Original link: https://www.lidingzeyu.com/real-impact/
In this episode, I interviewed Cao Xiang on Bilibili. Cao Xiang is currently in charge of the AI Platform and the next generation of creative products at Station B. Before this interview, I had heard about Cao Xiang’s unique personal experience from many listeners and friends, including his experience at MSR, Lenovo and his own business.
In this issue, our starting point is how to proceed with the career path after the PhD in HCI Human-Computer Interaction. Cao Xiang believes that the important point of career planning is to constantly find the direction with the smallest angle between it and his own goals. And he thinks that he is very lucky that he knew what he wanted very early on. Once the goal is determined, the so-called career planning can be simplified into searching, or waiting for the emergence of such a minimal opportunity.
We also discussed, how to make real impact? How to combine scientific research with making money? Of course, if it cannot be combined, how to make money? We also talked about how hard it is to change the DNA of a company, and thus the car that software companies overturned trying to enter the hardware industry.
Finally, I still have to ask a topic I am very interested in: traffic password, how can I make this program recorded by Cao Xiang and I more popular at station B? There are no unexpected answers, but at least I tried.
Well, if you are a doctoral student, an engineer, or a researcher, you should be able to extract ideas suitable for you from Cao Xiang’s sharing.
Youtube: https://www.youtube.com/watch?v=KZQ3cbJpj54
Bilibili: https://www.bilibili.com/video/BV14B4y1q7WQ/
Text Version: Scroll down
Audio version Search “Li Ding Chat Room” in your podcast app
for example:
– Apple Podcasts https://liding.page/apple
– Podcast Casts: https://liding.page/pocketcasts
– Overcast: https://liding.page/overcast
– Spotify: https://liding.page/spotify
Guest: Cao Xiang Bilibili https://caoxiang.net/index.html
Li Ding’s job search list: https://www.liding.page/jobs
Chat with me for half an hour: https://www.liding.page/1on1
0:00 Notice of this issue
3:20 Research done by MSR and its current implementation
6:40 The experience of the Mavericks startup
20:20 How to make REAL impact?
26:10 The company’s genes
32:20 Important decisions in Cao Xiang’s career
42:42 How to make money?
50:04 Preview of the next issue: Differences in workplace communication between Chinese and American companies
Production team
* Li Ding Zeyu
* Pengdi Zhang
* Guo Yu
* Lin Xuhui
* Zhang Tianjin
Next program preview
In the next episode, I chatted with Chen Ran from Tubi/Bitu Technology about the differences in workplace communication between Chinese and American companies. Chen Ran shared a lot of his experience and Tips. Chen Ran is also a big v on Zhihu, so he has a lot of experience in communication and sharing.
https://ift.tt/0QmnpX1
In terms of language, we talked about how to use the sandwich method to express denial comfortably. Why do many questions in English are actually imperative sentences? How to deal with the inflation of positive words in the North American workplace?
We also discussed some misunderstandings caused by the “cultural” differences between China and the United States. Should you directly ask your manager for a promotion and salary increase? How can you communicate more efficiently with your colleagues 1:1?
This is a very teaching experience sharing program
Interested listeners can subscribe and turn on notifications to be notified as soon as possible.
Text version full text
Li Ding: Cao Xiang, hello.
Cao Xiang: Hello, Li Ding.
Li Ding: I read your homepage before, and your experience is very interesting. Would you mind sharing your high level picture mental journey?
Cao Xiang: Because I may really mean that the background is more complicated, that is to say, I have done it in academia, and then in the industrial world. Because in fact, after I graduated from my PhD, I worked in MSR for about five years, and then I also worked at Lenovo for a period of time, but doing it there may be more inclined to not academic research, but more of this kind of research. This innovation in industry. In addition to starting his own business for more than five years, he joined Station B last year. Therefore, it may have been done in academia, some in industry, and in entrepreneurship. Now it may be done more directly to serve users.
Li Ding: Then we can ask backwards. Now your recent work at station B is actually very interesting to me. I also asked it in our audience before, and everyone has this kind of question in various aspects. One of the questions is that you are at station B. Because I don’t see that what is written on the homepage is particularly specific, I would like to ask what type of work you are doing at station B?
Cao Xiang: I am currently in charge of two projects at station B. One is a team similar to the AI platform that most traditional companies are doing. Then it will provide support for AI-related capabilities for various businesses of the company, such as machine vision voice and so on. Then the other one is the company’s one, which is considered an innovative product. It will specifically target the up master of this B station and provide some common support that is not available on the market. This is still in the research and development stage.
Li Ding: In fact, it sounds like, at least for example, as you said at the time, compared to the work in MSR, it must be more about products in terms of implementation. Because you said that both of them seem to have nothing to do with academic research, is that so?
Cao Xiang: Yes, in fact, when I came out of Microsoft, I should say that I am not particularly active to do this kind of academic research. But I have been paying attention. Every year, I will watch these videos of CHI VIS to find some inspiration.
Li Ding: Why did you switch, say, after MSR for a few years?
Cao Xiang: Actually, I think there may be quite a few people who learn HCI, and they may have this kind of restlessness in their hearts to make products. I think it should be because of HCI itself, its entire training. In fact, it is to put it plainly. He is teaching you how to design a product, and he wants you to consider it from the perspective of user experience. Whether it is from a technical point of view or a design point of view, including a human psychological point of view, the whole set of training, and even user research, the whole set of training is actually teaching you to think about a product. thing. Well, since you have this training, of course, you can’t help but want to make a product yourself, which is normal.
Li Ding: Do you think that your research at that time, for example, whether it was research done in PhD or research done in MSR, when you converted to this kind of product like you said, for the user century, How to do the conversion? What’s the challenge? Or some of the lessons you learned.
Cao Xiang: I think the whole point of departure is not the same. That is to say, even in academic research, HCI should start from the user experience. But I think one thing in common with all scientific research is that when you are doing scientific research, you are more likely to push the boundary, that is, to find new possibilities for new technologies. Not only the possibility of technology, but also the possibility of design and so on. So in fact, you may not pay special attention to this thing, how much market value or economic value it will actually bring. More to tell you, there is such a possibility, there is such a possible experience. So of course, research is also a common point, that is, it must be very new, that is, whether this new is a novel, that is the most basic point of whether your paper can be published. But when you are actually making an actual product, you may not be so obsessed with every point of it, whether no one has done it, but more about the problem I want to solve, the most suitable solution for him What is the way. If it is someone else’s existing way, it is suitable, I will use other people’s way, if it means that there is no existing way, I have to come up with a new way. So its whole starting point is still not the same.
Li Ding: This is indeed true. Many times, when a product is made for users in the industry, it does not care whether it is the latest technology, as long as the user experience is the best, even if an outdated technology is used, as long as it can guarantee users The experience is very good, in fact, this is not much of a problem. We then went back and asked, like you were talking about some of your experiences at station B, what did you do with your startup after five years?
Cao Xiang: At the time of the startup, our company name was Xiaoxiao Niu, and our technology was called Wonder Painter. In fact, it can be understood as a bit like a realistic version of the magic pen Ma Liang, just draw a picture on paper, shoot it with a mobile phone or whatever camera, based on this picture, it will be converted into a 3D animation, that is, it will be automatically For 3D modeling, it may be automatically given motion to it, or it may be combined with some preset animation scripts. Let him use the picture you drew to perform an animation script, or even put it into a game, you can control you to draw this picture to become a game character. The whole is such a set of technologies, which will be applied to such scenarios of various interactive media. For example, at that time we would talk to LG in the United States and South Korea, then they would use it for interactive animations on this kind of TV. Including cooperation with some domestic education companies, may do some such interactive teaching and so on. Of course, many users may actually be children, because children like this kind of painting the most, so basically integrate their own creations into the content we want to convey to them, so that the entire content can be There is more of an immersion. Probably such an idea.
Li Ding: Actually, you are already answering my next question. The next question I want to ask is that since you are a start-up company, as you mentioned just now, you actually want to give users a good experience in the end. As for your users or customers, you just mentioned LG or this kind of education company, so your main customer is this kind of company, or your main customer is actually an individual user, that is, individual user children or their parents .
Cao Xiang: In fact, we have also tried different approaches. In the beginning, what we actually did was to directly face individual users, that is, we made it into an APP. There are also some supporting this APP, because it is some drawing board brushes, or even plasticine, which is used for painting. Pack it into an interactive toy, then go directly to sell it to such consumers. But it is actually quite difficult to find out, because you will find that most of the time here is actually no longer doing the technology itself, but actually how to package the toy. At this time, it will involve a lot of links, including production, including this kind of logistics, including many links of sales, and I find it difficult to focus on really doing the experience itself. Therefore, it was later transformed to more of a ToB idea, providing technical solutions for those who have interactive media needs.
Li Ding: So I started from ToC. Later, I found that there might be too many things that I didn’t want to deal with or things that I couldn’t handle, and then I started to work on ToB. Does it make any difference when working with such a large company? Just follow your instructions directly to individual users.
Cao Xiang: If you talk about common ground, you really need to understand your users or customers, that is, what they really need. If we talk about understanding an individual, it may actually be the set of things we learned in HCI at that time, doing user research and analyzing him. But if you understand the demands of an enterprise, that’s another matter. What you really need to understand is how he makes money and what kind of business model he himself is. Can your technology help him, or to make more money, or to attract more users, or to save costs, etc.? It is said that you actually need to understand his industry, so after working for so many years, although you have not really entered the content industry and the education industry, you will still understand a lot more than before, at least There is a lot of dealing with people in these industries. Of course, we all know that the domestic education industry has been relatively miserable in the past two years. Many of my friends back then are not having a good time now, but at least I have dealt with them at the time, so I can better understand some of their difficulties. Including station B now, it is actually a relatively typical content industry. In fact, Station B is not a very typical Internet company, it is more like a content or cultural company operating on the Internet platform. Well, this thing is actually very related to the technology we used to do this kind of entrepreneurship at that time. To put it bluntly, although it seems that my general direction may jump from academia to industry, to entrepreneurship, and now I am back to industry, but in fact, there has always been a line called creativity support according to the HCI circle. It is a technology used to support creativity, so this line has been doing all the time, including some papers in this area at Microsoft at that time. In fact, this is what I did when I started my own business, so now I am doing this at station B. thing.
Li Ding: I have two follow up questions. One is that in the process of converting from individual users to enterprise users, you must have fewer trivial matters that you should handle yourself. You don’t need to deal with logistics and other things, just leave it to the enterprise to handle. Has your company’s revenue increased? Or less? Or the same as before?
Cao Xiang: In fact, the revenue is very dependent on your ToB business. It will be very dependent on your specific customers, because of how big his own customers are. In theory, of course, the bigger he is, the more likely he will be able to pay for one and you. The ability will also be stronger. The other is that, including him, the international market is actually not the same as the domestic market in this regard. To be honest, at this moment, at least at this moment, the domestic market still does not recognize the premium of this technology. In contrast, in the international market, they will pay more for this thing because of its technical content. This thing has relatively formed a tradition, and this aspect still needs education in China. So it is to sell to foreign companies to make more money, at least if the unit price is usually better. But of course, ToB actually has a big challenge, that is, your entire state will be very dependent on the state of your customers, that is, when the entire industry of your customers is not good, you will also be indirectly affected. . So in 2020, after the entire epidemic came, in fact, all of these, whether it is the education industry or some offline entertainment industries, are actually greatly affected. In fact, for us, the entire market is actually much more difficult to do than before. So basically the company stopped at that time.
Li Ding: I don’t know if the audience actually knows about the Wonder Painter app after hearing this. At that time, I forgot whether it was on YouTube or SIGGRAPH. I saw one of your demos, and I thought it was very interesting. It’s also a coincidence that I can chat with you on the spot. At that time, your company, Xiaoxiaoniu, was a one-person company, or did you also build up a team?
Cao Xiang: There are about 20 or 30 people in the company. In fact, although the sparrow is small, it has all the internal organs. That is to say, it is not in a pure sense, it is just the initial stage, it is purely just doing technology to wait for acquisition. This is really doing business by myself, so there is development, sales, then market, then marketing, and then HR finance, needless to say. In the end, the whole company is actually a very complete company. Of course in the early days, when there were only a few people, all the dirty work had to be done by yourself.
Li Ding: Just like me now, I am doing this channel by myself. Many things are done by myself anyway, and now there are some friends who help.
Cao Xiang: So I have worked in all the positions that can be found in a company.
Li Ding: When you joined Station B, because you were the founder of Xiaoxiao Niu, when you joined Station B, Xiao Xiaoniu did not exist? Or is it that Station B acquired Xiao Niu? or something else?
Cao Xiang: This company is currently out of operation.
Li Ding: So those colleagues then went to find new jobs?
Cao Xiang: Yes.
Li Ding: The second question I just wanted to ask is, you also mentioned that Bilibili may be less like an Internet company than other large domestic companies, it is more like a content creator. company of. Actually, I think I don’t really understand what this means? Because station B is a User Generated Content, it is a UGC thing. In theory, these contents are all users. What does the content of station B mean?
Cao Xiang: I think this may be understood from several angles. That is to say, one refers to Station B. In fact, it has a relatively obvious cultural tonality, which means that it is a completely general content platform compared to YouTube. So station B is actually, everyone who knows station B knows that station B was originally a two-dimensional community, and then slowly began to break the circle. But in general, station B still has a relatively clear position on the tonality of these contents of the entire culture, that is to say, it places great emphasis on positive energy, including creativity. I think this may be the kind of company that is completely different from the kind that is only used as a platform. In fact, there is still a difference. Then I think another point is that station B is not just a content platform, station B is actually doing a lot of PGC stuff, that is, station B is investing in animations and documentaries, which means that it is actually an internal There are also many people who are really cultural. So the whole so why does he feel like he’s more of a cultural company.
Li Ding: Just now you talked about the experience of the Mavericks in the past five years. In the end, why did you think about it, why did you think of looking for other opportunities? Why choose station B?
Cao Xiang: In fact, the reason is very simple, or because I have been doing this kind of creative-related things. In fact, the entire positioning of station B is very dependent on the creativity of the up master itself, so he has a very strong demand for support of this kind of creativity. This means that he needs to do this, but conversely, it means that he has been working on this kind of creativity for many years, regardless of education, technology or product, and one thing is that you only need to go It is not enough to be a creative technique or tool. Because you don’t have a wall, and there is no creator’s ecology behind you to support you, it is actually difficult to create an ecology based on tools after it is made. Or it is difficult to create an ecology based on a technology. But if there is already such a creative ecology, it is very smooth to use tools to further empower it. Well, station B just has such a soil. I think it should be the most dynamic creator ecology in China.
Li Ding: Knowing your logic is equivalent to saying that you have made a tool that no one uses, so you hope to have a platform to support it. Whatever the tool you make, more people can enjoy the creative tool, maybe my understanding is biased?
Cao Xiang: In simple terms, this is the case, but it is not just a platform in the technical sense. In fact, it is more important to the group of people behind the platform.
Li Ding: Actually, I feel that many companies that are like station B and achieve this scale will have this problem. He definitely doesn’t want to be a standalone gadget, he definitely wants to be integrated with his ecosystem ecosystem. Including a lot of software made by our company, it is also the software you make videos that can import Illustrator, and the After Effects you make in the later stage, you can also import it into different software. This is also an ecosystem imagined by our company. Of course, it means that because our company does not do a content platform. For example, I don’t have YouTube. Go for collaboration on this platform. So I feel that station B is also considering something of this type of platform, whatever that means, it may be that users may be tools, some kind of platform. The next question, this question may be a question collected in the audience. Many young or relatively senior researchers, including engineers, have heard a lot of words in the company, into make impact, many of them are impact driven and have influence. In this kind of academic lab and this kind of industry lab, including different paths in entrepreneurship, what does this kind of impact look like? What’s the difference? What kind of common ground does this have?
Cao Xiang: I think this is a particularly good question. I see that you originally wrote the title of make real impact on your side.
Li Ding: We can just say real impact. I missed the word just now.
Cao Xiang: I think this question is very interesting. Because I think in fact, it could be anything. In fact, as long as you are sincerely doing one thing, you are actually making real impact. That is to say, I think everyone may say, but it may be more entangled whether the impact is big enough. Just to tell the truth, it’s like we are chatting here today, as long as one audience hears and feels inspired, it is real impact. It’s just that it may not be a very large impact. Then on the other hand, in fact, whether you are doing academics or products by yourself, or in a company, you are doing academics. I want to assume that I am doing academics because I really like doing academics, not because I am forced. I think your original intention It must be said that I feel that I am making a fun thing. You do something that others have not done, and then inspire others. This is a very real impact. If I make a product and someone uses it, it will be real impact, so I don’t think this point is particularly bothersome. But the real possible challenge is whether your impact can be scaleable, and whether it can be scaled. In fact, this matter is not that easy. In fact, it means that whether it is an academic system, an entrepreneurial system, or a large company system, in fact, he answers this question in different ways. For example, in the academic system, we have this kind of conference, journal, and the entire publication system, that is, the purpose is to let others see and see if you make something meaningful to you arrive. If you rely on a large company as a platform, then because it has a large enough user group, it means that it makes sense for the company to make this product. Then it can be used in a relatively rapid stage, so that more people can use it. It is the same as why startup companies need to raise funds. If I just want to make something that I use with my friends, I can make it beautiful, but there is no way for him to make it available to more people. Because if I can rely on my own profits to roll, and many users may be in a hurry, then financing will allow me to quickly make a copy for him. Well, in fact, he answers this question in different ways, how to amplify something that you think is at least real impact. But I can especially understand why people keep entangled with this problem, because many times when you make it scaleable and scale it, you will find that you may go astray. It seems that you and your original intention to do real impact are not the same thing anymore, that is to say, you are being led astray by this system. For example, when you were doing academia, you initially only wanted to let everyone see this paper for me. Everyone knew what a cool thing I had done. In the end, I became chasing how many papers I posted every year. Then the same reason, for example, in a large company, that is to say, what you may have thought is that many people use this thing like me, and it turns out that you are pursuing a KPI of the company and pursuing promotion in the company. But a big company, by the way, a big company can provide you with a very good scale platform, but you have a problem with being a big company. One thing that has to be considered is that every company has its DNA. It means that if what you want to do matches the company’s genes, it will be easy for you to promote and implement. If his genes are mismatched, it’s actually very difficult. So when you are choosing a company, you must think, really first understand what the company’s genes are, and whether the genes are suitable for you to do this, because once a company’s genes are established, it is almost impossible to change its genes. impossible. Then a startup company may be a benefit in this regard, that is, its genes are your genes, and your genes are defined by you, but entrepreneurship and companies actually face the same problems. That is to say, you are only working on this product at the beginning, and then you are very happy. When you are doing it, you are financing for financing, financing for survival, financing for investors, and supporting your team. You will find that It may gradually mean that you spend more of your mind on these things, which means that on the one hand, you find that your scale is expanding, and on the other hand, it seems that what you are doing is no longer the one thing you originally wanted to do. So why everyone is talking about not forgetting the original intention is because he is very difficult. So in fact, I don’t have this question, and I don’t have a clear answer. I can only say that it may be constant. Every night, I think about whether I am still doing the thing I wanted to do. If not, maybe consider doing it another way.
Li Ding: I feel that I very much agree with the two points you just said. One is that they may find that they are not doing what they want to do, and the other is that every company has genes, which I also fully agree with. For the first point, I would like to add a framework of this kind of thinking. In the last episode, I happened to be talking with Jiaming, the founder of another startup company, here to talk about the difference between the means and the purpose, which is to use back, for example, what you just said. Your little calf’s entrepreneurial experience, at the beginning you may want to make a good APP, then write this APP, this is your purpose, just write this APP. Later, with the expansion of the company, as you want to sell it to more individual users, your goal may become to sell to more users, and then your method may be to write APP, through recruiting Man, through these things, through this transformation of means and ends. It’s a progression that’s not about who did what, but a natural progression. So when writing an APP becomes an action of writing an APP, you can think of it as writing paper, or doing research. When your goals change, your original intention is no longer your original intention. It’s just a means. At this time, you will find out whether you will optimize it, such as asking others to write the APP, or rushing to do some work, because this is only part of the goal, it is not the goal. I feel that the thinking framework is also very applicable here, but I feel that this is actually a very unsolved problem, because as you expand with the company, whether it is a company that does not strap, or if you take VC money Well, if it is still a big company, there will be such a driving force, whether it is internal or external, that will push you to constantly update your goals. It is difficult for you to say that my goals have never changed for ten years. Your company or yourself will definitely say that your goals should change. Once the goals change, the so-called original intention will be difficult to maintain.
Cao Xiang: I personally feel that I have to make a trade-off, that is, you still want to understand what you care about the most. Or to put it another way, what have I thought about in my life, and when I die, I think why I have never lived in vain. That is to say, think about this matter, and you may not care a little more when faced with some choices.
Li Ding: I also very much agree with what you said about genes. The genes of each company are really very different. What impressed me the most is that every time I use Google’s hardware products, I find it very difficult to use, because I feel that this company is a very software-driven company. I don’t know if it’s because the transition from software to hardware is really hard, but the gene is there, so every time I touch any Google hardware I find it really hard to use, and I have to Switch back to another company with more hardware genes.
Cao Xiang: Yes, there is a huge gap between software and hardware thinking, because I have touched both sides a little, because although I was doing research at Microsoft at that time, many things I did at Microsoft were related to equipment. Including my own company later, I also made physical products. Although it cannot be called hardware in the strict sense, because it has no electronic components, it will deal with the entire production chain, and I have played a lot. The ideas of making software and making hardware are largely opposite. Making software is about a quick iteration, small steps and quick iterations, because the cost of software iteration is extremely low, and its only cost is the consensus of the programmers. So to speak. And the cost of hardware iteration is extremely high. This means that for my version, I design a version of the circuit version, or if I open a mold when making a machine, if there is a little problem in it, it will be completely scrapped. It may be dozens of Tens of thousands of dollars will go to waste. So it’s impossible for you to do something as casual as software, iterating on very small versions. Therefore, he will be extremely cautious when doing it. Of course, there is another thing to consider about the hardware. That is to say, the software does not need to consider mass production. I just need to make the software, and then copy and paste it. . But the hardware I make a prototype, and I can make the factory mass production are two completely different concepts. That is to say, there are many things that I can make by hand, and many things can’t be made by the factory, because he has to consider whether the mold can support it, and his whole is so fine that when I open the mold at this time, can that piece be pulled out? come out. These are all things you don’t need to think about when you’re doing hardware and prototyping. So from making a single thing to being able to mass-produce it, you have to even consider packaging, and you have to be able to withstand the test of logistics. All these things are really if you only do software, you really have no concept at all.
Li Ding: Since you have done it on both sides, for example, will you want to engage in more opportunities in this kind of hardware in the future? Or do you feel that your background is more suitable for software?
Cao Xiang: Personally speaking, I am very open. That is to say, it may be back to that sentence. When you are young, you are called a team, and when you grow up, it is called this platform. Its own genes are not there. It is because the things that an individual can do are actually very good. Individuals can do research. I can solder a board, build a component, and send a paper. But to be a product, it is very dependent on where the capabilities of the entire organization are. So it still depends on where I am.
Li Ding: Just now you mentioned that what you did at MSR was research-oriented, then started a business, and then went to station B. There are actually several big ones. I think it is the so-called life choice. , or a career choice. What do you think are the most important moments? Do you think you made the right decisions looking back on those decisions now, or could have done better.
Cao Xiang: To be honest, I really don’t say that I was stimulated one day. It’s really not the case when I suddenly think about changing my path. I think it may be more about the general direction I want to do based on the state at that time and time, which way I think I do and which method I want to do. Because I want to make a general direction, in fact, basically after I graduate, maybe two or three years after graduation, I think more clearly. All I want to do are things related to creative, so maybe under that circumstance and opportunity at that time, I think I should choose a path that has the smallest angle to the direction I want to go. Basically, it is unlikely that you will encounter a road that is 100% in the direction you want, but you can find the road with a smaller angle. When it comes to entrepreneurship, in fact, it is similar to what I said at the beginning, saying that all of my HCI people may have a restlessness to make products. I think basically this sentence can also be translated into, all doing HCI have a restlessness to start a business.只有因为创业的时候，你是可以百分之百决定你做什么产品的，当然就是其实也说了，因为HCI 的训练，我觉得相比之下，在就是至少跟计算机整个大方向底下其他专业的训练相比，它是比较偏全面的一个训练。因为它本身跨学科，跨技术，跨设计，跨心理，就是他已经相对完整的教会你去想怎么去设想一个产品了。但是他并没有教会你怎么去卖一个产品，他也并没有教会你怎么去开个公司，怎么去管理一个团队。这些东西依然是你只有真正去做才能学到的。包括刚才说的，就是说如果过了一段时间以后，发现真的想去做跟这种，如果想让真正去服务到更多的创作者，那么还是需要去跟一个已经比较在那儿的一个创作生态去结合，那么这也我觉得都是比较在机遇下一个顺理成章的选择。他还真不是说特别大的一个思想转变。就像你说的，可能更多是目的没变，手段一直在变。
李丁：说的更加的实际一点，也切回我们今天这个主题，也是很多听众也可能很感兴趣的。因为毕竟我感觉我的听众大部分都是属于职业早期，最多到中期吧还不会像你这么senior 的。像对于这种类型的这种听众的话，你觉得他们应该怎么去想自己这么长的一个职业规划？
曹翔：其实我觉得实话实说，我觉得一方面其实，我一直觉得我自己人生就没什么规划。但是一直知道自己喜欢什么，所以就是说碰到机会的时候就去挑跟自己喜欢的夹角最小一条路，基本上是一直在做这件事。所以你可能看具体每一条路，它并不一定那么直，看起来会可能会非常蜿蜒曲折。但是你把它zoom out 的时候，你可能还是看到方向还是在那的。那么其实我觉得不管自己是不是心里有一个特别明确的方向，但是说事实情况，现实情况就是说外界，就是你即使心里想的很稳，外界也不会那么稳。就是说整个外界的环境是不断的变化的，不管是一个方向，它火不火。好比说HCI 至少在我入行以后，我觉得已经经历了好几次，火了又不火，这是小的专业的火不火，然后大的环境那就更不用说了。我想两年之前，应该没有人的职业规划里会想到有疫情这么回事。但是事实它就是直接的影响了所有的职业走向，所以我倒是觉得可能不需要特别纠结于说我长期一定要说我这步怎么样，那步怎么样。但是我觉得可以做的一件事就是说，你始终在为更多的可能性去做准备，而不是说完全只是在做我此时此刻做的这件事，需要我做的那些能力。我这件在做这个阶段的时候，可能有意无意的在为下个阶段再做准备，再培养一些下个阶段需要的一些能力。当然说你不可能完全预预测下个阶段是什么了，但至少是说你想着说，我能够扩充一些其他方面的东西。当然真的可能到下一个机会来的时候，你可能就可以很自然的去把它抓住。
李丁：我从哲学层面我特别同意这个观点。就是说在做现在的事情的时候，在做好现在事情的同时，当然能够为未来做准备是很好的。从执行层面来说，我觉得挺困难，像你说的这很困难，你没法预测未来，而且你现在的工作，也不一定真的就能够给你这样子的机会，去往一些比较新奇的方面成长，可能他就真的是只能让你做现在想做的事情。比如我现在自己，比如说我PhD 当中做了很多这种仿真，这种fabrication，这种声音方面的非常hardcore graphics 的这种东西。我入职了之后发现我其实对video 这种方面更感兴趣，我可能对这种纯虚拟的这种simulation 并不是特别感兴趣，我现在就等于说我在往video 这种video tools understanding 方面转。其实像你说的，我也不知道未来这东西能不能够就给我铺一条路，但至少这个东西是我感兴趣的，我感觉现在我做的开心的，至少短期优化是好的，未来希望的像你说的，如果能有一个夹角比较小的机会出现，无论是公司内部还是公司外部，这个的话会给我一定的，比较独特的一个视角和背景专长。
曹翔：我觉得还有一点说，跟我可能刚才说的，可能是一个硬币的两面。就是说就一方面你可能不断去给自己找一些新的能力，但是另一方面其实你之前做过的每件事，即使你已经不再做它了，其实都是你的一个家底一个财富。因为其实我自己，其实我Master 我硕士的时候其实学的是计算机视觉，当然是很传统的计算机视觉了。因为很久以前了，后来我发现了有HCI 这个方向，觉得HCI 真好玩，我就做HCI，然后就转了。一开始其实非常想跟计算机视觉划清界限，但是后来发现根本就划不清界限，就是我后来做的很多的事情，不管做PhD 做的一些东西，包括后来自己创业做的东西，都是因为我有CV 的背景。所以我能够想到，我能够从HCI 角度想到一个东西，但是我真的用自己的方式把它做出来。这个东西可能是很多，没有背景的HCI 的人可能他就是做不到，但是他们有一定有他们自己的优势，有他们的背景。所以就是说包括我现在在B站，其实还是在帮公司做一些AI 的事情，就是实际虽然心理上好像一直想跟CV 划清界限，但实际上从来没有划清界限过。所以以前做的每件事真的都是有帮助，所以不用纠结。
李丁：其实我刚在你一开始讲到你在公司引领AI 这个平台的时候，当时我就其实有点困惑，因为我当时就看你的比如paper 也好，或者说你做的事情也好，我感觉跟HCI 可能更接近一点点，跟这种general AI 其实感觉没有特别大的关系。现在你说出来了，okay 是make sense 的，这都make sense。你刚才说到之前做过的每件事情都是一个加分的点，这其实跟之前Steve Jobs 讲的一个talk 也有关系，大家可能talk 记得最多就是stay foolish stay hungry，但是它其实还有几个其他的bullet point。一个you can only connect the dots backwards，你只能在经历过这一切事情之后，无论你是功成名就还是碌碌无为也好，你才能会去看这些dots 是怎么联系到一起的。你很难有一个对未来的预测，说我预期到就这12345个点将会很好，所以我就做。如果你真的这样的话，你就是一个先知了。现在咱没有先知这种东西的存在，所以说大家都只能look backward 试图去分析我怎么成功的，我怎么失败的，你怎么成功的，你怎么失败的这样子。 Hopefully 这些我们抽象的过程中，可以帮助到未来的人帮助自己。
曹翔：而且实话实说成功失败，这个东西永远都是个概率。就是即使你把所有事都做对了，你也不一定就一定能成功，一定还是最终都都是打个概率。所以我觉得可能有些人可能一成功就开始写自传，就开始说总结，真的你再来一遍你还能成功吗？不一定的。但是把所以所有这些自己的一些能力去增强，把自己心态放长一点，就是从概率上从统计意义上讲，你的成功概率会提高的。
李丁：这个挺好的。我看到今天这个话题其实差不多了，我感觉大部分我们没有挨个的聊，但是我感觉主心骨是聊到的。最后的话你有什么最后想跟听众分享的吗？
曹翔：我看还有一个问题没问是怎么赚钱。
李丁：对，我对刚才没串进来，但是我们也可以问对怎么赚钱，你觉得科研跟赚钱怎么结合起来。当然如果结合不了的话，只讲如何赚钱也可以。
曹翔：这个问题问特别好，我觉得特别实在。科研和赚钱怎么结合起来，我觉得你没法刻意的去结合，就是因为它本身它的出发点完全不一样。刚才也说了科研是为了去探索未知，赚钱那就是为了赚钱，赚钱的前提就是有人愿意买单对吧。所以说如果我想做一件事的时候，就想着我又要科研又要赚钱，基本上是做不到的。最后做的大概既没科研也没赚钱。但是它虽然方向很不一样，但是方向不是180度。它是所以说有的时候可能真的就是机缘巧合，你会发现我做的这件事，碰巧科研做好了以后发现他还有赚钱的机会，那就正好结合起来了。或者说我在做一个商业的产品，做的过程中发正好发现我需要解决一个科研问题，才能把它给搞定，那就结合起来了。所以它不是一个事先能规划的事情。然后只讲如何赚钱，实话说我也没太弄明白，如果真弄明白了，我可能今天就不是这个身份就在这聊了。
李丁：今天你就小小牛上市创始人的一个身份了。
曹翔：但是我觉得你把它回到特别基本的逻辑，其实特别简单，商业的基本逻辑是什么。其实我觉得就三句话，一句话叫便宜买，第二句叫贵卖，第三句叫多卖。这就是特别简单，但是每一句做到都很难。就是说当我说便宜买，这个是一个广义的买了，就是说不只是说我花钱是去别人那买东西，包括我自己去去投入成本去做研发，包括人力成本这是这些都叫买。我怎么能够做到便宜买，基本上你是要有个得天独厚的优势，你才能做到比别人便宜的，或者是我在这方面已经有非常多的积累了，或者说我有垄断优势，我有什么独特的人脉。但是我觉得对大部分人来说，是可遇不可求的事情，那么更多的时候你能够得到，你能够以比别人更低的成本去做成一件事，基本上就是因为你有积累，积累包可能包括你个人的积累，包括你专业的积累，包括你可能人的积累对吧。然后贵卖，我怎么能把这东西卖得比别人贵是吧。当然最好的方式就是垄断，全世界只我这一家，当然就可以，理论上我可以想卖多贵的卖多贵。你做不到垄断的时候，可能就是说要把东西本身体验做得比别人好，可以卖的贵一点。但这事很难，就是因为体验这个东西其实它很难量化，而且比别人好一点，你不一定能够多卖很多。因为消费者都是很精，什么东西能够让你相对来说卖的贵的比较多，就是所谓的品牌。所以为什么marketing 很重要，品牌的溢价很重要。为什么苹果会比别人卖的贵那么多，它不光是好用，它真的是说很多人，我说现在可能没有那么严重了，至少在前些年我手里拿个苹果，我带个手机壳没问题，最后一定要把苹果标露出来了，让别人所有的别人知道我是苹果党，这个品牌的溢价是在那儿。你怎么做到这个品牌溢价其实也很难，其实需要前期大量的积累，是需要可能市场的一些投放的策略，然后建立消费者对你的认识等等其实也很难。然后要卖的多，怎么卖的多，其实这个你会发现有时候会很纠结，你做的东西越定制化，你可能能给一个人卖的越好，可能卖的越贵，但是就没有越更多人会买。但是你想卖的更多的时候，就会发现一些东西可能慢慢会失去它本身的特性一些。然后包括你要去找合适的市场，说可能一样的东西，你换个地方卖就能卖的多一些少一些。所以看起来逻辑特别简单，但是每一点做到都是特别具体问题，特别具体分析，我所以是赚钱其实还是挺难的。
李丁：问这个问题的听众也是个特别实在的人，他也是真心想知道怎么赚钱。最后的话我的问题这边问题问完了，对我还有一个问题，就是说既然你是B站的员工，我这视频每期也都会传YouTube，B站和其他一些播客平台。你觉得我的这一期视频怎么样可以在B站得到比往期更多的播放量呢?
曹翔：你这是直接问流量密码（笑），其实我也没有流量密码。但整体来讲我觉得B站做的比较好的一点，其实B站是很重视对于这种小创作者，新创作者的扶持。这就是说如果真的内容好的话，是可以在比较短的时间内去得到关注的。我觉得这一点可能是在其他很多那种纯数据驱动的平台上做不到的一件事情。
李丁：挺好的。我这边问题问完了，最后的话我不知道之前忘记跟你说了，如果你有什么想打广告或者招人啊，你有什么要打广告的吗？
曹翔：欢迎大家都来上B站，B站上有很多好看的东西。而且我来了B站以后，我觉得自己被科普了很多二次元内容。
李丁：我们这期的录制就到这里，非常感谢曹翔来这里跟大家的分享。然后欢迎各位在B站的下面也留言发弹幕，我们下期再见。
曹翔：拜拜。
本文转自： https://www.lidingzeyu.com/real-impact/
This site is for inclusion only, and the copyright belongs to the original author.