Reading experience of "The Supervisor's Strongest Work Technique": When AI can give the answer, why does the supervisor have to define the problem himself?

*He writes to people who lead teams, and I write to people who don’t write procedures. It went on the market on the same day, but answered the same question. *

We published books on the same day, one for people who lead teams, and one for people who don’t write procedures.

On the morning of September 2nd, I stared at the cover of “Zero Basics Vibe Coding (opens in a new tab)” that had just been put on the shelf for a long time. It was my 21st book. On the same day, my junior fellow student Peng Jianwen’s fourth book “Supervisor’s Strongest Work Skill: Learning Management from TSMC, 18 systematic thinking tools will help you do it right, do it faster, and do it better!” ” was also published on the same day.

Well, it’s hard not to think about this coincidence for a few seconds. The themes of these two books are very different. He is written for supervisors who are chased by problems every day, and I am written for people who can’t even distinguish between front-end and back-end. The readers of the two books have almost no overlap, and they probably will not be placed together on the bookshelves. But after I finished reading his book, the more I thought about it, the more I felt that we were dealing with the same thing.

Five years ago, I just finished teaching a class for a supervisor at TSMC and opened his book.

When I opened the new book, I remembered a record I had written.

In August 2021, I wrote in my diary that I would give lectures to several TSMC executives that afternoon and give them suggestions for their presentations. Two months later, I wrote an article Reflections on Reading “The Yield of Thinking”. The first sentence at the beginning was: I just helped a group of TSMC executives conduct educational training. In the past two days, I started reading the latest work of Mr. Peng Jianwen, the former TSMC Operations Efficiency Manager.

Today, five years later, I am reading his fourth book. In the meantime, he also published “The Process of Thinking” (opens in a new tab), and I also wrote about the OGSTM Strategy Tool he introduced on the website.

Putting these time points together, it has been more than five years. In the past five years, his Pinshuo innovation has grown rapidly, but there is actually only one thing to do: take apart TSMC’s set of things that no one can explain clearly, layer by layer, and turn them into steps that others can follow.

The scene in the prologue explains the gap between grassroots and supervisors.

The prologue of this book is very interesting, so good that I think this chapter alone is worth the price of admission.

He wrote that his first position at TSMC was as a production management engineer. Once there was an issue with the production line and the product could not be shipped. He immediately notified the equipment engineer to deal with it and the production line returned to normal. He said that if he had to grade himself at that time, he would give it 80 points or above.

Then he walked into the supervisor’s office to report the good news. The supervisor nodded and asked two questions:

The product cannot be shipped and the customer’s delivery date will be delayed. Have you notified the customer immediately?

Since delivery will be delayed, do you have a plan to catch up with the product?

He stood there unable to answer.

The reason why this section is great is that it doesn’t make sense, it only gives pictures. In his eyes, the problem is a production line failure, and repairing it is a perfect score; in the eyes of his supervisor, this matter involves at least three aspects: production line repair, customer communication, and delivery catch-up. He only did a third of it.

The book closes this scene with one sentence: The grassroots see the problem from one point, and the supervisor needs to see many lines from this point, or even the entire surface.

I paused when I read this. Because this is not just a matter for supervisors, anyone who takes on cases, starts classes, or runs a one-person company makes the same mistake every day. The client said that the website was broken, but we fixed the website and felt that we had taken care of it. As for what to do with the activities that customers originally planned to put on the shelves today, whether to proactively inform them of delays, and how to avoid them next time, they are all out of sight.

Three levels of training: do the right thing, do it quickly, and make good methods without relying on heroes

The skeleton of the book is what he calls three layers of practice, and 18 tools are hung under these three layers:

The first level, high efficiency, do the right thing: First confirm whether this is a real problem that should be solved. Tools at this level include AIG quantitative description, structured 3 questions, same-benchmark comparison method, Why-Why analysis, and funnel model. He put it very bluntly: If you do something wrong at this level, no matter how fast you do it later or how much you invest, it will just be a waste of effort in the wrong place.

The second level, high efficiency, do things quickly and do it right the first time: This level includes smart problem-solving TS method, data insight skills, layer disassembly method, comparative difference analysis method, 3×5 Why, and strategy 3-level method. Among them, 3×5 Why traces from the occurrence source, the outflow source all the way to the system source. I think it is the hardest chapter in the whole book.

The third layer, efficient innovation, so that good methods can be copied: This layer jumps out of the individual and deals with the organization: reverse thinking dichotomy, comprehensive thinking system strategy map, harvest mentality management method, PRA Model, rigorous logic 18 aspects, and finally collected in OGSTM strategy to undertake.

In the author’s preface, he gave a definition of innovation that I like very much: Innovation is not just about inventing new things, but about making good practices no longer dependent on a single powerful person, and can become a system that the team can copy and continuously optimize.

This sentence is worth thinking about for those who lead teams: the problem of many teams is never that there are no masters, but that only that master can do everything.

What I feel most about is the first floor, because a one-person company is most likely to die here.

Among the three levels, the first level is the one I read the slowest.

The reason is simple. For the efficiency issues on the second level, I can rely on tools; for the organization issues on the third level, to be honest, I don’t have much use on my own. But on the first level, “What problem are you trying to solve?”, there is nothing that can help me fix it.

The first chapter of the book talks about getting rid of “almost” management and using AIG’s quantitative description to change returns from feelings to measurable. The so-called about the same refers to sentences such as “Customer complaints seem to have increased recently” and “The results this month seem to be poor.” It sounds like it’s describing the problem, but it says nothing at all: How much more? Since when? What type of customer complaint? Compare with whom?

I checked myself. Last month I actually said to myself, “The email open rate seems to have dropped recently,” and then it passed. If that sentence is changed to a description with time, numbers, and comparison benchmarks: from which letter did the open rate start to drop, by how many percentage points, and whether it dropped from new subscribers or old lists, what to do next will almost automatically emerge. The difference is not in hard work, but in whether the problem is described to a level that can be taken action.

This is the most dangerous thing about a one-person company: no supervisor will ask you a few questions after you finish your report. If you say that the production line has been repaired, no one will really ask you whether you have notified the customer.

The two books answer the same question: AI removes the threshold for doing, leaving the threshold for thinking clearly.

When I write this, I go back to the coincidence at the beginning.

He wrote in the author’s preface that nowadays, AI is becoming more and more powerful, information is getting easier and easier, and many tasks can be done faster, but what managers really cannot outsource is thinking and judgment. He also added: AI can organize knowledge, but it cannot walk through nearly 30 years of practical management for me.

What I wrote on the day the new book was launched is: It’s not that I suddenly learned to write programs, but that I discovered that as long as I can explain something clearly, AI can accompany me to make it step by step.

Put these two sentences together, and the answer comes out.

What AI really changes is the threshold for “doing” this thing. In the past, if you wanted to build a website, you had to find an engineer, and if you wanted to analyze a batch of data, you had to be able to write a program. Now, as long as you explain clearly, the machine will help you complete the task. My entire book is about how this threshold disappears.

But after the threshold disappears, what remains is more naked: what exactly do you want to do, why is it this and not that, and whose problem is it intended to solve? In my book, I call it “making something clear”, in his book, he calls it “defining the problem.” The name is different, but the stuck place is exactly the same.

I even think that because AI makes execution faster, first-level errors become more expensive than before. When you go in the wrong direction, you used to work on it for three months before you realized it, but now you finish the wrong thing in three days, and then do another one, and then another one. Speed ​​is an amplifier, it will amplify what is right and what is wrong.

This is why I have always said that the biggest threshold for AI product development is not technology, but speaking clearly. He walked over from a supervisor’s perspective, and I walked over from a hands-on perspective, and we finally met at the same intersection.

Who to recommend to?

If you have just been promoted to a supervisor and do not have a framework in hand, and rely on intuition and overtime work every day, this book will be very useful. It breaks down the method and you can try it at work tomorrow.

If you are a mid-level executive who has had enough of the same questions popping up over and over again, the second level chapters are worth reading slowly.

If you work alone like me, my suggestion is: jump directly to the first level and read the six chapters first. You can put the third level aside and read it later. Because, if no one will ask questions behind you, you have to be the one asking.

---

Further reading:

• Reading experience of “The Yield of Thinking”: Strengthen four kinds of thinking skills and develop business thinking
• Make good use of OGSTM strategy tools and build an action blueprint
• The biggest threshold for AI product development is not technology, but speaking clearly
• “Zero Basics of Vibe Coding” is released today: that moment of “It turns out I can do it too”

---

📚 Book Purchase Link: “The Supervisor’s Strongest Work Skill: Learn Management from TSMC, 18 Systematic Thinking Tools to Help You Do It Right, Do It Faster, and Do It Better!” 》 (opens in a new tab)