A brief background of the important news
On May 5, according to Reuters, citing six people familiar with the matter, computer scientist Li Feifei is building a start-up company with investors including well-known Silicon Valley venture capital a16z and Radical Ventures.
Li Feifei is a distinguished Chinese computer scientist and professor at Stanford University. One of her important contributions in the field of artificial intelligence is the creation of the ImageNet database, a large dataset of more than 14 million labeled images, which has played a key role in the latest developments in deep learning and AI. Therefore, she is also known as the "AI Godmother" and "AI Goddess".
Li Feifei once spoke at the TED conference in Vancouver that the cutting-edge research involves an algorithm that can reasonably infer what images and text will look like in a three-dimensional environment and act on those predictions, using a concept called "spatial intelligence". According to sources, "spatial intelligence" is the research direction of the startup established by Li Feifei.
So, why does Li Feifei value "spatial intelligence"? What are the prospects for her startup?
The following is Wang Yuquan's important news commentary:
Recently, there is a news that is very noteworthy, that is, a leading figure in the field of artificial intelligence, Stanford University professor Li Feifei started his own business.
Li Feifei: I think everyone is familiar with ImageNet, and she started ImageNet. For the first time, artificial intelligence has shown the ability to be comparable to that of humans, that is, it can recognize all kinds of content in images, which is what Li Feifei has promoted. In other words, this person is pivotal in the field of artificial intelligence, and now she has also started a business.
Those who are familiar with us know that we don't really support professors starting their own businesses, because after all, there are specializations in the art industry. But the field of artificial intelligence is an exception, why do you say that?
There was such an entrepreneur before, that is, the godfather of artificial intelligence named Geoffrey Hinton, who founded a company (DNNresearch) back then, and then the company was quickly sold to Google and made a lot of money. Of course, he founded this company with a few of his protégés, who joined Google and later became leaders in the field of artificial intelligence.
Li Feifei also has a lot of high-level apprentices, and she must also start a business with a group of students, at least this start-up company will be guaranteed in the future, why? Because there must be large companies willing to collect, on the one hand, the company's technology, and on the other hand, these artificial intelligence experts in the company.
Li Feifei has just started a business, and it is said that he has not yet confirmed his name, but he got angel investment, and it is a very famous investment fund in the United States, called A16Z.
Moreover, this technology now seems to be in demand and in demand. In fact, we said several years ago that the development of artificial intelligence must first be the text of the content, then the two-dimensional, visual, and finally the three-dimensional spatial analysis.
The previous paragraph of text has broken through, the picture has broken through, and Sora Vision has also broken through at the beginning of this year, so the next step is likely to be a breakthrough in three-dimensional space, and Li Feifei is also aiming at the next strategic focus to do research.
So they are really different from other start-ups, because they don't have to make a product, whether it's ToB or ToC, and then sell it all over the world to achieve revenue and profit. They are more like Google's acquisition of DeepMind back then, just because this technology is very good, and it can enrich my technical team and enrich my company's technical strength.
Li Feifei is at the forefront of artificial intelligence, and she must know where the next breakthrough should be, so she will lay out in advance.
So do you think this layout is important or not? Frankly speaking, we also think it is really important, as I just said, we predicted that three-dimensional intelligence must be the future development direction.
But do you think it will break through in the next two years? Frankly, I don't think so, it's not like it's pouring cold water.
In the past, there was indeed a demand for this aspect, whether it was autonomous driving or robot training, it was indeed necessary to do it in a three-dimensional environment. But we all know that after Sora came out, there has been a leap in autonomous driving, especially a leap in pure vision-based autonomous driving, what does it mean? It shows that it is possible to use a two-dimensional system to train a three-dimensional model, and it is valid.
Autonomous driving can leap, and the same robot training can also leap. We have all expected that from now on, this year and next year, we will see that the flexibility of the robot hand will be significantly improved, and even begin to compete with the flexibility of the human hand.
It is because of such a background that two-dimensional can train the robot hand, so three-dimensional intelligence is not so important. At this time, of course, it is not impossible to break through, and it will be much more difficult.
We have a view that is not quite the same as the popular view nowadays, we have always believed that it is actually the market pull that can make the technology break through faster, if the demand is not so strong, then the technology breakthrough is not so easy.
However, I actually want to mention another very interesting topic, that is, I believe that Li Feifei has become a Stanford professor, she is not short of money, and she has already achieved fame, so why do you have to jump out and go to a start-up company? Why not do research at Stanford?
In fact, this mentions another embarrassing thing about the research of artificial intelligence in colleges and universities, that is, the research of artificial intelligence is becoming more and more expensive, and it is really unaffordable to have no money.
Whether it is a large language model, a large visual model, or a large model called three-dimensional that Li Feifei will work on in the future, what is one of its characteristics? All need strong computing power support, this kind of computing power is either large companies have, or there is sufficient financial support, start-up companies led by special experts, basically there is no third way, colleges and universities must not be able to afford it.
I think maybe why Li Feifei wants to start a business is really because he saw that this is a direction and wants to break through, but he is really powerless in Stanford, so he has to simply come out and start a start-up company by himself.
In the future, you care about Li Feifei's entrepreneurship, and what topic you have to care about, that is, how much money she has financed and how many GPUs she has bought. Therefore, in the future, if you look at artificial intelligence or even technological breakthroughs, pay attention mainly not to come from universities, because colleges and universities can't afford to play.
Where do you want to come from? It depends on the big companies, the big companies play the role of research, Google has money, Facebook has money. There is also a star startup, OpenAI is rich, and now Li Feifei and they may still be rich. Other companies really don't have to play if they want to do this kind of large-scale model development, you can go to the application, but you don't have the ability to do basic development.
Therefore, we know where to look at the frontier of science and technology in the future, of course, it depends on their progress, and the development and application of its technology should be based on their scientific and technological frontier, so as to seize the real market opportunity.
In addition, this year is the 8th anniversary of the science and technology training camp, and the friends of the innovation map have prepared the 8th anniversary benefits for everyone, from May 1st to May 31st, that is, if you buy a member of the science and technology training camp this month, you can get two months of school hours.
Interested friends should seize the time to scan the QR code in the event poster below, sign up, and let us understand the future together in the science and technology training camp.
The above is today's content, more detailed industry analysis and underlying logic, I will share it in the science and technology training camp. Welcome to pay attention to the global outlet WeChat account and sign up.
Wang Yuquan has to be heard, and we'll see you tomorrow.