Original video in Chinese.
AI really can help you make money, as long as you find the right approach.
Right now I’m using AI to help me make money from three directions: content, investing, and cross-border e-commerce. In this episode, I’ll use an example from the investing side.
I remember that after OpenClaw blew up at the beginning of the year, quite a few people used it to build what they called investment intelligence systems. In other words, they would gather a bunch of information every day and then generate a report. But the signal system I’ve recently built with Perplexity only runs once a month.
Why design it this way? Aren’t you afraid of missing something important?
Because I invest in the AI industry chain. In an industry chain, the variables that really determine the long-term direction don’t change very fast.
For example, it usually takes years for a new production capacity line to be built and put into operation; it also takes many years to complete a power grid expansion and grid connection.
These things that truly affect the core assumptions are slow variables. But prices and news are fast variables, jumping every minute, with new headlines every day.
When you use the frequency of fast variables to track the truth of slow variables, the result is that you treat a pile of noise as if it were a signal.
So low frequency isn’t laziness; it’s actively aligning with the rhythm of the industry itself.
What I asked Perplexity to build for me is not a system that summarizes news every day, but a system that verifies investment assumptions every month. It keeps helping me answer one question:
Have the core assumptions behind why I bought into this industry chain changed?
The actual method is very simple: write down the assumptions first, then verify them.
For example, I believe demand for computing power will continue to expand, I believe electricity will become a long-term bottleneck, and I believe scarcity in the industry chain will keep shifting.
I write these judgments down in advance so that fuzzy intuition becomes a sentence that can be proven wrong.
A good signal system starts not with data, but with assumptions. Data is only a tool for testing assumptions.
After the assumptions are written, AI looks for evidence from four angles every month: industry data, capital flows, technological changes, and market sentiment.
At the same time, it distinguishes between official data, reasonable inference, and mere market rumors. When there isn’t enough evidence, it doesn’t force an answer; instead, it marks it as “under observation.”
In the end, this signal system only needs to answer four questions:
Where is the money flowing?
Where is the bottleneck?
Who has pricing power?
How long can this situation last?
If the answers haven’t changed, then I don’t need to change my judgment because of daily news and price swings. If the core assumptions really do change, only then will the system remind me to research again. That’s the role of a low-frequency signal system:
it doesn’t predict tomorrow’s gains or losses; it stops market sentiment from making decisions for me.
I’ve posted a more specific introduction to the system, as well as the reports it generates, in Knowledge Planet newtype. Everyone can go take a look.
Back to the question at the beginning: if you want AI to help you make money, one decisive prerequisite is that you need to have something of your own.
The investment signal system I just introduced was built for me by AI. But the basic premise for this is that I have my own investment strategy and my own understanding of investing in the AI industry chain. The signal system is just a concretization of those strategies and that understanding.
If I had nothing and understood nothing, then making something that gathers news every day and generates reports would definitely be completely useless.
As I said in the Planet: AI is an amplifier. Whether AI can help you get results still depends on your ability.
OK, that’s all for this episode. If you want to understand AI, want to reclaim individual sovereignty, and want to find like-minded people, come join our newtype community. See you next time!