Required Product for AI Product Manager: Systematic Thinking



First, experience and methods

1, the difference between experience and methods

Need to give directions to others, if the other is a person who will not see the map at all, we can only take a “landmark-based” way, a detailed description of all the landmarks throughout the journey, such a guide The method is simple and easy to use, but such a way of doing things loses their meaning once these landmarks are lost in a new environment. And if one learns to read the map, there is no hard road to him anymore.

“Follow the Landmarks” is our experience. The experience sounds complex, simple to be done, the learning cost is very low, easy to use and easy to use and effective, but rigid, it will quickly become obsolete as the environment changes. “The ability to see the map” is what we call the way of thinking. It sounds simple, complicated to do, extremely abstract, not easy to understand, more difficult to learn, flexible, versatile, and less obsolete.

People tend to learn experience because it is fixed enough and simple enough. But experience changes constantly and requires us to constantly learn new experiences. Therefore, if we only learn from experience and do not learn the methods, we are dumbfounded at the moment when the change is coming. Even if we learn a lot, there is still no strong ability to deal with the problems.

2, the relationship between experience and standards

In the field of e-commerce, Alibaba has survived, so Alibaba has shaped the standard of e-commerce. However, in the field of AI, there is no experience, there is no standard, who survived, who is experience, who is the standard. In order to deal with a completely new field, we must learn the method, just as we should know clearly from the perspective of walking in the northwest and south-east. Although the method is abstract and difficult to learn, it has a longer life span and stronger applicability. Once we master the method, we can Use your strengths to create experiences.

To distinguish between what we learn is experience or method, there is a quick and easy way to identify what we learned can be used across scenarios. In practice, most of the lessons we have learned are in fact wrong without the help of thinking methods. These wrong experiences will make us learn more silly. Standards are the product of successful experience. Do not use complex to pursue complex, but to use simple to explain the complex.

3, with a way to think of ways to replace the guess

To illustrate the problem, the teacher cited an example of Kodak film. Kodak film was what beat? Many people replied that it was a digital camera. However, to the surprise of all, the world’s first digital photo was even invented by Kodak. It was not a digital camera that beat Kodak.

In the face of the teacher’s questions, the present students gave a variety of answers, however, they were not accurate nor directional. From here we can find that in the actual life situation, most people first through intuition guess, and then through their own depth thinking verification. This poses a problem, many things seem right, but in fact wrong. When the snail denied the “digital camera” answer, it entered a cold session, followed by trainees guessing again and again. However, these assumptions did not help the real answer.

Faced with a problem, the correct solution should be the above logic in turn. The first step, combined with the previous experience delineated a range of thinking. The second step is to guess within this range, and keep narrowing the range by guessing.

In the face of the question of “what Kodak was defeated,” the first thing to think about is “what can defeat a company.” Such a mode of thinking relies on methods instead of relying on experience. Kodak’s core business is photo processing, and once the business is compromised, Kodak is defeated, so it’s easy to think of a product that has revolutionized rinsing. The advent of printers has brought new ideas to consumers, such as beat-and-shoot, and at such efficiencies, it takes less than a week for the films to be washed out of the market.

Second, know and learn

Half, one half, is more terrible than all

Most people’s questions seem right, but they can not stand the test. Individuals resolutely oppose the fragmentation of learning, if not all the points in a line, but affect the cognitive.

The use of information has two elements, access to costs and identification costs. Before the advent of the Internet, the cost of using information was mainly obtained. However, in the information explosion, the cost of identification has greatly increased. Especially for the older generation, they are in an age where the quality of information is high and the speed is slow, so the capability of screening information is very weak. However, the time is exactly the opposite (media are not re-emphasized in quality), so they are easily lost in mistakes In the information (deceived).

2, the product manager must have the ability to accurately define one thing to be able to reproduce

There is a difference between knowing and learning. Know that refers to the history of Apple’s hometown, learned that you not only know everything about Apple, you can copy to yourself. The teacher cited an example of hunger marketing by millet company. In the hungry marketing, which word is the most important? Hungry is the result, not the reason. Millet 1 2011 listing, htc with the configuration to sell more than 5000, but millet sold in 1999. Hunger is the beginning of all this. The key point of hunger marketing is that products should create a bright spot that allows users to thirst, far beyond the user’s expectations of products, reduce the supply of hunger point to maximum, and then lead to marketing. This logic can be used in anything. When we can use something else elsewhere, we learn it.

3, learned from the case of millet hungry marketing, how to apply

China Mobile has had such a case, in a region, some people spend 80-90 yuan, the main point of payment is 3G, 4G traffic, this part of the population accounted for 40% of the local population, how to consume per capita by the way of marketing To 120? Through the lottery to create a user-hungry highlights, the annual bill free of charge. Such activities far exceed the user’s expectations, and then through the withdrawal of only 10 individuals a month to reduce supply, the hunger point zoom. Finally, through the monthly bill over a hundred talents can pumping, and share to a friend’s virus marketing, the activities will be released.

Product managers to develop a habit, precisely define all the problems around, some of the phenomenon in the end is what? Even friendship, communication, can be defined. The definition of the standard is that in any one scene can be reproduced.

Third, systematic thinking

1, systematic thinking and fragmented thinking differences

There is a fish tank with fish inside, a camera at angle A and a camera at angle B (vertical to direction A and B). There are individuals in the house that can only see fish through A and B. Systematic thinking sees “A and B are different angles of the fish,” and the fragmented mindset sees “a mystical connection between A and B.” When such two people communicate, it is difficult to understand each other. Systematic thinking refers to looking at the problem, not only to see the performance, but also to see the back of the link.

2, physical thinking, from simple to complex cognitive system

When learning physics, the initial model of motion was “an object moving at a constant velocity on a smooth surface”; when friction was added, force and acceleration were introduced into the system; and in circular motion, we further learned velocity, acceleration , The force of three in a constantly changing movement system. If not step by step to learn, but to learn a circular movement from the beginning, will be very difficult. This is a systematic way of thinking, its process is from simple to complex (step by step overlay).

3, find the main contradiction in the physical system

System is composed of several parts of the interaction, with a specific structure, the specific function of the organic whole, these systems are progressive, the next system is an extension of the previous system.
There is a relationship between systems, the strongest relationship is the relationship between father and son.
When multiple systems are affected by a single structure, there must be an association between these systems.
The relationship and structure of systems are relative, and the direction that determines them is the solution to the problem.
Galileo throws two different quality iron balls on the Leaning Tower of Pisa, landing at the same time. However, throwing feathers and iron balls, the two will not be landing at the same time. The former main influence factor is the acceleration of gravity, the latter is the main influencing factor of air resistance. If we do not have a systematic mindset, it is easy to see only a single factor in this experiment and fall into a paradox. When making a product, you should use simple, pure logic as much as possible, with simple simulation of complex, rather than complicated with complex pursuit. Consider the necessity of negating general rules with a few examples, noting the scope of the product and the main problems to be solved, rather than refuting ourselves with factors that seem relevant but not actually relevant.

4, through the construction of the core system framework to explain the complex phenomenon

In the analysis of the target system, we must first determine the target system’s system framework, the core description is what? For example, the core system of Newtonian mechanics is (F = ma). The linear motion with friction is the smallest version of Newtonian mechanics. Continuously removing the weakly connected subsystems in the target system until the subsystems are very strong (threw the iron ball without throwing feathers), these subsystems constitute the core system (or the smallest system) of the target system. Observe the core system and define the structure of the core system, then the structure is the core framework of the target system. The system framework is not a fact, it is a theory we use to explain facts and there is always a deeper system framework. (That is, do not complicate the complex pursuit of complex, simple to explain the complex, the system framework is simple, we can abstract complex things)

Compared with AI, human ability is from simple to complex, abstracted ability, but AI is complex to complex.

In the field of the Internet, technology continues to close the bottleneck, there will be continuous business model has been developed. The essence of business is to buy and sell one, equivalent exchange, the birth of the currency instead of the general equivalent. Now there are free business models that exchange customers for free mode, making money by selling these aggregated traffic. The process of physical learning is like playing mahjong and playing chess, constantly observing the field phenomenon, and then extract the more general law of winning. When the discovery of new rules can not confirm their past rules is wrong, you have to continue to observe the test validation.

When we learn to think systematically, we still have to go to some old areas to borrow some system frameworks, abstract things can be reused, and then as a new world to explore the rules. Scientific exploration is an approximation that links far-flung systems to some basic theory. The underlying framework is reusable.

So the process of high school physics learning can be translated as:

When we want to learn a complex system problem, we must first learn the system of the smallest subsystem, when we have the smallest subsystem, the minimum subsystem to add a condition, resulting in a new system to learn and understand the new system. Repeat this process, and to ensure that the conditions of the system is still understandable, we can gradually understand more and more complex systems.

Fourth, how to apply physical thinking

Our understanding of the world is systematic and will not be constrained by any single dimension.

1, Erlong Musik – refining the system framework or the smallest subsystem

When you think, you should (as physics does research) evaporate the problems you encounter until the sedimentation of the most fundamental principle of the problem, rather than through analogy. Physics can really deduce many new things counterintuitively similar to quantum mechanics.

SpaceX example. Although material science and control technology has been rapidly developed, aerospace has not significantly advanced. Once the multi-stage non-recoverable rocket is a temporary solution that needs to usher in some changes. Musk confirmed the possibility of SpaceX through a series of inferences and calculations. Why we have not done so, because we are stuck in the past experience. Without careful reasoning, the rocket in the past under what conditions designed, and now this condition has changed? The sight of us all is shackled by the thinking of the rocket in the past, user research will never be able to find something new in the new scenario, the past experience is useless.

Why do Tesla, even if the charge is 50% loss, is still better than the gasoline engine. Engine utilization of fuel only 30%, but the power plant utilization of energy up to 90%, so that the electric car has a very large energy use advantages.

2, iPhone as the representative of the new mobile phone to get rid of Nokia as the representative of the underlying causes of the old cell phone?

The teacher first prompts us to think about how to think about this issue. Instead of setting a position, comparing two systems, removing all the same things, the rest is the answer, and we need to be able to learn from the answers.

After evaporating a lot of subsystems, we can find out that the difference between the iPhone and the traditional mobile phone lies in the physical keyboard and the virtual keyboard. The multi-touch input provided by the iPhone has brought revolutionary changes to the UI and interaction design, which is also a good source for the iPhone user experience.

The decision-making power of the product manager to the user experience is like a shopping guide selling the product at a premium to the elderly. Product manager’s design capabilities can be about the user experience, but fundamentally determined by the input and output devices. Old-fashioned touch screen is a single touch, the mouse is a single point, the experience is particularly like PC, a single point is not good enough to replace the keyboard, because the mouse with the keyboard is the best experience. So the final single-point touch screen as iPhone.

From the iPhone example we can learn:

User experience, interactive logic and input and output devices to match;
The difference between Mac and Windows is that the Mac supports multiple screens, one screen per software, and different underlying designs.
VR input is not mature enough, the input and output do not match.
Five, with physical thinking to see AI

1, when we are talking about product design, what are we talking about?

The standardization of product value, once standardized, can be copied; valuable workflow, and then the process of interface, and then reproduce; standards include two levels: the underlying standards and business standards; industry just appeared chatting technology, technology is almost The chat business, web2.0 are basically chatting behind the business, no one chatting technology, Internet business people are concerned about the business.

2, artificial intelligence in the end can do?

Dry things innocent things dry, dry people do not want to do dirty living. Look at the X-ray, driverless. The biggest change that deep learning brings is that machines understand and handle abstract concepts.

3, natural language processing in the end is what?

Natural language processing is a language of interaction, NLP is a new way to interact, we can enhance the understanding of NLP by analyzing the evolution of the GUI.

Gossiping robots are like Windows desktops, which allow you to do more than Dos, do not need to know the command line, and you can see many folders. We need to add a program icon to the desktop. The application is currently represented by the new NLP The most interactive lack of things.

4, Chatbot technology itself has basically been ready today (enough, does not affect us to do demo verification), the key is the need to add more program icons (app). That is, today is not the lack of technology, but to use the right scene, the lack of App, product managers have to learn to assess the scene (not landing

No Comments

Leave a Reply