我認真用 AI 過了一個月的生活，老實說有點回不去了

Let's cut to the chase: I'm not one of those "AI evangelists." Initially, I thought it was silly to ask AI for help deciding what to have for dinner. But after using it for a month, there are some things I just can't go back to. This article isn't a list of tools; it's a record of my real-life experience over the past month.

The Beginning: I Was Just Too Lazy to Think

Here's what happened. On a Wednesday evening, I was standing in front of my fridge, staring blankly at its contents: half a head of cabbage, three eggs, and a block of tofu that was about to expire. I had no idea what to cook. In the past, I would have just ordered takeout and then regretted the expense. That day, I randomly took a photo of my fridge and sent it to ChatGPT, asking it, "What can I make with these ingredients?"

It gave me three options, complete with steps. I made one of them, and it turned out to be surprisingly delicious. At that moment, I thought to myself: Hey, this might actually work.

What Really Worked: Those "Too Lazy to Think" Tasks

After a month, I found that AI was most valuable to me in situations where I was "too lazy to think, but still had to do something."

Planning a trip was the most obvious example. My friend and I were going to Tainan for three days, and in the past, I would have spent an entire evening researching attractions, planning the itinerary, and deciding which restaurants to visit without having to wait too long. This time, I sent my requirements to Perplexity (because it provides sources, which made me feel more at ease), and within ten minutes, I had a rough plan, which I could then adjust myself. The time I saved wasn't just time; it was also the frustration that came with planning.

Another example was writing "difficult messages." I needed to inform my landlord about a leak, but I didn't want to be too aggressive or too soft. I asked Claude to help me with the tone, and after two revisions, I sent the message, and the landlord came to fix the leak the next day. In situations where I "knew what I wanted to say, but didn't know how to phrase it," AI was able to help a lot.

What Didn't Work: Don't Kid Yourself

However, I must be honest and say that there were many things that were a complete waste of time.

Asking AI to help me decide "whether to exercise today" was pointless – of course, it would tell me to exercise, but I would still end up lying down. Asking it to help me plan a "daily schedule" was also useless – it would create a beautiful plan, but I would only follow it for two days before giving up. For things that require "willpower" rather than "information," AI was of no help. It would only give you a perfect plan that you wouldn't follow, leaving you with more guilt.

My takeaway is that AI is good at handling "information problems," not "human problems." If you don't know how to do something, it can teach you; but if you know what to do but can't do it, it can't save you.

An Unexpected Side Effect

After using AI for a month, I discovered something: I became better at "asking questions."

Because I needed to get good answers from AI, I had to clarify my questions – my budget, my limitations, what I cared about. The process of "thinking through my needs" actually forced me to clarify many things that were previously vague. Sometimes, I would have the answer before I even finished asking the question. This was probably the most unexpected benefit.

So, Should You Use It?

If you ask me, I would say: Don't treat it like a omnipotent assistant; treat it like an "intern who is always available, has a lot of knowledge, but needs clear instructions." The better you are at assigning tasks, the more useful it will be; if you treat it like a fortune teller, don't be surprised if it's not helpful.

After a month, I didn't become a master of life, and my fridge still often only has three eggs. But for those "too lazy to think" tasks, I did become a lot more efficient. That's enough. If you want to find the right tools for yourself, you can check out our AI tool categories and start with the free ones.