Commercialization of ChatGPT products: ToC companies actively embrace it, while ToB companies cautiously wait and see

"When the times abandon you, you won't even say goodbye." How sad and decisive. GPT-4 was officially launched last night, only 6 months since the last iteration and only 4 months since the launch of ChatGPT. Its multi-modal and real-time learning capabilities have been greatly improved, making CIOs and CTOs of many companies feel like they are facing a formidable enemy.

Technology and markets are changing rapidly. Cloud computing, big data, blockchain, artificial intelligence... familiar and unfamiliar words flash before our eyes every day. ChatGPT has become a "top trend" in the past month, touching the heartstrings of countless technology enthusiasts, fearing that they will miss it.

ChatGPT seems to be capable of anything: chatting with people and having active thinking, being able to write and draw to open up your imagination, modifying the code so that bugs have nowhere to hide... While it has attracted widespread attention, it also seems to have caught the attention of capital and entrepreneurs. "The next trend". Baidu, Alibaba, JD.com, and iFlytek have all stated that they will launch similar products as soon as possible. On February 7, Baidu claimed that it would launch related products, and then Baidu said that the "ChatGPT-like application" ERNIEBot will complete internal testing this month.

One stone stirs up a thousand waves. Nearly 300 companies in industries such as the Internet, media, finance, automobiles, and insurance, including Postal Savings Bank of China, iQiyi, and Artron Culture Group, have announced their access to the Baidu Wenxin Yiyan ecosystem. Baidu is full of confidence in the commercialization of AIGC (AI-Generated Content), and CEO Robin Li can't hide his excitement, stressing, "The technology has reached a critical point. How can technologies like ChatGPT be used in search scenarios?" , we will have great opportunities in the next year.”

Jiang Zijun, CTO of Yachang Culture Group & General Manager of Yachang Art Network, said in an exclusive interview with a reporter from Zhiding.com: "Wen Xinyiyan will be launched in March. Baidu is a domestic enterprise with advanced AI accumulation, so we We will give priority to such partners."

"We are now focusing on laying out the Metaverse track. Since ChatGPT is not open to domestic institutions and due to security and compliance considerations, Baidu's Wen Xinyi said: It is a relatively reliable choice at present. Some people think that when AIGC comes, the Metaverse will not be popular. I think this is a one-sided view. The Metaverse is a big concept, and Web 3.0 is its infrastructure or branch. Just like in the PC era It is called Web1.0, which solves the problem of information connection of PGC; the era of mobile Internet is called Web 2.0, which is actually UGC, used to solve the problem of information efficiency; in the era of Web 3.0, or called the metaverse, it is necessary to solve the problem of information fairness. Most of the content is generated based on AIGC in the virtual world."

"So AIGC technology has great potential. It is the same infrastructure as Web 3.0 including virtual digital people, and they do not conflict with the metaverse. , can be fully integrated. Just like mobile networks have 4G, 5G, and there will be 6G in the future, but it does not mean that 4G will no longer be usable after 5G comes out. In the end, a converged system composed of heterogeneous networks and heterogeneous terminals will be formed."

AIGC is overestimated in the short term and easily underestimated in the long term

"People always overestimate the capabilities of technology in the short term, but underestimate it in the long term." Jiang Zijun commented on ChatGPT. However, he pointed out: "No matter how intelligent AI is, it still relies on people, algorithms and computing power, and more importantly, data. As for ChatGPT, it relies on data before 2021, and its pre-training cost is very high. The real-time performance is also very poor, and it does not have real intelligence."

"Because it relies heavily on the data you feed it, its answers are a bit like parroting, and more often than not, it is serious nonsense. Including some simple 'chicken and rabbit in the same cage' problems can stump it. So in essence, it will not replace people, let alone programmers. If it can replace it, it will be the most entry-level and low-level program. However, it can serve people better as an auxiliary tool."

"It does not create science, but is more of a breakthrough in engineering. Just like when the Metaverse first came out, artists are also very scared of it. , but as a technology, art and technology have never been separated. From a technical perspective, it is volatile and constantly evolving. The pigments used by artists have thousands of years of history, but pigments It has been changing and progressing. In the 19th century, cameras appeared, which today are called photography technology. Later, there were movies. In the future, the metaverse will also have newer technologies. This will definitely have some psychological impact on artists. Even panic. But from an artistic perspective, human beings' emotional pursuits and the exploration of humanistic spirits such as man and nature, man and the universe, man and society, are eternal, and technology will not have a big impact on it."

"From a historical perspective, there will be no revolutionary impact. Whether it is ChatGPT or the Metaverse, it is possible to replace it in certain scenarios. But in terms of expressing true feelings , the machine does not yet have this ability.”

Consumer Internet actively embraces AIGC, while the industrial Internet is cautious

Unlike the consumer Internet that actively embraces AIGC technology, although industrial Internet companies are very interested in this technology, they are not willing to apply it to the company's production The attitude of the scene is very cold.

Mr. Jiang, CTO of a well-known domestic energy company, said in an interview with reporters: "We currently have no specific plans for the application of AIGC technology. It is not very helpful to energy companies like us, and there is no suitable solution in the short term. Application scenarios."

"We have also tested ChatGPT. The product is very rudimentary and not easy to use. Especially in terms of the reliability of the results and performance, we do not have enough trust in it. ”

But AI still has a lot of room for development in enterprise production. Mr. Jiang believes: “For new energy companies, our energy mainly comes from light energy and wind energy, and we are highly dependent on weather. Therefore, accurately judging the weather is very important for our production. Current weather forecasts mainly use historical data, Satellite data and ground observation data are continuously monitored and deduced, and future weather conditions are predicted through algorithms. Now it seems that the accuracy of AI predictions is still far from our actual needs."

"But at the same time We have seen that ChatGPT applies large model technology. We will also consider using large model technology to accurately predict weather, and hope to make a breakthrough in this regard."

"We have seen that it is used in word processing, For example, it has some advantages in legal provisions. It can accurately judge various compliance provisions, and can provide answers very quickly when reviewing contracts and extracting key data. In the future, there should be more opportunities in corporate compliance, contracts and finance. Suitable application scenarios. However, it may take 3-5 years for large-scale application in enterprise production to see a clearer route."

The AIGC industrialization road is long and long, and we still need to be down-to-earth.

We are talking about AIGC today. It represents not only technology, but also reflects the anxieties and shortcomings of the entire industry. Jiang Zijun pointed out: "Today we are talking about ChatGPT and hope it can change something, but it still faces many challenges: not only technical issues and product experience issues, but also cost, compliance issues, copyright issues, security issues... ”

“We have explored a lot at the technical application level in China and have a lot of successful experiences, so we have seen the phenomenon of chasing various hot spots. Chai Yuan Universe, blockchain, Web3.0, ChatGPT …This shows that the growth of the Internet today is weak and development has encountered bottlenecks. Everyone is competing for existing stocks and there is a lot of homogeneous competition. All industries are seeking new breakthroughs regardless of capital, technology, or products.”

"This also reflects our lack of accumulation of underlying technology and our unwillingness to invest in infrastructure and underlying technology. We are only keen to pursue short-term economic benefits and are oriented by financial reports. We will definitely not be able to make great products and create great companies. ."

"The future will still be driven by technology. Success cannot be achieved by just relying on a simple business model. We must do a good job in lower-level basic work to avoid low-level competition. Like Amazon's previous long-term losses, If you want to do logistics and cloud computing, only underlying technologies such as chips, networks, storage, and operating systems can form barriers and become your true competitiveness."

The above is the detailed content of Commercialization of ChatGPT products: ToC companies actively embrace it, while ToB companies cautiously wait and see. For more information, please follow other related articles on the PHP Chinese website!