Today I watched Tuobuhua talk about the “Law of Recklessness” on WeChat Channels. One point strongly resonated with me:

Often, we do not act because we have thought everything through. We understand things after we act.

Especially in the AI age, we see new tools, projects, and opportunities every day.

Some people want to create content, some want to learn AI image-making, and others want to build mini programs, edit videos, or develop agents of their own. But when it is actually time to begin, we always feel unprepared:

Have I chosen the wrong tool?

Is this direction still worth pursuing?

Without the relevant skills, will what I make be terrible?

What if I invest the time and get no result?

So we save a pile of tutorials, study countless examples, and make plenty of plans. The only thing we do not do is actually start.

Increasingly, I feel that many people do not lack opportunities. They are simply used to thinking every possibility to death before acting.

Why does more thinking make it harder to start?

We tend to assume we are hesitant because we have not considered things thoroughly enough.

But often, what keeps something unclear is not a lack of thought. It is a lack of information that only action can supply.

Without writing, you cannot know whether you can keep writing.

Without filming, you cannot know whether you enjoy being on camera.

Without making a mini program, you cannot know whether you are really interested in products, design, or the process of solving a problem.

Before beginning, all the problems you face are imagined ones.

Only after beginning does the question change from “Can I do this at all?” to “How exactly should I change this step?”

These two kinds of question may look similar, but their nature is entirely different.

The first only wears you down. The second might actually be solved.

The greatest value of action, then, is not to prove immediately that you made the right choice. It is to turn a hypothesis hanging in your head into a question reality can answer.

How does action reveal the next step?

Recently, I registered a new WeChat public account called “Growing with AI.”

Before actually doing it, I could have kept wondering:

Is the name good enough? Is the positioning too broad?

Can I keep posting?

Can I move articles over from the old account?

Does starting a new account from zero have any point?

However long these questions spun in my head, I would only have been circling my own imagination.

Only when I actually registered the account, chose the name, made an avatar, set the account ID, wrote the welcome reply, and published the first article did the questions become concrete:

Which topics do readers respond to more?

Where should personal experience appear in an article?

How can views, examples, and experiences with AI tools form a consistent content structure?

I still did not receive an answer saying, “This account is certain to succeed.”

But I gained something more useful than that answer: a sense of what to adjust next.

Starting will not immediately give you the final answer. It will give you the next step.

Looking back, I actually do this quite often in everyday life.

When I buy a cup, I try whether ice water makes condensation form, whether coffee makes it too hot to hold, and whether it feels comfortable to pick up. Instead of repeatedly reading reviews, I would rather really use it once.

When cooking rice, I adjust the ratio of rice to water. After a few meals, the texture that suits me naturally emerges.

When learning watercolor, if I see a new wash or a way of preserving white paper, I usually want to try it that very day. Only after painting it myself do I know whether it merely looks attractive or can become part of my own expression.

These are small things, but they make me increasingly certain:

Some understanding cannot be thought into existence. It grows from your contact with reality.

Creating AI-related content and trying new tools are the same.

You can spend a long time researching which direction has the best prospects. Or you can first write three articles, make a mini program demo, or edit a video with AI.

Afterward, look at where you were most absorbed, what others responded to, which difficulties you are willing to keep solving, and which things only briefly attracted you because they were trending.

Then your judgment of a direction is no longer just an imagined one.

“Recklessness” is not impulsiveness: three steps to lower the cost of trying

At this point, someone might ask:

Does that mean doing whatever occurs to you without considering costs or consequences?

Of course not.

Quitting a job, borrowing money, investing heavily in assets, or making any decision that could cause major loss to yourself or others is not a suitable way to test the “Law of Recklessness.”

As I understand it, “recklessness” is not blind risk-taking. When the cost is manageable, the result observable, and you can stop at any time, it means no longer substituting endless thought for action.

In practice, it comes down to three sentences:

STEP 01

Make the action small. Instead of asking at the outset, “Should I change careers and work in AI?”, ask, “Can I make one small piece of work in a week?”

STEP 02

Keep the cost low. If you can test something in your spare time, do not start by testing it through resignation. If you can make a demo, do not rush to develop a complete product.

STEP 03

Review quickly. As soon as you finish, ask yourself: Was I engaged? Did I get real feedback? If I did it again, would I want to keep improving it?

Discovering that something is worth continuing is an answer. Discovering that it is not right for you, and stopping in time, is also an answer.

Stopping after a low-cost attempt is not failure. It is obtaining a piece of real information at a manageable price.

Recklessness or selective focus: which should we choose?

Interestingly, my previous article was about “actively choosing what not to do.”

In an investor discussion, Liang Wenfeng mentioned that DeepSeek focuses only on the main path toward AGI. It would not pursue areas such as 3D and video generation that were not closely related to that path toward intelligence.

One idea asks us to do less; the other encourages us to begin. They seem contradictory, but concern two different stages.

Borrowing terms from AI and algorithms, growth always involves two activities: exploration and exploitation.

Exploration means encountering new possibilities and finding directions worth pursuing.

Exploitation means concentrating resources on directions already tested, so that advantages continue to accumulate.

The “Law of Recklessness” serves exploration. “Actively choosing what not to do” serves accumulation.

When you have no main direction, you need action to find it. Once you see it, you need choices that protect it.

With only recklessness and no selectivity, you keep chasing trends. With only selectivity and no exploration, you may commit too early to a path that has never been tested.

An effective sequence for growth, then, should be:

Use action first to discover what is worth doing. Then use selective choices to decide what is worth doing over the long term.

I do not yet know what “Growing with AI” will eventually become.

It might become a place for continually recording collaboration between people and AI. New columns and connections might grow out of the writing, or it might take me in directions I cannot yet imagine.

But none of those answers could appear before the first article was published.

What I can do is write first, publish first, make contact with real people, then keep adjusting according to the feedback.

If there is something you have thought about for a long time without starting, perhaps set aside the question of whether it can succeed.

Ask yourself first: Can I shrink it into an experiment I could begin today and afford to fail at?

Write three articles, record one video, paint one picture, make the simplest possible demo, or just give it a serious seven-day try.

Beginning does not mean you have to continue for the rest of your life.

It simply gives reality a chance, at last, to tell you whether this is worth continuing.

“Recklessness” does not mean giving up thought. It means no longer using thought as a substitute for action.

Because a real direction is never fully figured out before departure. It grows, little by little, through action, feedback, and choices.

AI collaboration note: The views, personal experiences, and final judgments in this article were supplied by the author. AI tools assisted with organizing the text and making illustrations.

Source note: In public materials, the “Law of Recklessness” is often summarized as “Get started, and you are halfway to success,” emphasizing progress through attempts, feedback, and correction. This article interprets and extends the idea through personal experience; it is not a word-for-word account of the video. The discussion of Liang Wenfeng and the AGI focus refers to a transcript published on July 23, 2026.

References: Materials on the “Law of Recklessness” · Transcript of Liang Wenfeng’s discussion · Tuobuhua’s WeChat Channels account