AI Coding: A Meta-Ability
Yesterday, while introducing our product to a friend, I suddenly realized something.
Underlying technology is evolving rapidly—how rapidly? It's changing by the day. But the general public—even many white-collar professionals—are adopting these tools far more slowly than imagined.
Take a recent example. OpenClaw went viral, and everyone was drawn to this technical breakthrough. Someone asked me: "Are you using Claude Code? Is that for helping people write code?"
I said yes.
Their first reaction was almost identical: "This sounds complicated. I probably won't need it in my lifetime."
Not What You Think
I demonstrated in the most intuitive way: installed in 5 minutes, just need an account to get started, and all you need is the ability to speak.
They stared at my terminal for a while, then fired off a series of questions. Is installation tedious? Will it be difficult? What problems can it solve?
I said, installation is just one line of code, 5 minutes. An account costs just a few bucks a month to get started.
They were stunned. "Is that really it? I thought something this powerful would definitely have a complicated installation."
It isn't.
What's even more surreal is OpenClaw itself. Online, there are circulating 15-step, 20-step installation guides that look headache-inducing. Many people complain: "OpenClaw is great to use, but installation is too difficult." On Xianyu (a Chinese second-hand marketplace platform), services even appeared offering door-to-door installation for 500 RMB.
I demonstrated it for them on the spot. When your computer has Claude Code installed, you only need to say one sentence—"Help me install OpenClaw on this computer."
It will find out what OpenClaw is, where to download it, and how to deploy it on its own. You tell it what the Feishu (Lark) App ID is, and it will help you configure and test it until it's running. You go grab a coffee, and by the time you get back, the deployment is done.
One line of code, 5 minutes. That simple.
A Meta-Ability
Someone asked me: "Does your software have out-of-the-box applications? Like text-to-image?"
I didn't answer directly. I was thinking about something else.
If we let everyone master an efficient AI coding tool—a coding agent—whatever you want, just have it install it for you.
Text-to-image? You say: "I want a text-to-image application, download the model and deploy it locally, so I can generate images with a single sentence." It might take one or two hours to help you download the model, deploy it, and get it running.
This capability has already transcended the concept of "applications."
It's helping every ordinary person—with extremely low barriers to entry—truly take control of their computer. By "control," I mean the level of control that exceeds that of 90% of engineers and senior engineers. And your computer is connected to the internet.
So that old mindset of "I need to install an out-of-the-box application" is actually unnecessary. Whatever you need, just tell it. It will do it for you.
What Can It Do?
Honestly, if I had to use a computer without this kind of software now, I wouldn't know how to use it.
It can utilize 90% of the capabilities on your computer that you normally don't use.
You have 20 Excel files to process? Put them in a folder, throw it to the agent, tell it how you need them analyzed. It will invoke Python, download plugins itself if they're missing, and solve problems itself when it encounters them.
Want to implement some interesting network functionality? It can help with that too.
As long as the model is strong enough and fast enough, things that originally required professional skills become within reach.
What Does the Future Look Like?
In the future, people will become increasingly lazy when using computers, unwilling to get their hands dirty. I rarely actually operate the computer myself these days—at most, I open a few Claude Code instances, switching between them to have them do different tasks.
Work patterns will change.
Before: Human → Tool → Computer.
In the future, probably: Human → Agent → Computer.
What's the point of building a complex interface? It's a burden for AI, and humans don't look at it either. Just give instructions directly to the agent and let it execute. Done.
Software infrastructure might change. I'm not sure what it will become, but the general direction is probably this way.
The Difference Between Claude Code and OpenClaw
Many people ask me this question.
Strictly speaking, OpenClaw is a layer wrapped around Claude Code's capabilities. Some people call this layer the "lobster shell"—I think that makes sense.
What Claude Code does: you give it a vague instruction, it executes end-to-end, basically meeting expectations during the process without needing you to correct it, can run for dozens of minutes or even hours, and finally gives you a result.
What OpenClaw does: adds another layer on top—distributing instructions, managing results, forming long-term memory, like a project manager or supervisor. It can interact with you at a higher level, becoming your personal assistant.
But the premise for all this is: the underlying coding agent must execute beautifully on every task, without you having to watch over it.
This is also why OpenClaw connected to different models is an entirely different species.
Currently, there aren't many models that can truly "execute well on a vague instruction." I've only seen Claude's Opus, and possibly Codex's latest model, capable of this. Versions connected to other models feel like toys—no guarantee of execution each time, making the coordination layer above even more difficult.
The reason OpenClaw amazed everyone is because it was connected to Opus. But Opus's coding plan isn't available now—too expensive.
This is the current pain point. I don't know when it will be resolved.
Originally published at https://guanjiawei.ai/en/blog/coding-agent-meta-ability