Recently, ChatGPT has continued to be popular. Not only shine in the secondary market, but also very lively in the primary market.
Take Internet tycoon Wang Huiwen as an example. After ChatGPT became popular, this tycoon posted a post late at night, saying that he would build China's OpenAI and bring 50 million US dollars into the game. However, an investor in the hard technology track told the reporter of the "Science and Technology Innovation Board Daily" that this wave of AI fever in ChatGPT has only seen splashes, but no investors have taken action.
"The AI track was a track that some VC institutions preferred to invest in in the past, but after the track is shut down in 2022, some VCs are eager to find new outlets to help AI projects continue to raise funds." During this wave of ChatGPT boom, the investor Frankly speaking, I didn't feel the continuous financing boom in the AI track. On the contrary, some VCs may be trying to save themselves.
However, another investor in the south who was interviewed by reporters said that any track cannot be achieved overnight. "When judging, don't kill AI with a single stick, and don't push it to the sky because of ChatGPT. Investment is a long-term thing."
Can't find anything new to pitch in in the AI field?
Although ChatGPT has made the AI track popular again, the above-mentioned hard technology investors believe that in the field of AI, there is nothing new to invest in.
"From the perspective of start-up companies, they have no ability to compete with big companies. Behind ChatGPT is a large model, which requires 2-12 million US dollars for one training, and has high requirements for chips. If a start-up company develops a ChatGPT-like project, it means that it It takes such high-cost investments as building a computing center.”
"However, aside from investing in large models, startups seem to be able to do something in the field of AI. Since large models are open-sourced and more and more infrastructure-oriented, startups can engage in research and development of related AI scenarios on this basis." Said the above hard technology investors.
But for now, startups are still doing the same things as in the past, without any special breakthroughs.
Taking the industrial scene as an example, the hard technology investor said that AI is still the largest application in industry. It can replace people and do complicated and repetitive work to improve efficiency. But in this regard, VC institutions have invested a lot in the past, whether it is a US dollar fund or a RMB fund, they have all invested in industrial AI. It's just that the subdivision scenarios of concern are different, which are essentially identification, detection, sorting, supply chain optimization, etc.
The investor believes that ChatGPT’s wave of AI fever has not made AI smarter. "It's just that the data model has been expanded, allowing AI to have the ability to generate, and then produce the track of AIGC."
Regarding AIGC, the investor said that this is a newly derived track, which is a bit more entertainment, but the real payment demand downstream is debatable. "For example, AI automatic choreography, no one in the downstream is willing to pay. Although AI painting and AI scripts can generate content, it feels that the TOC end is highly competitive. Any traffic that appears cannot escape the follow-up of major manufacturers. Special effects on AI paintings, short video platforms Douyin and QQ have launched related applications, and Meitu, which specializes in picture beautification, also has related applications, and they are all free.”
"At the same time, painting requires artistic value. Whether AI painting will make traditional artistic creation lose its value is also worth thinking about."
As for the AI intelligent virtual digital human that has come into sight again, the concept of metaverse is superimposed, including AI-driven AR/VR technology applications, etc., but these applications are all virtual worlds. The investor believes that human nature is still the needs of the real world, and the hardware investment in the metaverse, AR/VR, etc. are difficult to wear lightly on the body, coupled with the purchase threshold, the number of C-end users will not be too many . At the same time, the grand scale of AR/VR game scenes and 3D rendering... These investments are the reasons why it is difficult for start-up companies to get out of the profit dilemma. "
So, in this wave of AI being hyped by ChatGPT, what new things are there to vote for? The investor's answer is: at least not for now.
The AI track has not completely disappeared
However, another investor in the South said in a chat with a reporter from the "Science and Technology Innovation Board Daily" that some people made money on this, and some people lost money. Therefore, opinions and attitudes are different.
As far as the AI track is concerned, he believes that the track has not completely disappeared. "In the past two years, the AI field has indeed been cold, but there are still some funds in the market that are paying attention and investing, including Lee Kai-fu of Sinovation Ventures and Lu Qi of YC. Their funds are dedicated to investing in the AI field. During the exchange, I asked them Why do you still vote for AI, but the answer is simple - you are optimistic about AI, so you are willing to keep investing.”
Such an answer reminded the reporter of "Kechuangban Daily" of the latest investment: ByteDance quietly invested in Lu Qi. The YC (Silicon Valley YC Incubator) represented by Lu Qi, a well-known entrepreneurial incubator in the United States, has been increasing its focus on the AI track.
Returning to the field of AI, the above-mentioned investors in the south said that the AI track has really become hot during this period, not only in China, but also in foreign countries. "The field of content output represented by AIGC has become hot all of a sudden, and many entrepreneurial fields have emerged here. I believe that a lot of financing will be generated this year, because where there are entrepreneurship, there will be financing."
On the application side, the investor also mentioned that there is no shortage of applications in the AI field for a long time. "However, the reality is that some institutions and companies are eager for quick success, expecting disruptions and explosions in 2-3 years, but the development and application of AI technology is not so fast, causing bubbles in valuation. Simply put, it is expensive to invest. Outlet Fund."
Facing the views of the above-mentioned technology investors on AI paintings, this investor also believes that the reason why AI paintings are popular is that the final result is not to sell paintings, but to enter the industry through AI art creation. "Especially in the field of game production and music, a lot of labor was originally required, but after AI came, it quickly produced content and replaced part of the manpower; including AI writing medical records, when the large model is trained strong enough, it may also replace Junior doctors ask patients, write medical records and reports, etc., helping doctors save time.”
But for AI to replace doctors and write medical records, investors in the above-mentioned hard technology field think that the application in this field has also been paid attention to. "However, AI is just a tool and does not bear legal responsibility. It is unclear whether there will be a legal threshold for how many related applications will be popularized."
"The most important thing is that many investors can't tell whether it's a software investment or an AI investment, because the valuation and cost of software investment may be reduced a lot; but for AI investment, the salary of hiring an AI master is as high as tens of millions. Will the increase in labor costs lead to a new round of bubbles?"
In this regard, the southern investor agreed with the above investor's statement. He believes that any new technology has such problems. "The key is not to set the project and product positioning too high, the company's valuation should be lower, and the expected development speed will be slower. In the process of not deifying, let the investment return to normal, and while the industry is developing slowly, it will no longer follow the AI flameout the old way."