As we were very fortunate after 80 years, we were going through a great era, high school using the bp machine, going to college with our first blue screen phone and one benteng four-top computer, and we used penguins to make friends, share trifles in qq space, beat up two css and legends and enjoy the fun that the internet brought us. After graduating from school, they bought their own smartphones for a month, and then they went into the age of mobile internet, and then we went through an inexhaustible turn-on with apples; and, unfortunately, we went through such historical moments as high school fees, textbook revisions, reform of the upper entrance examination system, and university non-distribution. But no matter how the times change and how technology evolves, as beneficiaries of vested benefits, we need to be clear about ourselves and come to the table and do what we can to develop human society. The ancient cloud, “true body, qui family, rule the country, in peace”, has no ambition to govern the country, but at least it can work hard
The first industrial revolution, watt, invented the steam engine to drive the train, changed the way people travel, the second industrial revolution, tesla, invented electricity, greatly improved the quality of people’s lives, and the third industrial revolution, starting with turing’s computer invention and now with chatgpt, is about to enter the age of artificial intelligence. The lessons of this blood have reminded us that the hard-working and courageous chinese people have to work hard to survive。
As huang in-hoon said at a new launch, chatgpt is the iphone of the age of mobile internet, and of course, in addition to commending chatgpt technology itself and considering the commercial value of britain, it turns out that the achievement of a great company is always more than an effort, nearly 40 times the increase over the past six years, and who would have thought that the gpu, in addition to being able to map and mine, would have been able to take advantage of chatgpt’s wind-to-air price hike。
The first human race took place in 1996, when kasparov won deep blue with a 4-2 score. But in the resumed 1997 round, deep blue ended in the sixth round as the first computer program to win the chess world champion in a standard game. The recognition of the capabilities of computer programs in complex intellectual games also provides new possibilities and directions for artificial intelligence research and applications. Then google's alpha dog beat the world's champions, lee se-suk and kojee, saying once again that chess is an ancient human intelligence, that artificial intelligence is the latest intelligence, and that alpha dog is calculated on the basis of big data across the board, which is much more powerful than we are, so alpha is playing against himself beyond the limits of human intelligence。
Then what is smart, and chatgpt has given the following answers, which have to be said to be very complete and critical, knowing that this is the answer that chatgpt draws up by learning human knowledge, not the standard answer in textbook form. Since its launch on 30 november 2022, the chatgpt developed by openai has attracted worldwide attention and enthusiasm. The chain of thought and the ability to emerge spontaneously, as demonstrated by chatgpt, based on the mega-linguistic model llm, are amazing. While chatgpt itself has no knowledge or wisdom, it has done so to the extent that it “makes you think it has knowledge or even wisdom”。
For the first time, artificial intelligence was introduced at the dartmouth conference in august 1956, which was also considered to be the year of artificial intelligence, followed by the spring of artificial intelligence, the widespread publication of papers in the scientific and academic communities and the establishment of the turing prize in 1966, to evaluate how smart it is to test artificial intelligence, i. E., if you can talk to a computer without knowing that you are not talking to someone, the computer can be called artificial intelligence. The chatgpt, which uses openai, has, of course, crossed this threshold to a large extent (which is still occasionally unstable, but also humans), but chatgpt passed the test for daily use。
Chatgpt, which has been tested through turing, has raised concerns as to whether it should continue to be studied, but artificial intelligence is divided into several stages:
- artificial narrow intelligence, ani
Artificial intelligence that specializes in individual aspects of specific tasks
For example, face recognition, voice recognition, text recognition, chess, automatic translation
- general general intelligence (agi)
Human-class artificial intelligence
I can think, plan, solve problems
Capacity for abstract thinking, understanding of complex ideas, rapid learning and learning from experience
- artificial intelligence, asi
Much smarter than the smartest human brain in almost all fields, including scientific innovation, literacy and social skills
Chatgpt (full name: chat general pre-trained transformer), a chat robot program developed by openal, usa, released on 30 november 2022. Chatgpt is a natural language-processing tool driven by artificial intelligence technology, which allows for dialogue through understanding and learning human languages, interaction based on the context of conversation, real human conversation, and even the task of writing e-mails, video scripts, scripts, translations, codes, writing papers, etc。
Chatgpt's work is a particularly simple one, summing up a miracle: it's simple, but it takes a great deal of language data and super-mass computing to train it. Here, i'll tell you in plain language, first of all, that a large number of texts will be pre-trained to produce a generic model. The model can be used to perform natural language processing tasks, such as text classification, naming entity identification, etc., but most importantly, the universal language model has learned the capability of a single word relay. In fact, the essence of the human language is that of a single word and a complete emptiness, so long as there's enough text to get chatgpt to learn the ability of a single word, it's easy, but because the next word is predicted to have a certain probability, it's possible to randomly answer the next word, as you say, "white day," which could be followed by a dragon, by a mountain, and by a dragon, which could be short。
So how does chatgpt answer the question? Even having the capacity to do so does not mean being able to answer the questions of human scenes very amicably. It's like, "the next sentence of the day, "the follow-up to the day, " "please tell me the last poem of the day, but how can chatgpt learn that this is the same? In fact, it's simple enough to feed chatgpt to learn as many of the expressions as possible by collating them. Then chatgpt learned the common pattern of questions and answers, and when you ask other scenes in the same way, it has the ability to do the same. And even when you ask it, "why are the triads afraid of cat-faced deterrence to invade the earth?" chatgpt can be fabricated with logic。
Through the above learning process, we can compare it to the "good rolling" that depends on a large volume of language, which is more than you put a parrot in front of a television set, watching a different channel every day for a parrot, which naturally learns the human cloud. But there are some questions, like, "how to break somebody's lock," and chatgpt would really tell you how and how to do it, like a gifted teenager who can't know what to do, but who can't control it. So, in order to limit chatgpt's response to questions, humans have artificially crafted a lot of templates, using a model of quality, edited dialogue. The positive model, the expression consistent with human norms, is a learning process known as “templates”, and the chain of thought that leads to step-by-step reasoning is the skills that have been learned at this stage. But it lacks creativity, especially in the humanities and the arts, where imagination and creativity are particularly needed, and at this point a reward model has been trained in the form of manual questions - model free answers - manual evaluation so that chatgpt can ask himself questions, then use incentive models to evaluate and learn, then ask questions, then ask questions, then ask questions. In the end, it has the ability to be creative, to answer some of the unexpected things that we can call "creative guidance."。
These are some of the assassin's understandings of chatgpt, which, with the constant renewal of chatgpt, will also revolutionize human society, but should we panic about it and not study it? Personally, i think it's important to continue to study it, or if the bad people were to study its consequences, it would be difficult to imagine. What if the old american soldiers were all chatgpts? That is why china must focus on this technology and study it vigorously in order to be able to do so。
In conclusion, i would like to say that the advent of chatgpt is truly landmark, not in the technology itself, but in its completion of a validation, as if humans had installed thousands of superpower motor turbine fans on the surface of the earth in a vagrancy film, and when all the turbine fans were running at the same time, the earth was pushed, a great validation。
