In My To-Do

I have a to-do list. Somewhere. Finding it has been on my to-do list for quite some time now.

I'm sure most of you have used a to-do list at some point in your life. It could be scribbling notes in a book, marking things on a calendar, or using applications like Todoist, Microsoft To Do, Google Tasks, or TickTick. And I'm equally sure that none of them would have satisfied you completely. There would have been those little moments when you wished, "Why can't this application do this? Why can't I organize things the way I want?" Well, we all have our own ways of managing things, and no application seems to get everything right. Now, what if you could build one yourself? One that works the way you want, with all those little things you always wished for.

Let’s build one then. Pick an AI system of your choice, whatever LLM you prefer. Start with a simple prompt: "I want to manage my day-to-day activities. Create me a to-do list." See what it creates for you. Don't worry about complicated features like login systems. You could run it on the web or your local desktop. Depending on your technical knowledge, keep prompting, exploring, and refining. Add the features you always wanted, remove what you don't need, and see how close you can get to the application you had in mind.

Now, here is something to ponder. How is building a to-do list different in the AI world? Earlier, if I had to design, develop, test, and host an application like this, it could have easily taken me six months, even with dedicated effort. Today, with AI taking care of much of the coding, you could have a working application in an hour or two, or maybe a day. That's quite a change. But have we really harnessed the potential of AI? Think about it. If we are building the same to-do application that existed ten years ago, only faster, what have we actually achieved beyond saving time? Now that we have intelligence at our disposal, shouldn't we be questioning the very way a to-do application is designed? What could it do differently? What could it do that was never possible before?

Go back to your application. Experiment with different prompts, add new features, remove a few, and see how far you can take it. Don't worry about getting everything right in the first attempt. The idea is to explore, experiment, and understand what you can build with AI. More importantly, see if you can think beyond what existing applications already offer. At some point, you might get frustrated with what AI creates, or you might end up building something much closer to what you had in mind. Either way, there is something to learn. Before you call it complete, here are a few questions I want you to ponder.

Reflections:

1. What problems did your application solve, and what remains unsolved?
2. What did you design yourself, and what did you simply accept from AI?
3. If you used this application for six months, where would it fail?
4. If AI can understand your habits and priorities, should a to-do list still be just a list?

Enough of reading now. Get started. Start building it. Get your hands AI-dirty!