I recently read a 2019 article "The Lesson to Unlearn" by Paul Graham, which touched me. As the founder of Y Combinator, Paul Graham not only invested in unicorn companies such as Airbnb and Dropbox, but more importantly, many of his views have profoundly influenced the entire entrepreneurial circle. He is nicknamed the "Godfather of Silicon Valley". In this article, he makes the point that the most important lesson we learn in school is precisely the lesson we need to unlearn.
What is this that needs to be forgotten? In fact, it is "examination ability = real ability" .
In school, we are accustomed to this logic: good grades = good students = strong learning ability. But in reality, there may be no relationship between test scores and real learning.
Many people may understand this, especially in college. How do we prepare for exams?
Study past exam questions and find out the pattern of questions
Memorize test points instead of understanding the principles
Practice your answering skills instead of thinking deeply
Within a limited time, pursue the highest “input-output ratio”
The ppt before the exam and previous exam questions are always the most important. We may even be very annoyed because the teacher suddenly changes the questions this year. If we compare a course to a software system, we have actually been trying to crack the exam part of it. So as students, we were all hackers. As for the school, it can only rely on this indicator.
Paul Graham put it bluntly: So what college admissions is a test of is whether you suit the taste of some group of people. (Therefore, what college admissions tests is whether you suit the taste of some group of people.)
This problem apparently exists all over the world. The better the school, the more powerful their students are. If you're not smart enough and can't hack, you can't go to a good school.
Or to put it more bluntly, anything that can be quantified can be hacked, but the difficulty is different. Just like there was a joke before: As long as there are enough attributes, anyone can be the first.
The trap of hacker thinking
If we only stop at criticizing exam-oriented education, the point of view of this article is not novel. In fact, "hacker thinking" has indeed helped many people achieve good results in school and later in work.
The key question is: What impact will this way of thinking have when we try to step out of the established track and try to create new value?
Especially when making products, we often unconsciously look for systems that can be "hacked". This way of thinking seems smart, but is actually dangerous.
Paul Graham shared a typical case: a product manager proudly showed how he doubled the number of users in a month. However, after three months, 90% of users were lost.
Isn't this just a replica of the "pre-exam assault" we are familiar with? An important point mentioned in "Lean Startup" is to distinguish between real user growth and false user growth. To put it more bluntly: You have to know how capable you are. You can lie to your friends, but don’t lie to yourself.
Why does this kind of hacker thinking fail in the product field? Because the product life cycle does not end with an "exam", users are not teachers who give you grades. This kind of thinking not only damages the long-term value of the product, but also causes us to miss real opportunities for innovation.
Of course, some people may say at this time: "I just want to make some money as a side job. Do I need to think about it so complicated? I think they have already made money."
This is where reality gets awkward, and we will discuss this issue in the next section.
The Dilemma of AI Products: Making Money and Product Power
Making money has nothing to do with product strength. To a certain extent, there is even a so-called "negative correlation". This is even more obvious with AI products.
Take a look at the current AI entrepreneurship circle: various media exposures, financing news come and go, and all kinds of talented young people are born. However, product homogeneity is serious.
First, the threshold for AI products has been significantly lowered. Since the emergence of ChatGPT, various products that simply package GPT API have been launched one after another, and AI+all things products have emerged in an endless stream.
Almost everyone can quickly develop an AI product. As we can see: chatbots with shell APIs, various AI writing assistants... these products have short development cycles and low costs.
Various large models have sprung up, but the performance difference is not obvious. They are all chats in a dialog box. If I don't tell you which model you are using, you may not have any idea. This leads to a more serious problem of homogeneity of applications in the upper layers.
In the early days when everyone generally had insufficient knowledge of AI, this made it easier for the market and users to be fooled by superficial gimmicks. These products may have a good wave of profits in the early stages (usually through financing) .
This type of "quick money" products have a common characteristic: they are good at catching immediate needs and shallow needs. It's like chasing a hot spot. In addition, a large number of marketing methods are used to package themselves. In this era of information explosion, it is indeed easier for such hot-button products to gain attention and be monetized.
But the problem is also obvious:
Highly replaceable: Anyone can do it, otherwise you wouldn’t worry about being replaced by big model upgrades every day.
Short life cycle: the hot spot cools down as soon as it passes.
One-shot deal: Many products cannot withstand scrutiny. They look good but actually do not work.
Even though their problems are well known to everyone. But those teams that are really thinking about the value of AI applications will find it difficult to prove their value in the short term. These products, which are quickly copied and packaged, can gain more attention and revenue in a short period of time. In many cases, these teams are forced to make such applications.
I often ask entrepreneurs or other AI developers a question: "What is your purpose of making this product?"
If you just want to make money as a side job, you don’t have to worry too much about product capabilities. Those so-called “hacker thinking traps” may not be so important. But if your goal is to build a truly valuable product, just chasing short-term gains is too shallow.
Make money while standing
Standing still, making a "good product" may require a lot of product ideas and observations, but it may fail on the road; if you want to make money, make a "hot product", which may seem low, but it can make money.
So, is it possible to make money standing up?
Such products are what we call great products. It usually has to meet several conditions:
1. Ambition
The product really aims to solve problems and benefit users, and it is a product with a high ceiling. A simple case product can also solve the problem, but its upper limit is also very low.
2. Long-term thinking
In the short term, they may not be smart, do not play tricks, and do not pursue quick results.
3. Ultimate user value
Importantly, great products always focus on one core question: Do users really benefit from this product?
We often joke that B-side products only need to satisfy the paying party, and the opinions of users are not important. But for a truly great product, users must really feel the benefits, and word-of-mouth can bring about spontaneous growth.
write at the end
Paul Graham believes that the best way to get investment is not to learn how to persuade VCs.
Because the reason why VC invests in you is very simple: it is a good investment.
They usually judge by the number of users, so how to obtain users? It’s not about being exposed through various channels, but really having an excellent product. When you have a good product, everything becomes simple: users will actively spread the word, investors will actively search for it, and the market will give positive feedback.
Of course, not every product can become a great product. Some products are destined not to be great products from the first day of design.
Not everyone wants to design great products either;
Not everyone can design great products either;
Not everyone can afford a great product.
And the great products belonging to the large model era may still be hidden in the darkness before dawn.