"Abandoning the meta-universe and deploying AI instead" has become a trend that ma
"Abandoning the meta-universe and deploying AI instead" has become a trend that major enterprises are scrambling to catch up with in recent days.
Microsoft allegedly dissolved the industrial meta-universe team formed four months ago; After entering the meta-universe two years ago, 360 ("360" for short) recently announced that it plans to launch a demo product similar to ChatGPT ..... Will the "AI craze" become the next bubble to burst quickly?
According to media analysis, the emergence of AIGC is like "a long drought meets the rain" for the venture capital industry. At present, venture capital companies are preparing to enter the market with 300 billion US dollars of gunpowder. At the same time, Silicon Valley, which is experiencing a wave of mass layoffs, is in urgent need of the next big event to make it "completely new".
However, the appearance of anything can't avoid the voice of opposition. With the heat wave of artificial intelligence ChatGPT sweeping all walks of life, the voice of concern and vigilance about artificial intelligence also followed, and "anti-ChatGPT" began to become a new force.
The media pointed out that the essential difference between the AI explosion represented by ChatGPT and Metauniverse and NFT is that for many people, Metauniverse is still a vague concept, but the underlying technology of AI has been applied in all walks of life for 20 years, and it is going through the stage of technical iteration and accelerated development, from software engineering to finance, to operation, to law, to logistics, to art. Therefore, although there is still a bubble in the AI explosion, it is undeniable that
Huatai electronics team analysis pointed out that whether to have a big model will become an important watershed for technology giants/technology platform enterprises. In the future, there will be a big model arms race among technology giants, and the long-term investment value of technology companies will be redefined.
Brought a new revolution?
After the release of ChatGPT in June 5438+065438+ 10, 2022, Microsoft announced that it would cooperate with OpenAI to launch bing search, Edge browser, Office suite and Azure with built-in ChatGPT, which doubled the pressure on Google, and the rapid development of AI also made the market unexpected. Reid hoffman, a member of Microsoft's board of directors, said last year that "magic really appeared", and Hoffman explained:
Copilot is helping software engineers write up to 40% code. Artificial intelligence will change all industries and everyone must consider how to use it. AI will appear in everything we do, which is the basis of everything we do.
A cloud computing executive said in a recent discussion at TEC Town Hall that the development of AIGC will not only affect search engines, but also a "cloud revolution" may be gradually erupting:
"This is a very important revolution, and I think it is as important as the industrial revolution. Companies that do not accept it immediately may be eliminated. "
Media analysis points out that the appearance of ChatGPT redefines machine learning, and the deep language learning model being developed and launched will reach another height:
Take the application in finance as an example. If we use AI to analyze 5,000 balance sheets, it can complete reading in a few seconds, extract all useful financial information, calculate risk scores, and make decisions on portfolio risks.
Whether "AI+" (a technology enterprise that empowers the industry with AI technology) and "+AI" (a traditional enterprise that adopts AI technology) can win is the biggest controversy in the last round of AI investment. The analysis points out that when the innovation cycle dominated by big model comes, taking GPT-3 big model as an example, it needs 654.38+075 billion data and millions and tens of millions of dollars of computing power input, and the barriers are significantly improved.
Wall Street informed research believes that due to the exponential growth of high computing power demand promoted by AIGC, the capital expenditure of cloud computing vendors on artificial intelligence training and cloud services may rise again.
With Microsoft as the driving force, capital expenditure shows an increasing trend for the AI function upgrade of search engines. Under the trend of declining market demand, other manufacturers may change their original capital expenditure plans.
Can create new employment opportunities?
The classic argument against the development of artificial intelligence is to describe it as a "job killer", but through media interviews, company executives pointed out:
Artificial intelligence will take over the repetitive work that humans should not do in the first place, and let humans do more important work.
According to the media, the current attitude of executives towards artificial intelligence is that artificial intelligence will not replace lawyers, but "lawyers who use artificial intelligence will replace lawyers." The analysis pointed out that:
"Now there is a view, especially in the professional industry, that if professionals such as lawyers and accountants do not use artificial intelligence, they will be replaced by those who use these tools, because those who are willing to use AI reasonably will become more efficient and can do more things."
In the technology industry, a CTO said that they started using AIGC to run the experimental results of service requests four years ago, which led to about 89% of the company's unplanned service requests being handled completely independently. The loss of the former team is zero:
"We have no layoffs. No one is unemployed, and I also think that people who improve efficiency through ChatGPT have replaced those who are inefficient. "
However, another executive said that the vision that some people can be liberated and engage in higher-level jobs is blindly optimistic and may have the opposite effect in the business world:
"When people begin to meet the challenge, people will become stupid because we just ask the computer everything."
The debate about artificial intelligence is becoming more and more fierce. Some executives believe that artificial intelligence-driven manufacturing and automation are essential for optimization and output, and if they don't use it, they will face elimination:
"We think it is very irresponsible for companies today not to seek artificial intelligence to provide solutions because they are very powerful now."
The media analysis behind AI points out that there are still many risks in artificial intelligence, including: a large number of false information spread by artificial intelligence; Prejudice; Threats to human work; Copyright issues; plagiarize .....
Huatai Electronics team pointed out that ChatGPT is a leap to AGI (General Artificial Intelligence), but it is still far from AGI. At present, there are endless disputes about using big models such as ChatGPT, mainly focusing on the following aspects:
1) information authenticity: the large language model lacks a true understanding of the facts and logical world, and the authenticity and interpretability of the generated content are in doubt;
2) Harmfulness of information: For example, although ChatGPT uses reinforcement learning (RLHF) based on human feedback to reduce harmful and untrue output, ChatGPT will still respond to harmful instructions if users guide it step by step;
3) Improper use: ChatGPT is not suitable for academic scenes. Many top journals have banned the signature of generative artificial intelligence tools.
ChatGPT is fun, but it's not the author. This is an editorial about artificial intelligence published by holden Thorpe, editor-in-chief of Science. It is pointed out that "originality" is the basis of scientific publishing papers, and the behavior of using ChatGPT to write words is equivalent to copying ChatGPT;
"We are updating our editorial policy, asking authors not to use text, data, images or graphics generated by ChatGPT (or any other artificial intelligence tool). Violation of this policy will be regarded as academic misconduct by scientific journals, just like tampering with images or plagiarism. "