On June 15th, Zhou Haijun, the instructor of the Dark Horse Growth Camp and the founder of 360 Company, once again came to the dark horse class to share with the students the latest views on Internet and entrepreneurial innovation in the past year. He believes that the prospect of intelligent hardware is still optimistic, and the future intelligent hardware will be closely integrated with artificial intelligence, and the intelligent hardware will no longer continue the free mode of the traditional Internet, and it is really profitable by the hardware itself.
The following is a recording of Zhou Hongyi's lecture:
I still have confidence in smart hardware.
Some time ago, there was an article that made a summary of the industry and believed that intelligent hardware is not the future direction. I think we can sum up what went wrong, such as whether the user's needs are not accurate, or whether the problem is in the hardware or in the cloud service, but it can't be because there is no special cow intelligence yet. The hardware company will deny this industry.
Now everyone says that smart phones are the most needed and largest market in smart hardware. However, when you look back, Steve Jobs launched the iPhone in 2007. It’s almost 10 years now. You think about Android’s development in the past seven or eight years. Going back, many people don’t see iPone. Not the first smartphone, you will find that after 20 years of development on this road, the result of today's smartphone is blooming. Today's smart hardware has only been developed for up to three or five years. I believe that smart hardware is still in 2008 and 2009, when Android came out.
I still have confidence in IOT's Internet of Everything, but I need a little patience. It will not be like the squad of the year, and it will recede like a tide overnight. On this road, I respect the companies that have failed. The market must be followed by many people, and the lessons of failure are invaluable wealth. The whole industry is the loser who gave birth to the later winners before they can move forward.
For smart hardware, everyone’s past ideas are wrong.
Smart hardware has not made such a huge market like a mobile phone. I have also been rethinking recently. In the past, everyone's thinking was wrong.
The first mistake: think that networking an item is smart hardware. This idea is wrong. Many smart hardware products are made. In fact, consumers do not have this demand. For a while, our family's fans, washing machines, refrigerators, because of the addition of a microprocessor inside, can have some simple program control functions, what computer washing machine, computer microwave oven is used overnight, selling more expensive It is said that this thing is purely computer controlled. Finally, everyone understands that it is to add a microprocessor to make some functions a little more convenient, but it does not bring about a real change. Did you say that after connecting this thing to the Internet, is it smart? Is it information home appliances? Is everything connected? For example, if you go to Japan to buy a rice cooker, what is the point of remote control of a rice cooker at home? Who is going to glutinous rice? Who poured the rice into the pot?
You will find that blindly connecting anything to the Internet is still a problem: when you are intelligentizing an object, you should think about it. Under what kind of scene do you solve the problem for the user?
The second mistake: We feel that putting a smart chip on a thing, there is an operating system inside, or if you have to rigorously assign an APP to a hardware, you think this is called smart hardware. In fact, this idea is also wrong. You have a physical device with an Android smart system that only means that it has this capability, but after you're really connected to your phone, you'll always have to figure out what scenarios users will use.
I am talking about an example. We also did a smart router. It did a lot of interesting functions. It also took a lot of effort to make an APP for the router, but later found that the router is set up, once it is set, the router is away. Your distance is very far, usually placed in the corner of the wall, placed under the computer desk, or placed behind the cabinet, you don't have much time to open the app. We still have a lot of hardware, the setup is not convenient, in order to let users open this app, it is said that this is smart hardware, it is against the user's nature of the attempt.
For another example, someone made a universal remote control, used infrared to remotely control home appliances, and then used the phone to control the hardware. There are also some people who do smart bulbs and smart switches. When they wake up at night, they first touch the phone, find the app, find their own lights, click and light the lights. Don't be ridiculous, including me, those of us who are smart hardware, this is an inevitable attempt on our road to innovation. To cite this example, you have to think about what kind of mistakes we made, smart and intelligent, and APP for APP.
Real intelligent hardware must be combined with artificial intelligence
I think that the real intelligent hardware must be closely integrated with the brain artificial intelligence brain and artificial intelligence services. In fact, intelligent hardware is more than a window and a bridge between intelligent services and users. Because the CPU of intelligent hardware itself cannot have too much artificial intelligence, the real artificial intelligence should be placed in the cloud, whether it is voice or not. Recognition, speech understanding, image recognition, in this way, intelligent hardware has the ability of cloud intelligence, it can really be called an intelligent hardware.
So, smart hardware, smart two words, in the past we thought it was Android, or Wi-Fi inside it called smart hardware, today I think this view is wrong, it must be with a real artificial intelligence Cloud services are tightly integrated. When the intelligent hardware really has this ability, it can produce revolutionary changes. Otherwise, it will be the same as the computer washing machine with the addition of the single-chip microcomputer in the same year. Finally, everyone found that it is still a washing machine, which cannot produce a real revolution.
We have to ask ourselves whether you really bring fundamental changes to users. Smartphones are truly revolutionary. With smartphones, they have replaced many devices in the past and solved various problems. But if a lot of hardware today, if only add a smart chip, just connected to a network, this is not a big innovation, or even a revolution, it is just an increase in the selling point.
Let me give two examples. We have also explored smart hardware and paid a lot of tuition, but we have not given up. For example, the smart watch we made, at the earliest, the function is very simple, is to let parents use, but children do not like to use, because children are particularly happy new and old. Later, the phone function was added, and the child can call the parents at any time. It turns out that the Apple Watch is just a vassal of a mobile phone. It is said that the next generation of Apple watches will be separated from the Apple mobile phone. However, children can't call their parents too often. We realized later that we still have to make fundamental innovations on our watches, so our latest version took about a year to re-do. Again. This version of the watch has just been released, and I have worn it for two children, so that children can do it.
We also have an artificial intelligence team that connects the watch with the voice recognition function and the chat function, so the child can talk with this watch. For example, how high is the child asking about Everest? Why is it so blue? Where did I come from? We can summarize the questions asked by children every day and send them to their parents to let parents know what the child is thinking. We created a very narrow but very practical scene that allows kids to communicate with this watch. After truly combining it with artificial intelligence, this product will give the child a feeling that is completely different from the original product.
Similarly, we are also doing car recorders. Car recorders are just needed, but there aren't many that really work, because not every day there is a scrap. We put a camera on the car recorder. This camera can see what happens on the road. On the one hand, it can lay a foundation for autopilot. It can remind you of the distance from the car, or if anyone suddenly crosses the road in front of you. Or, there is a sudden obstacle in front of the prompt, because the CPU's ability to respond is still faster than the human's ability to respond. I was thinking, if it can judge that there are some abnormal conditions on the road surface, you can automatically take a picture of the road surface, or take a short video to pass, we can get a lot of road information. So, I'm rethinking smart hardware recently. I used to say that the time to reinvent the wheel is here. That's right, you can reinvent the wheel, but not simply add a smart chip inside.
Driverless cars are the most difficult in intelligent hardware. Its navigation service is definitely in the cloud, but the response speed of the car is very fast. The current 4G network and even the 5G network may not be able to be reached. In case the network card is a bit unimaginable.
To do auto-driving, put a number of computers with strong computing power in your car. How to reduce it now, this has not been solved. All the trunks of the self-driving car are stuffed with stuff because they have to put a lot of computers. But the car is actually an exception. The smart hardware we usually refer to is some simple intelligent hardware.
I also very much agree with Kai-Fu Lee's point of view. In the future, after the wireless Internet, it must be a smart Internet, and it is artificial intelligence. You will find that the combination of artificial intelligence internet and intelligent hardware is a particularly natural combination. Why? In addition to the intelligent functions I have just mentioned, you can make your hardware products truly revolutionized. There is also a point that many smart hardwares are different from mobile phones in that they do not have the control interface of mobile phones.
Why has Apple's Siri been so many years ago that it has not become a mainstream application? Because the easiest way to get your phone out is that you don't really talk to your phone. The best way is to touch your finger. For example, many people always fantasize about the mobile phone. I want to order a meal today. Many young people order food today. He didn't think about what to eat. He opened a book ordering app. The first page of the recommendation may have attracted his attention. . Including when you want to book a ticket, you will find that touch, click, and slide are the most direct communication methods at this time. However, many smart hardware may not have a touch screen, nor a keyboard, nor a screen.
For these intelligent hardware, the most advantageous way is the interaction of voice, the intelligent hardware can be seen, the interaction of voice enables the intelligent hardware to understand, and of course the speech synthesis, really need to do, still need a lot of details. For example, watching a mobile phone in a car must be the most inconvenient. Voice must be the easiest, but many people have voice functions in the car. Why not use it? You will find that artificial intelligence is not doing well. Sometimes when you talk to a computer, you will feel like a fool. You have to wait for it to finish. You can say that this is not the case in real life. You will not have the patience to wait for it, one round at a time, one round. To communicate. Therefore, to do artificial intelligence well, it still takes a lot of real details.
Artificial intelligence is not a new thing, it has long been proposed, but the original computing power can not be done, it is not true artificial intelligence. For example, there is a child, you say this is a cat, he can remember at a glance, this is the real human intelligence. You want the computer to know the cat. It is not enough to show it a photo. You have to show it 1 million. The next time you look at the 1.01 million photo, you can recognize that it is a cat, so you need big data. In the absence of the Internet and the absence of big data, the data base does not exist, so it is necessary to improve the cloud data and improve the computing power. The old deep learning algorithms play a role in some special fields, but they are not so mythical, and it is impossible to change the world in one to two years. But I think artificial intelligence is also good, smart hardware is also good, in the future world of interconnected things, these two will have a huge space.
This is my reflection on smart hardware. Everyone is doing smart hardware and may think it is too simple. There are many cottage factories in Shenzhen. It is really amazing. It has been doing for many years. The hardware speed is very fast. There are also many sales in a year. Although the profit margin is not too high, at least the hardware is quite capable. However, many teams in Beijing don't know how to make hardware. After the APP is matched, there are not a few things to sell. With the crowdfunding, they immediately call out the investors. Many companies become unicorns, and the valuation is hundreds of millions. . I don't understand. Compared with the company in Shenzhen, you don't have a supply chain yet. You don't have the ability to purchase. You just have to do more than others. The result of this app is not meant to me. Use your app. Isn’t this bubble bursting for granted? If such a company says that it has succeeded and obtained a high valuation, this is not right.
Polygonal Steel Pole, Polygonal Steel Electric Pole, Polygonal Steel Light Pole
YIXING FUTAO METAL STRUCTURAL UNIT CO.,LTD( YIXING HONGSHENGYUAN ELECTRIC POWER FACILITIES CO.,LTD.) , https://www.chinasteelpole.com