Recently, “AI God”, Dai Wenyuan, founder and CEO of Fourth Paradigm Company, was a guest in Gelonghui's high-end interview “Dr. Gelong's Living Room”.

When talking about how large the fourth paradigm expects to be in the future, Dai Wenyuan admits that the ideal target is probably a very huge company.

From an industry perspective, the fourth paradigm is to serve all B-side enterprises through a broad industry model, which is considered to be providing tools.

In response, Dai Wenyuan said that when they first started their business, several co-founders talked about this topic. If placed in Baidu, this can help Baidu improve its efficiency by 8 times in 4 years, and now it is far more than 8 times. If you put it in bytes, like the average online time on today's headline platforms, through recommendation engines, it can be increased 5 times within 1 to 2 years. All walks of life, if not five times, eight times, ten times, even just 50%, space is unimaginable.

But he also admits that to do business, you still need to be down-to-earth and step by step. For Douyin, Baidu, or Today's Headlines, this is called a headshot scene, a huge scene, and for the entire market, it probably consists of many small scenes. Therefore, not solving a single problem brought in hundreds of billions of dollars in output value. It is possible that solving a problem only brings a few million production values, but it is unimaginable that millions of such scenarios need to be solved.

Opportunities and challenges coexist

As far as the fourth paradigm is concerned, Dai Wenyuan believes that opportunities and challenges coexist. A challenge is not about solving a single problem; it may require solving millions, tens of millions, or even hundreds of millions of problems; to solve millions, tens of millions, and hundreds of millions of problems, it is also necessary to have the ability to solve so many problems.

Dai Wenyuan pointed out that the methodology behind everything they have done to this day is the fourth paradigm methodology. But you still need to recognize that your abilities are lacking. “If we had 10,000 customers at the same time today, we wouldn't be able to serve them at all. It's not that it can serve hundreds of millions of scenes; we can't even serve 10,000 scenes.” he said.

However, he still said that he will always be optimistic. Although there is a huge gap between now and the value he wants to achieve, he feels that he is much better now than half a year ago. He believes that in half a year, he will be able to do it again.

Everyone is in the fourth paradigm

Referring to the position of the fourth paradigm, Dai Wenyuan said that if you don't position yourself, you must be To B, or you must be To C. He believes that even to B is “AI for everyone.”

The fourth paradigm concerns almost everyone; everyone is in the fourth paradigm. For example, go to a restaurant such as KFC to order food. These foods are probably recommended by the fourth model of technology. Another example is when you go to a bank transfer, what people perceive is that the money has gone to a friend, but in the process, they may have gone through the fourth paradigm of anti-fraud and anti-money laundering systems.

Therefore, Dai Wenyuan believes that even though the fourth paradigm now has almost no To-C business, it is actually already “AI for everyone.” Of course, this is far from enough.

Currently, the vast majority of AI companies are facing loss problems because developing large iterative models, whether in general or vertical industries, requires extremely high R&D investment.

Regarding how to balance R&D investment and commercial profits, Dai Wenyuan said, first, Fourth Paradigm, as a technology company, must continue to increase R&D investment. Only with a certain amount of investment in R&D can we ensure leadership in this field of technology. This is the most basic basic model, and at the same time, the fourth paradigm also focuses on (profit), because in any case, it is a business and a commercialized company.

He believes that in the end, it is still necessary to return to the business itself and back to the commercial enterprise itself. The so-called loss or burning of money is phased; it cannot always be this way. Therefore, the company will balance investment in R&D, the growth rate of revenue, and the pace of loss.

Dai Wenyuan said that the company has been reducing losses since two years ago. Last year, the adjusted loss ratio had dropped to single digits. An AI company definitely needs investment, but in the end, it also needs to become a profitable company. Therefore, the fourth paradigm has always had very clear steps and profit rhythms.