Two days ago, I saw a ** interview with Mr. Liu Qing, about the huge impact of AI (artificial intelligence-artificial intelligence) on education, he has a prediction:
"Children who are in elementary school now, when they get to the college entrance examination, the current college entrance examination may no longer exist...
Source: This reading "Life. Screenshot of the fourth season of "Growth".
I don't know what you think, but when I read this paragraph, I still felt a great shock in my heart.
November 30, 2022, ChatGPT35 was born!Too many people were shocked :The era of AI has come overnight.
Source: openai
Now, a year later, let's take a few minutes away from the recent winter flu and focus on this once-in-a-century event in the tech worldChatGPT was born, and 2023 will become the first year of AI.
How big of an impact will powerful ChatGPT have on the future?
In the age of AI, will children lose their jobs?
Will the college entrance examination be cancelled?
What kind of inspiration does the growth experience of top AI talents have for cultivating children?
The future has come, look up and glance into the distance. I hope that these topics can not only be used as popular science content for AI, but also bring you some thoughts on the future of education.
Approaching ChatGPT, Pandora's Box is opened!
ChatGPT's 1st Anniversary Self-Evaluation
Q:How do you rate your impact on humanity over the past year with 3 keywords?
Innovation, education, interaction,These three words summarize the main impact of ChatGPT on all aspects of human society in the past year, and education is one of the three, which has been the consensus of the industry.
How popular is ChatGPT?
On November 30, 2022, OpenAI grandly launched its chatbot ChatGPT. It attracted 1 million registered users in just 5 days, breaking the history of the Internet.
Just two months later, in January 2023, ChatGPT's monthly active users soared to 100 million, an unprecedented rate. Facebook used to need 4It took 5 years to reach this milestone and it took 3 for WhatsApp5 years, Instagram used 25 years, while Google is close to a year.
OpenAI's traffic has also shown a sexual increase, from about 20 million per month last fall to a staggering 1.8 billion in April 2023.
Source: Internet.
2023 is the first year of AI. "AI" was named the most talked-about word of 2023 by dictionary publisher Collins. "AI" was named the most talked-about word of 2023 by dictionary publisher Collins.
How awesome is ChatGPT?
Q:What are some common uses of ChatGPT?
Little Half Dad said
The most important reason why ChatGPT has caused many shocks to people is that the evolution of artificial intelligence it presents has entered a new stage of developmentRapid development period.
This rapid development is reflected in the fact that AI can accurately understand the needs of users and continue to be competent in some of the tasks that should be completed by humans with high quality.
Looking at the long list of features above, this substitution is no longer a plot in science fiction, but an occurrence that already exists in real life.
Although the AI still needs to be perfected, there is no doubt that Pandora's box has been opened!
Will smart chatgpts make humanity finish?
AI is smarter than we think
In the early days, it was widely believed that high-level mental work, such as the work of doctors and lawyers, could not be replaced. However, recent trends have shown that AI is not only capable of doing jobs in these areas, but is also far more efficient than humans.
As a result, people have turned to the belief that professions that require creativity, such as painters, designers, and writers, should be relatively safe. But it wasn't long before it became clear that computers could also draw, design, and even write press releases.
Subsequently, there was an opinion that emotion recognition is a unique ability of humans and that robots cannot possess such an ability. But now, robots have been able to accurately judge a person's emotions by analyzing subtle changes in facial muscles, even more accurately than humans.
This means:The possibility of AI replacing human jobs is not limited to any particular industry, but is comprehensive.
On November 29, at the Dealbook Summit hosted by the New York Times, the host asked Iron Man Musk, "Although you are known for exaggerating the speed at which things are moving, I still want to ask you how soon you think AGI will come."
Musk thought carefully for a while and deduced that if the standard is"Smarter than the smartest humans in everything, smarter than the machines augmented humans of the future," is really hard to say.
But to put it mildly, it will be less than 3 years before writing about AI that is as good as JK Rowling's ability to discover new physical laws or invent new technologies. ”
Source: Interview with Musk at the Dealbook Summit hosted by the New York Times.
In less than 3 years, if it is really so fast, will humanity be ready by then?
In the eyes of ChatGPT: Will humans be unemployed in the era of AI?
Q:As ChatGPT features continue to evolve, do you think humans are at risk of losing their jobs?
In general, ChatGPT is currently good at information and data processing tasks based on language input and output, and text, voice, images, and programming are essentially different modes of language.
While AI technology like ChatGPT may lead to the reduction of certain occupations, it also has the potential to create new career opportunities and improve productivity.
We humans are going through an evolutionary process from the beginning of making the machine bionic, to the machine becoming a part of us, and then to the machine becoming another self and replacing us for work.
Every major advance in science and technology has the potential to accelerate the differentiation of human society.
AI will inevitably reduce or even disappear certain professions, but those who have mastered and used AI technology as early as possible will become the group of people who will grasp the dividends of the times.
Because it is difficult to improve the speed and height of AI evolution, our focus should not be on researching which jobs will not be replaced by AI, but on changing our perspective, considering how to coexist with AI, and paying more attention
1. What will people still do better than machines in the near future?For example, open creative work.
2. What do we still want to be done by people even if the machine is better than us?For example, work related to interpersonal communication, daily communication and coordination, and management work are also included.
Premise: The operation of human society is still controlled by humans, and there are no perfect robots.
Eventually: in the future AI era, people will no longer be a separate competitive unit, and the pairing of humans and AI will be a basic combination.
How will AI change education?
How AI works, don't do it in education
AI is so powerful, we might as well come to popular science to learn the training principle of ChatGPT.
The full name of ChatGPT is Generative Pre-trained Transformer, which mainly consists of three steps.
Step 1: "Collect sample data and train a supervised model". From thousands of questions, some questions are marked by humans, and the parameters are written to the AI, which then understands how humans express and learn.
Step 2: "Collect comparative data and train a reward model". People score and rank different responses to the same question from the AI, and through this comparison of results, tell the AI what is good and what is not.
Step 3: "Continuous optimization of the supervised model based on the reward model". It is to iterate and repeat the first two steps, so that the text output by AI is more and more in line with human cognition.
Source: OpenAI Blog
It's familiarChatGPT is now the world's most intelligent population. Its training method is the same as brushing questions, first do example questions, then compare and familiarize yourself with the answers, and finally a lot of repetition to improve speed and accuracy.
We can draw two points from this: firstly, brushing questions is indeed useful, AI also needs to brush questions to make itself more "smart", and brushing questions is not anti-intellectual;In addition, AI has done "no way out" to improve intelligence through brushing training.
With the advent of the AI era, our education strategy should reflect differentiation and do the opposite.
Since AI is already a master of brushing questions, the education given to children must be reversed, light brushing training, heavy thinking training, and mathematics must be learned well!
What major will the child study in the future?It was difficult before, but it will become even more difficult in the AI era, because the technology is changing too fast, and AI requires both PK and cooperation. The current profession should all encounter challenges, need transformation, and need +AI.
Down-to-earth, the current test-taking questions, for the time being, have to continue, to avoid losing in the present;But at the same time, it is important to tap the child's potential, awaken his real interests, and help them win in the future.
As for whether the college entrance examination will disappear, Mr. Liu Qing's judgment direction is right, AI will definitely have a huge impact on the college entrance examination, the content and method of education will change, and the selection method of talents will inevitably be adjusted.
The college entrance examination may not be canceled, but the form of the college entrance examination should usher in a major adjustment in 5 10 years, and changes in college education will take the lead.
Two representatives of the AI era
The Godfather of AI - Geoffrey Hinton
Source: Network Geoffrey Hinton and his students Alex Krizhevsky, Ilya Sutskever (from.
He never officially took a computer course.
Undergraduate: Yes, at the University of CambridgePhysiology and Physics, during which it was turnedPhilosophy, but in the end it wasPsychologyBachelor's degree in direction;He was tired of studying to work as a carpenter, but after a setback, he returned to the University of Edinburgh and got an unpopular majorArtificial intelligenceDirection of Doctorate;After becoming a professor, he always had to consult his graduate students about neuroscience and computational science that he didn't understand.
Geoffrey Hinton's academic path seemed faltering, but he eventually made it to the top, being revered as the "father of deep learning" and receiving the Turing Award, the highest honor in computer science.
In 1973, Hinton pursued a PhD in artificial intelligence at the University of Edinburgh under Langer Higgins. At a time when almost no one was optimistic about neural networks, even his mentor advised him to give up research on the technology. However, the doubts around him never wavered from his belief in neural networks.
After nearly half a century of technical precipitation and life grinding, in 2012, he cooperated with students Alex Krizhevsky and Ilya Sutskever, and proposed AlexNet, which shook the entire industry. This marks the reshaping of the field of computer vision and opens a new era of deep learning.
The father of ChatGPT - Sam Altman
Source: Altman posted on social media on November 20.
Born in Chicago, Illinois, USA in 1985.
At the age of 8, he owned his first computer and became obsessed with programming.
At the age of 16, he came out of the closet frankly to his parents.
At the age of 19, he dropped out of Stanford University and co-founded Loopt, a social networking company, with a friend.
At the age of 26, with $43 million**loopt, he earned the first pot of gold in his life.
At the age of 28, he was favored by the founder of YC and became the president of YC. At the age of 29, he founded Openal with Musk and became a co-founder.
At the age of 30, he became a member of the board of directors of Reddit. At the age of 33, he became the CEO of Openal.
In 2022, at the age of 36, the shocking artificial intelligence ChatGPT will be launched.
In November 2023, at the age of 37, he became the protagonist of OpenAI Gongdou, was fired from the company he founded, and returned to serve as CEO less than a week later.
From Professor Hinton's academic research and Altman's life experience, I have summarized two key words:Interdisciplinarity & toss-up.
Professor Hinton's background spanned physiology, physics, philosophy, psychology, and although he was not good at mathematics and computers, he eventually became a leader in artificial intelligence algorithms.
Altman, like many big entrepreneurs in the United States, chose to drop out of school. Dropping out of school, starting a business, changing industries, OpenAI Gong Dou, which attracted worldwide attention last month, was asked back by capital and more than 700 employees after being fired, and his tossing ability is a must.
Human wisdom has accumulated to this day, and in order to make significant breakthroughs and contributions, the mastery of interdisciplinary knowledge is becoming more and more important.
From Hinton and Altman, two representatives of the AI era, we can get a little enlightenment on the talent training in the AI era: either they have interdisciplinary innovation capabilities, or they can unite a group of people to make interdisciplinary breakthroughs.
Ask a thought: Did they win at the starting line?
Write at the end:
As a parent of a junior high school student, when preparing this article, I also asked ChatGPT questions related to the study of 3 subjects. ChatGPT's answers are smart and insightful (especially Q3's answer on memory).
I've put the chat history below so you can take a look at it.
Q1: With ChatGPT, do children don't have to learn English?
Q2: With ChatGPT, do children don't have to learn to code?
Q3: With ChatGPT, do children no longer need to memorize knowledge?
This article is authorized by EastWest Edu (ID: EastWest Edu).