I wrote several developer notes about AI and how it is changing our lives. Today, I want to formulate and share some basic ideas on how we can improve our AI usage skills in everyday life. Using AI tools has become as routine as using a mobile phone, and it has penetrated all spheres of our life. AI tools boost our productivity, but like any tools, we should use them correctly and efficiently. Once we use any AI tools, we tend to think we are experts, but we should train them like any other skill.
In one tech podcast, I heard a good suggestion: practice using AI tools at least 15 minutes every day. The main focus isn’t just on usage, but on thinking: how can I get some particular task done using AI? Identify a problem that can be solved with AI, find the right tools for it, and finally, implement it in practice. For example, I often need to create models that are for parsing JSON. The simplest way, just write a prompt and add JSON that need to convert and copy the output code and paste it in the correct place. Where is the possible optimization here?
First, this task is not unique, it is repetitive task, you can do it at least several times when you work on one project. As a result, you should write almost the same prompt every time. It is very complicated to get optimal result from the first time, it means that you should do several iterations because you get on expected result. After that, you can find out how the desired prompt should look at the initial step. As output of this work should be not only several classes but good structured prompt to get this result, and it is essential about what you care when you reach the goal. You should store structured and store this prompt for future using.
Second, you should use relative tool for specific task. The described result you can reach by regular chatbot, but it looks logically to use a specific IDE or plugin for this to reduce copy-paste work. What is the best for specific task, you can find just by practicing and playing around with different tool.
What is the key thought of all this, you shouldn’t think about using AI to get a quick result only, you should think as architect: What problem are you solving? What way are you doing this? How can you use it in a suitable and scalable way? You and I should not just use AI by training our skills in using it, and the right way do this by practice and in regular base.
You can find me in X, GitHub, medium or LinkedIn. Thanks for your time and see you in next post.
