#March Finance New Forces ## Artificial Intelligence ## Internet ## Internet of Things ##chat GPT ##chatGTP Artificial Intelligence ## Make AI within reach ##AI吐槽 #
Just during the first session of the 14th National People's Congress, Zhou Hongyi, a member of the National Committee of the Chinese People's Political Consultative Conference and founder of 360 Group, proposed that "the huge leap in artificial intelligence large model technology represented by Chat G P T will set off a new industrial revolution, and the mainland should catch up and create China's own "Chat G P T". The mainland has a foundation in the development of artificial intelligence large model technology, and it will reach a relatively good level of following within two or three years. It is suggested that while we pay attention to innovation at the technical level, we should also pay attention to the innovation model on which technological breakthroughs depend.
At this year's two sessions, I brought a proposal on promoting the technological innovation and development of artificial intelligence large models, specifically mentioning three points: first, by establishing a collaborative innovation model of production and research of large science and technology enterprises + key scientific research institutions, to create China's "Microsoft + Open A I" combination to lead large model technology research; The second is to support the establishment of long-term open source projects with multiple national artificial intelligence models, and create an open innovation ecosystem with open source crowdsourcing; Third, it is suggested to start from the height of building a national strategic scientific and technological strength, select production and research institutions with a good technical foundation, and form an industrial synergy combination with complementary advantages. ”
Not long ago, Tesla CEO Musk also sent a push, saying that we are not far from the powerful and dangerous artificial intelligence.
In addition, Bill Gates, the founder of Microsoft and former richest man in the world, also helped to say that this A l is no less significant than the revolution in personal computer and Internet technology that I set off.
What are these big guys talking about?
It is the recently popular chatbot app Chat G P T, which is the artificial intelligence agency Open A I that Musk co-founded a few years ago, and launched at the end of last year. This software looks unremarkable, and there is only one dialog box, which is a huge gap with the shape in science fiction movies. So why is this thing so popular that even street grandpas and aunts care? Because artificial intelligence does several fierce things, this really makes everyone break out in a cold sweat.
First of all, Chat G P T's IQ test result exceeded 80 points! And the IQ of a normal adult is generally more than ninety points, which is very close. And within two months of the app's launch, the number of users exceeded 100 million, breaking the record of the app store. In addition, American college students have begun to use it to write homework. School professors and teachers are stunned and can't check what is done by humans and which is done by artificial intelligence. This directly forced the New York City Department of Education to ban Chat G P T directly. There are also programmers who let him write code. As a result, it easily passed the entry test of Google Coding Level 3 engineers, and directly raised an annual salary of $180,000.
Not long ago, the Israeli Prime Minister gave a public speech, part of which was written by Chat G P T. Therefore, those bosses' secretaries have recently begun to worry about it, for fear that this job will be gone.
After testing by a wide range of users, or having fun, people have found that Chat G P T can not only chat with you, but also help you work. Whether it is helping to write news reports, writing love letters, writing legal opinions to clients, writing a report PPT to the boss, or even writing business techniques for investors, taking judicial exams, and writing a legal mini program. Just as handy. There are even traders who have customized quantitative trading strategies with artificial intelligence, which can be called the ultimate problem-making machine.
But the problem is, a lot of people don't understand, how is this done? It feels like Chat G P T is smarter than Xiao Ai classmates, Xiao Du, Tmall Genie, Siri and everything. This should be explained that Chat G P T uses a model called deep pre-training transformation. This thing sounds complicated, but in layman's terms, it is a cultivation system. His knowledge base comes from all the information on the internet before 2021. There are statistics on how big this information is, as many as 3-500 billion words. And this information comes from search engines, Internet encyclopedias, blogs, the official websites of major organizations and media in the world, and so on. From pure knowledge to official speeches to netizen bragging, there is only one. Of course, to do this, it is technically stacked, after all, there is all the information on the Internet. However, this is only the first step.
Then, in order to make Chat G P T perform better, the staff will adjust the ordering, training, reward and punishment model for each round of dialogue with him. Simply put, tell him that your answer is inaccurate, that your reaction is not normal, and that you do not understand this question correctly. Chat G P T in each interaction, according to the staff's set of standards, select the most suitable answer, feedback. It is a bit similar to children learning to speak, just next to adults every day contact with these parents in the short listening, children themselves will slowly speak. For artificial intelligence, it can also slowly simulate the tone and answer that humans can accept. After the launch, more than 100 million users around the world were chatting and interacting with Chat G P T all the time, and at this time, users replaced the staff and became his powerful trainer.
Netizens jumped in their minds and had many ideas, which stimulated the rapid growth of his language ability. So the result is that Chat g p t simulates the learning ability, and will no longer be like those artificial intellectual disabilities in the past, always producing back questions or dead duckbills. Not only can you complete some creations, but you can even see the wind and steer. There have already been attempts to use him to create a novel, as well as paintings, music, and poetry. At first glance, there is nothing wrong with it. You can even create a language. Another person asks him how much 0.5 multiplied by 0.6 equals, and he will tell you that it is 0.3. But if you tell him again, my boss said it, it should be equal to 0.03. Not only will he tell you that your boss is right, but he can also help you analyze how to calculate that it can be equal to 0.03. So that some people were surprised, this guy is simply a real person! He actually lies!
The next question becomes, Chat g p t is so capable, will it really rob us human jobs?
Essentially, Chat g p t is like a search engine for us users. It can understand human problems to a certain extent, organize and summarize the information on the entire Internet well, and at the same time absorb information as a reference in the process of dialogue with us, and organize and output it in a language that it thinks is easy for humans to understand. So the first ones most likely to be impacted are the giants of several search engines. It's no wonder that Google Baidu recently rushed to announce that it also has an artificial intelligence application similar to Chat g p, and Microsoft Bing directly integrates it into its homepage.
Plus, some experts who are always talking may be offended because Chat g p t's ability to say the right nonsense is equal. From time to time, some absurd remarks that pop up are stunning. It doesn't seem that this artificial intelligence is very annoying. Of course, although some people predict that he will have unlimited potential in the future, the reality is still so skinny. If you encounter a slightly professional problem, you may be able to do some simple math problems wrong. Or it may spill a big article, but you can't give references and sources, like those smart people on the Internet who talk and talk nonsense. Even it says that I can't feel it, I'm just collecting data. So, for now, it seems that replacing us is far away.
But why should we care about such an AI dialog? Are we mythical too much about Chat GP?
There's a saying on the Internet that when you complain about Chat g p T running trains with their mouths full of nonsense, it may be a bit like seeing a monkey on the beach writing 1 + 1 = 3 in stone. It does miscalculate, but that's not the point, because it will one day be right.
What do you think?
Welcome to leave a message to discuss! Thank you all for liking, collecting, and forwarding!