ALL in AI: 為什麼你要立即開始Vibe Coding?

Recently, two things happened that made me feel like I had to make a video to tell everyone,

why every one of us needs to start so-called "vibe coding" right now.

I know it might sound a bit anxiety-inducing, but I believe after watching this video, you'll understand that

what I'm sharing is a reality that many people haven't realized is already happening.

The first thing is about myself.

You might have noticed that I haven't updated in a while.

Not only that, but I also recently canceled two confirmed sponsorships worth over $10,000.

The reason is simple: they would get in the way of vibe coding.

To me, the value of vibe coding far exceeds these short-term gains,

and time is the most precious element.

So what exactly is vibe coding?

Why am I so obsessed with it?

And what have I been doing?

We'll get to that later.

The second thing happened when I was in Austin for the first trial run of Tesla's Robo Taxi.

I was having dinner with my friend Lao Yu, who was traveling with me, and he mentioned that many people around him have lost their jobs,

including programmers from big Silicon Valley companies and middle managers.

What's worse is that they are finding it very difficult to get new jobs.

I wasn't surprised to hear this, because it's a result I had anticipated long ago, and I believe

this process will only accelerate.

In fact, the experience of Lao Yu's friends,

and my decision to fully immerse myself in vibe coding, both stem from the same underlying understanding.

An epiphany I had right after ChatGPT first appeared.

(Original news channel background sound)

For thousands of years, a complete value creation chain has roughly been: idea,

to resources and tools, to the final result.

Within this framework, we primarily play two roles at various levels,

owner or user.

For the vast majority of people, our role is often the latter.

To put it more bluntly,

most of the time, we act as the resource or the tool itself.

For example, there's a term called "human resources." The word "resources" refers to us.

Similarly, many of us are essentially "tool people."

I know that defining people as "tool people" might feel offensive to you.

But to understand all of this, we first need to see a fact clearly.

(Rhythmic background music)

The core of historical technological revolutions has always been automation, replacing human labor with machines.

However, whether it was the steam age, the electrical age, or the information technology revolution, they all achieved only partial automation,

because past machines couldn't operate completely on their own.

They needed people as cogs in countless links to operate, maintain, and connect them.

The social division of labor, therefore, became increasingly fine.

A person in their entire life,

could gain a space for survival by mastering a tool in just one specialized field.

If you think about it, the essence of most of our jobs revolves around a specific tool.

Uh, just like many people in college think that learning Photoshop well can get them a job,

and some elders even simplify programming to being good at typing,

saying that if you're good at typing, you can find a job.

This reminds me of a scene from "The IT Crowd" where the female protagonist applies for a job in the IT department.

The boss asks her if she knows IT.

She says of course, I'm a computer expert.

(Original video sound)

Although this is a comedy, it actually reveals the essential relationship between many of our jobs and tools to some extent.

Of course, you might say that our value is far more than the tool itself; we also have knowledge and experience,

these are intangible assets, right?

Hmm, not necessarily.

Because when a knowledge worker must continuously and actively work to exchange knowledge for

income, then their knowledge is more like an advanced tool that they need to operate personally.

Once they stop working, the value creation also terminates.

Only when they can package their knowledge into a product, a book, a course,

an automatically running IP that can continue to create value without their direct involvement,

does knowledge truly transform into an asset.

From this perspective, the so-called knowledge of most of us,

is essentially still a type of tool, and we are still "tool people."

I'm not saying this to deny the value of "tool people." On the contrary, we should be grateful for this era,

it has provided every one of us with an opportunity to survive.

But what I want to emphasize is that the golden age of "tool people" is being ended by AI.

The technological revolution brought by AI is completely different from the past. It's not another partial automation,

but the starting point towards full automation.

First, is the full automation of the virtual world. AI can call tools on its own,

completing end-to-end closed loops in the digital world and knowledge work.

This is why the first to be hit in this wave of unemployment are the once-stable knowledge workers. AI,

this omniscient universal tool, has caused the professional barriers we used to rely on for a living to collapse in an instant.

Second is the full automation of the physical world.

With the emergence of humanoid robots, the manual labor we once thought was safe is also in jeopardy.

Many people don't understand, we already have robot vacuums and washing machines,

why do we still need to build humanoid robots?

Can it wash clothes by hand faster than a washing machine?

Of course not, but that's not the purpose of humanoid robots at all.

The purpose of humanoid robots is not to replace machines,

but to use the vast existing machine systems designed for humans, washing machines,

coffee machines, computers, cars, and everything in factories. Its purpose is to replace people.

So do you see it now?

In this revolution, tools will still exist,

but the "tool people" who heavily rely on operating tools for survival are being rapidly eliminated.

.

Large-scale structural unemployment is an inevitable result.

When Lao Yu and I were talking, he mentioned a detail that made my heart sink. He said that those tech companies,

it's not that they don't have money. After laying off programmers, they use the saved money to buy "cards."

The so-called "buying cards" is to double down on investment in AI computing power. For companies,

buying cards has a higher return on investment than hiring "tool people."

In other words -- cards and AI are now the resources,

and "tool people" will soon not even be resources anymore.

(Rhythmic background music)

Here, we must shatter a widely circulated beautiful illusion, which is that many people believe

technological revolutions will always create new and even better jobs for those who are displaced.

Sam Altman recently posted a well-intentioned tweet, depicting a future AI world where

humans will still find new opportunities and engage in beautiful creation.

But I want to say, Sam here is more like playing dumb while knowing the truth.

I previously recommended a book on my channel,

called "The Art of Doing Science and Engineering."

Its author, Richard Hamming,

made accurate judgments about the development of technology and its impact on society in this book decades ago.

He clearly pointed out that the so-called idea that technological revolutions bring new opportunities for the displaced,

is more like a, uh, beautiful vision that lacks evidence.

We can understand this point with basic common sense. Hamming gives an example in his book,

it's hard to imagine training a large group of displaced coal miners to become programmers.

I can actually attest to this to some extent.

My father was a laid-off coal miner. In his life trajectory,

he didn't transition to being a programmer or any more advanced knowledge work.

So,

the real situation is not as simple as providing some training to allow people to leap from low-end jobs to high-end jobs.

simple.

Many people now imagine that when AI replaces repetitive labor,

people can generally turn to creative work.

This idea is somewhat romantic.

As Hamming said in his book, in any era, the people who can truly engage in creative work,

probably only account for a very small minority of the population.

So the truth of history may be more cruel. Every step towards automation is inevitably accompanied by a painful period.

Most carriage drivers did not become car drivers,

and most laid-off workers did not become programmers or artists.

For history, this may just be a brief moment, but for every individual in it,

this moment could be the rest of their life.

So why is this the case? Hemming actually gave a deeper explanation.

This is consistent with the views of Geoffrey Hinton, who has always been warning about the risks of AI.

That is, as a biological organism,

our speed of evolution is far from keeping up with the speed of technological evolution.

Okay, so here's the question. Technological evolution will inevitably lead to a large number of people being eliminated.

So where did these eliminated people go in history?

From a macro perspective,

the increase in total social wealth brought by automation can indeed guarantee people's basic living through welfare programs.

This gives the displaced a buffer period to find new social roles.

For example, from a carriage driver to an elevator operator, from a farmer to a migrant worker,

but this is essentially just a horizontal shift of roles, from an old "tool person" to a new "tool person,"

and they will eventually face the risk of being eliminated again.

And in this era of AI, which is tending towards full automation, the possibility of this kind of horizontal shift will become even smaller.

Therefore,

the truly new opportunities in history are not prepared for the majority of the displaced as we imagine.

It is a rarer, more essential leap.

It's not from being an old "tool person" to a new "tool person,"

but a leap from a "tool person" to an owner of tools and resources.

Just like in the information revolution, the real winners were not the office clerks who learned to type and typeset,

but the group of entrepreneurs who mastered the new tool of the internet,

and used it to create platforms and gain capital and social status.

So, facing this revolution led by AI towards full automation,

every one of us is standing at a huge fork in the road.

To the left is being passively eliminated, searching for the next, possibly even less valuable, way of survival in anxiety.

To the right is to actively face the difficulties and seize this once-in-a-century paradigm shift,

to complete the leap in social roles.

I choose the latter.

That's why I'm so urgently invested in Vibe

Coding, because my goal is very clear: to become the master of new tools,

new resources.

My personal strategy is also divided into two levels around this goal.

First, become an owner of resources.

To put it bluntly, it's investing in the core resources of the future in the new era.

For example, I've invested a significant portion of my assets in companies like Tesla,

because I firmly believe it plays a crucial role in the AI revolution.

My goal is to become one of the owners of the future autonomous robots and self-driving car fleets.

one of them.

Of course, to deploy the world's best means of production, a powerful investment platform is indispensable.

Here I want to thank Interactive Brokers for their long-term support of this channel.

It is the tool I use to execute this strategy.

If you are interested in investing, feel free to click the link in the comment section to learn more.

Your click will also bring revenue to my channel.

My second strategy is to actively complete the transformation of my personal role and become the master of AI.

This transformation, for us who are used to executing tasks as "tool people," requires overcoming two major obstacles.

First, the mental obstacle, which is to transform from a passive executor to an active master who gives orders.

giving orders.

Second, the capability obstacle, which is to give up the obsession with a certain specialized field,

step out of the comfort zone, and learn to master more macro and general new abilities.

And Vibe Coding is the specific way I practice this second-level strategy.

Wait, you might ask,

uh, learning to command AI, doesn't that just make you a new kind of "tool person"?

Not at all.

Becoming a commander of AI is completely different from mastering any kind of tool in the past,

because it has two core characteristics.

First of all, we must accept a premise, that is, the evolution speed of AI is exponential.

Whether you like it or not, it is inevitable that it will replace most of our jobs.

And in the face of this inevitability, there is actually a counter-intuitive strategy, which is that you should not avoid it,

but rather, early on,

use AI deeply and actively to accelerate this process. Why is that?

Because we found that in the process of using AI deeply, the role we play and our way of thinking,

will undergo the most fundamental changes.

[speaker__0] And this discovery comes from a practice in our community.

Last month, we launched a community theme called Vibe Code,

and even created a dedicated group for it, with the purpose of leading everyone to practice it deeply together.

During the event, many friends verified this point,

that is, if you still command AI with an employee mentality, you will find yourself easily getting stuck in, uh,

getting bogged down in implementation details, repeatedly tinkering, and being inefficient.

But once you put yourself in the perspective of a manager or a boss to think about strategy,

and make requests and set goals, hey, your efficiency will instantly get a qualitative leap.

So your way of thinking quietly changes in this process,

you start to think, "how can I solve a bigger problem all

by myself with AI?"

This is a process of evolving from a "tool person" to an entrepreneur,

a resource owner, and a problem solver.

The second core feature is that the window of opportunity to become a master of AI is extremely short.

The so-called "master of AI" is not a new stable profession, but a, uh,

a fleeting and precious transition period.

Two years ago, "prompt engineer" was a legendary, uh, high-paying job,

but now few people say that anymore.

Maybe this year "vibe coding" is a popular trend,

but maybe next year AI will have evolved to the point where you don't need to "vibe code" like this anymore.

So, you must, at this moment, in this short window of opportunity,

quickly complete the primitive accumulation of personal knowledge, ability, and capital by mastering AI,

and completely transform your personal role before the window of opportunity closes,

from a commander to a true resource owner,

distributor, and initiator of ideas.

This is the whole reason why I'm obsessed with vibe coding at this stage.

So, what exactly is vibe coding?

And what have I done?

Uh, there are many definitions of it online, and this is my definition. Vibe coding is,

through natural language, combined with code when needed,

the process of collaborating with AI to solve practical problems.

Its core is to allow those who, uh, are not necessarily proficient in code or complex tools,

to also be able to harness the powerful capabilities of AI through clear logical thinking and natural language.

At the same time, it also allows people who are proficient in code to jump out of the narrow thinking of a "tool person,"

and use natural language to command AI to solve more essential and higher-dimensional problems.

For me, it has three specific forms.

The first is, uh, an application.

Here, it refers to creating an application with a fixed user interface to solve a specific problem.

For example, uh, our community has a monthly "coffee with a member" event, where members are randomly paired.

I used an AI tool like Replit,

to create a small application through vibe coding to accomplish this.

For another example, our community has always wanted its own private podcast.

So I vibe coded a, uh, web app,

that can not only host our own private podcast, but I can even open it up,

so that anyone can use it to create their own private podcast for free.

I will share this app with you in a future video, so please stay tuned.

Second, an automated workflow.

I want to say,

with vibe coding, you don't necessarily have to make an app. It doesn't have to have a user interface.

You can create a workflow to help you handle those time-consuming, laborious, and tedious tasks.

For example, let me give you an example of my own.

I used n8n and Gemini CLI to build a whole set of processes,

that can automatically convert the, uh, video of the community live stream into text, extract highlights and clips,

and export the corresponding short videos.

This process, which used to take hours of manual work, now only takes a tenth of the original time,

or even less.

What's needed behind this is not some advanced programming technique,

but my ability to describe the steps I want to achieve to the AI in clear language.

The AI will help me write a whole set of instructions, and these instructions are usually just textual processes,

not code in the traditional sense.

The same thinking can be applied to various places.

For example, when I organize and analyze community content, I use er Cloud

Code to create another workflow that can automatically crawl data from the community website,

and convert it into a format that large models can read,

which allows me to achieve many content insights that I couldn't even imagine before.

Recently, I've even started using AI to, uh, file taxes, do bookkeeping, manage notes, and assist with video production.

That's right, those, uh,

animated characters in my videos are the result of me and my editor creating AI workflows.

My goal is very clear: to let AI replace all those time-consuming,

but uncreative tasks, and focus my energy on what's truly meaningful.

Third, a single solution.

This is what we often call prompt engineering.

Compared to the first two, which are complex collections of multiple steps,

it focuses more on solving a single task.

For example, I have many prompts that allow AI to stably generate high-quality images, animations,

podcast scripts, or help me polish articles, handle daily community work, and so on.

Although these are just single tasks, because they solve practical problems,

they are also vibe coding.

More importantly, this ability to solve single problems,

is the foundation for solving grander and more complex problems in the future.

So, my definition of vibe coding is very broad,

but I think that's exactly what it means,

which is to solve problems with AI.

(Sound effect)

Here, I must remind everyone,

you will see many so-called "vibe coding" cases online, but in my eyes, uh,

many of them are not really vibe coding.

I'm referring to those, uh, experiments conducted to explore the potential of AI,

without any practical purpose.

For example, someone shows off that they can make a beautiful landing page in a few minutes, but the page itself has no function,

it doesn't sell products, nor does it collect customers,

nor has it been applied to a production environment to

save time and create value.

This so-called "trying" is a huge trap. It's more like a three-minute passion for a new toy.

Many people will mistakenly think that this is vibe coding, and after a brief try, they will declare, "uh, it's nothing special,"

or move on to the next hot toy. Is trying necessary?

Yes.

But we should skip the long, purely experimental stage,

and go straight to getting your hands dirty, to use AI

to solve a real, pressing problem that we are currently facing,

and let it directly generate value.

(Light background music)

Here, I must also shatter another illusion.

Web coding is not as simple as everyone imagines, as if you just sit in front of the computer,

drink coffee, say a few words to the AI, and a great product is born.

For me, or for everyone who has truly practiced it,

this process is actually more like a desert island survival.

When you really start, you will find yourself as if stranded on a desert island.

The problem you want to solve, the goal you want to achieve, is the distant continent that represents hope.

And between you are heavy waves,

these are the various difficulties you have to overcome.

At the very beginning, the tool of AI, for many people,

may be a small wooden raft that looks like it could fall apart at any time.

Many people, after seeing this small wooden raft and then looking at the distant continent, think, "Hmm, this is impossible,"

and so they choose to give up.

But I want to tell you, this is an illusion, because,

this so-called small wooden raft in front of you can be reinforced through your efforts,

and upgraded. It can become a speedboat, a ship,

or even an airplane that can take you across the ocean.

But the premise is, you must not give up.

When you encounter difficulties, you must remind yourself that there must be a way to achieve my goal.

AI can do it, others can do it, I can definitely do it too.

So, in the end, it all depends on how strong your will to survive is.

Only when your desire to solve the problem is as strong as surviving on a desert island, can Web coding

truly become a powerful tool for you to create great value.

And you, in this process of upgrading the raft and crossing the ocean,

have completed the ultimate transformation of your personal role.

So, how should we actually start Web coding?

The answer is actually hidden in my definition of it,

find a problem you really want to solve.

This problem can be big or small. It can be a, uh, mentally draining,

annoying chore in your work,

or it can be a grand vision that serves many people.

My advice is to start with your own specific small problems. To do this,

we must first abandon a deeply ingrained and inefficient learning model,

which is the so-called "learn the tool first, then solve the problem."

For example, many people want to become a YouTuber, and their first reaction is to systematically learn Final Cut

Pro.

Wanting to become a designer, they dive into Photoshop tutorials. That's fundamentally wrong.

The tool itself is not the goal, it's just the means.

And your real task is to create valuable content, not to become a so-called software expert.

The correct path is exactly the opposite.

It embraces a more efficient philosophy: get your hands dirty, go straight to the point,

and learn and master the tools in the process of solving problems.

Please remember, the problem is the core, the thinking is the core, the methodology is the core.

The tool always serves these three.

At this point, someone might ask, "What if I don't have any problems to solve?"

Then I would say, having no problems might be your biggest problem in the AI era,

because to some extent it shows that you are used to living for other people's problems,

used to passively accepting tasks, and having others think for you.

You are used to executing mechanically like a "tool person" in a narrow field.

This may be the worst situation an individual can face right now.

At the end of the video, I want to share a famous quote from Steve Jobs.

(Soothing background music)

I believe this sentence is more appropriate than ever today.

AI may lead the world to an unpredictable future, but as individuals,

the only thing we can do is to personally turn it into our opportunity. OK, thank you for watching to the end of the video.

I believe you have also discovered that, for me, Web coding is not a temporary content topic,

but a long-term career that I am determined to invest in.

In fact, I'm basically all in on AI now.

Through the sharing in the "A Bite of New Rice" experimental space over the past two months,

I realized how valuable it is to design such an AI community for ordinary people.

So, I worked with my creator friends to create a brand new community called "Generate with Love,"

to help ordinary people like you and me, entrepreneurs, and content creators,

to realize their own ideas and solve their own problems through Web coding,

and complete the transformation of their identity in the AI era.

The community link has been placed in the comment section.

Uh, this website was also made by me, a novice who doesn't know how to code, through Web coding.

As of the release of this video, the community has already started operating, and the community modules are fully equipped.

The high-quality courses we have prepared for everyone are also gradually being launched.

I am very confident in the value of this community,

This is because all the friends I co-created with are from the "A Bite of New Rice" community,

which in itself proves the value of the community. It can bring different people together,

and make the impossible possible.

So I believe that by joining this community, you will not only gain resources, but also new connections and new opportunities.

Now that the community has just launched, we have launched a special early bird program. If you join the community now,

and become our founding member, you will lock in the cheapest price for life.

As our courses are gradually launched, our prices will also increase quickly.

There are already more than a hundred friends in the community. If you also want to seize the opportunity of the AI era,

then please click the link below to join us.

Well, thank you for your support and trust. See you in the community, bye