Li Xiang, founder of Ideal Auto, said: "Anyone who spends just 3 days installing Claude Code will open up a new world."

Recently, I’ve been using Claude Code Opus4.6 more and more, and I strongly resonate with this statement. From my perspective, it truly opens up a new world.
The more I use Claude Code, the more I feel: there’s nothing that can’t be done.
This is the confidence that AI gives people.
Because, in theory, as long as we ask questions precisely enough, there’s no problem that AI can’t solve.
Every time I use AI to solve a problem I previously thought was unsolvable, I become even more convinced of this.
The more I use it, the stronger my belief becomes, and the more willing I am to use it. The positive feedback I receive increases, and my confidence grows, creating a positive flywheel.
In the future, the world will be divided into two types of people: those who use AI and those who don’t.
Among those who use AI, there are two categories: those who use Claude and those who don’t.
Recently, at the Daily Innovation Club, I’ve been leading my friends to use AI. The first task I assigned everyone on the first day was: don’t ask anyone, search and research on your own, install Claude Code, and use the Claude Opus4.6 model max version.
I truly believe that if we’re going to use AI, we should use the best large model; otherwise, it’s just a waste of time.
Because, for the same question, a good model and an average model will give vastly different answers.
As a result, the feedback we get is also vastly different.
Some people think AI is amazing and powerful, and they want to use it well.
Others think AI is garbage and that the results are not what they expected.
Of course, the results depend on how you ask questions, the prompts you use, and the constraints you set.
But first, fix the factor of the large model—then worry about everything else.
Otherwise, you’ll be stuck in self-doubt all day, lose confidence, or make excuses: AI is not good enough, it’s not me.
Once you use the best large model, there’s no external reason left.
If the answers are bad, the results are poor, or the code doesn’t work, it’s all your own problem.
Only then will you give up and focus on yourself—improving your logical thinking, aesthetic sense, fundamentals, and so on.
The compound effect of AI is much greater than we imagine.
The earlier you start using Claude Code, the sooner you can accumulate a sense of AI, understanding, and confidence, reaching the compound inflection point earlier and achieving rapid progress.