AI Writes the Answers Now, So What Should Programmers Do? One 1024 Day Answer: Start Small

---

theme: fancy highlight: monokai

In the past, writing code really meant typing it out line by line yourself. When you hit a problem, you'd ask a senior first, then search Baidu; if you ran into a bug you couldn't solve no matter what, spending the whole night tossing and turning was not unusual.

Now? Halfway through writing code, you come across something you don't understand, casually ask AI, and a few seconds later you get an answer. Staring at the screen, I sometimes suddenly think: so what can programmers still do from now on? Should I rush to start a side hustle, learn a new direction, or just switch careers?

This question is pretty real. But the moment it comes to mind, if you open Zhihu and start scrolling through 'how many more years do programmers have' and 'how ordinary people can earn X per month with AI', it's easy to get more and more anxious. Today we're not talking about industry trends, and we're not looking for a standard answer. Let's just treat it as a few colleagues sitting down to chat.

You can think about a side hustle, but don't treat it as a lifeline

When side hustles come up, many people's first thought is 'earning an extra salary after work'. But once you actually try it, what you might get first is: working during the day, changing requirements at night, replying to messages on weekends, and in the end your hourly rate is worse than just resting properly.

If you want to try, I think it's most natural to start from what you already know. For example, build report automation for a small team or community you're familiar with, help someone turn repetitive operations into small tools, or write up problems you've solved at work as articles or talks. They may not make money right away, but at least you can see whether anyone really needs them.

Taking a small project is also fine, but it's best to make clear from the start: what will be done, when it will be delivered, how many revisions are included, and who maintains it afterward. What wears you out is usually not how hard the features are, but the small change agreed on at the beginning slowly growing into a project with no end. I've been burned by this myself; poor communication just creates trouble for me later.

As for projects that require a large upfront payment, stocking up on goods, or promise easy money, I'd stay wary; basically, I don't trust them. A side hustle can take risks, but there's no need to stake your savings and after-hours time before validating demand. Also remember to draw a clear line with your company's business, code, and client data; don't bring yourself bigger trouble just to earn a little extra.

When learning something, it's best if you can use it right away

When you're anxious, it's especially easy to sign up for courses and bookmark tutorials, feeling like you need to learn every new language, framework, and AI tool. Your bookmarks get fuller and fuller, but the work in front of you stays the same.

Instead, switch to a smaller goal: next time you encounter repetitive work, let AI process it once for you; or pick a business problem you're familiar with and try to build a small tool that works. After you finish, see whether it saved time and whether other people use it, then decide whether to keep learning.

Besides tools themselves, programmers can also learn more about the industry they're in. Many requirements sound like 'build a feature', but as you keep talking you find that what's really stuck is the process, the data, or how departments hand things off. Whether technology can help often depends on whether you've understood what the other person is frustrated about.

If you really want to switch careers, you don't have to resign today

If you already really dislike your current job, or just want a change of environment, switching careers is certainly worth considering. But phrases like 'work in AI', 'become a product manager', and 'go freelance' sound fresh; once you actually step in, there are new troubles to face.

You can start by trying it in your spare time. If you want to do product, try taking a small tool from requirements to the point where someone is willing to use it. If you want to build AI applications, don't stop at calling APIs; see whether it can solve a specific problem. If you want to take freelance work, first take one small job with a clear scope and see whether you like the communication, quoting, and delivery.

If you try it and find it interesting, and someone is willing to pay for the result, then gradually scale up. If you try it and find it doesn't fit, it's not wasted effort; at least you've avoided one impulsive resignation based on imagination.

When it comes down to it, it's normal for AI to make people uneasy. If I really had to pick something to do right now, I'd choose something small: this week, find a task that annoys you every day and try to save half an hour with AI or a small script. If you can save it, write down the process; if not, you'll at least know where the problem is stuck.

Okay, that's all for now. There's still half an hour before I get off work; I'll tidy up today's daily report. Finally, happy 1024 Day to everyone.