering practices, like planning system design and software architecture, which provide valuable context for the AI tools to accurately predict code needs and enhance productivity.
“Programming is easy. Asking the right question is hard.”
2. Find the tool that meets your requirements
There are various AI-based tools available, each for a different purpose. Experimenting with these tools can be beneficial. Like GitHub Copilot, which provides coding suggestions, and ChatGPT and Google’s Bard, which can answer API queries or generate code snippets. These tools can be helpful for learning new programming languages, frameworks, or initiating small projects and creating prototypes.
3. Develop good prompt engineering skills
One of the experts said “One important technique is making sure your prompt includes all the necessary context and information.”
“Programming is easy. Asking the right question is hard.”
AI coding assistants require clear instructions to generate the desired outputs.
A useful strategy is to break down a problem into multiple steps, known as chain-of-thought prompting. Additionally, it is recommended that coders gain a basic understanding of AI and machine learning concepts to better comprehend how Language Model Models (LLMs) function.
4. Exercise caution and understand the risks
It is important to carefully review the outputs generated by AI tools to identify potential bugs. LLMs can occasionally generate code that is inaccurate or incorrect. It is essential to consider copyright and security implications, as the generated code may have vulnerabilities.
In conclusion, as a coder for you to excel in an AI-driven world, it is crucial to integrate AI tools into your workflow while being aware of their limitations. You should continue to rely on your unique human coding skills and not lose sight of the broader context of software engineering.