Some learnings and thoughts around Artificial Intelligence
To all the software professionals out there………
I know that we all are getting overwhelmed by the rise of Artificial Intelligence around us, especially with the advancement of generative AI. Hearing terms like Prompt Engineering, Model context protocol(MCP), models like GPT, Claude, etc, Agents, and so on makes us feel hectic wondering if it may replace us or will be difficult to learn. Even I had the same feeling.
But I want you all to relax here a bit. Yes, the advancement is happening, but I want you to slow down a bit and explore Artificial Intelligence without any worry, considering the hype around it. You might wonder what this guy is saying. But, if you read this article, you will realise there is no need to rush into.
Before learning AI, I would like to highlight this LinkedIn post that I wrote a few months back, which can set the stage for all to explore AI. Here is the link: replace “Software” with “AI”
If you can understand my message, you will get a feeling that you won’t get replaced. Instead, you will try to learn and help others, either creating or using AI tools. With this message in mind, start to learn about AI. Start from the basics, from what AI is, how it has started, why it’s required and then go further. Most importantly, learn about the different biases that AI has, which provide awareness for us to handle carefully. Be aware that even AI has bugs: Bugs around AI and AI-related tools. In this way only, my learning experience has started.
I got introduced to AI during my undergraduate in Mechatronics as a subject where I learned what it is all about. I understood some basic concepts, starting from Turing Tests, AI-related Search algorithms, first and second-order logic, and expert systems. Then, after a long time, I explored Intro to Testing Machine Learning models from Test Automation University. Here are some of the key learnings from the course: Key learnings from Testing Machine Learning Models.
When I heard about ChatGPT, I started using that and took some courses related to prompt engineering. I also conducted a small experiment comparing the responses between ChatGPT and Google Bard(Now it’s Google Gemini). Based on my current learning experience with AI tools like ChatGPT, Google Gemini, Microsoft Copilot and Github Copilot, I can say that it can help us as a digital assistant, but it requires proper instruction to get the required output from it. That’s where learning about Prompt Engineering comes into play. I learned about different kinds of prompts and started practising with them. Also, I have analysed the output and tuned prompts to receive a proper response. I also had an opportunity to test an AI chatbot during my current job. Based on all these experiences, I understood that
Domain/Subject Knowledge + Programming Knowledge + Prompt Engineering = Enhanced Output/Improved Productivity
I didn’t try prompting with LLM APIs, which I am planning to explore soon. After this, I will go for RAG and Agentic AIs. Also, along with this, I am looking to learn about ethics around AI, which is really important to handle AI-related tools properly.
Still, there are a lot of things to learn. But, I am not willing to rush myself here as I need to properly learn to understand where it can help me to solve problems based on my needs and which Large Language Models can help me to solve those problems. I can say that with the help of GitHub Copilot, I was able to learn Jenkins Pipeline scripting quickly, which helped me to create the CI/CD pipeline for my automation work. Also, it helped me as a second pair of eyes during debugging the pipeline and Python scripts that I have created. Based on my experiments, I understood that I need to use proper tools for the proper context. If I get any ideas in future, I will try to create some AI-based tools.
Finally, I would like to say that explore AI at your own pace, keeping the message in mind that I have shared. Soon, it will have an impact on both our professional and personal lives. But don’t rush. Try to understand it properly and then see how you can solve problems either using or creating AI tools. AI can’t replace us if we embrace it as part of our lives, like how we are using our smartphones and smartwatches. So,
Be calm and embrace AI