Author | Wang Ruiping
Reviewer | Yun Zhao
No one could foresee the arrival of ChatGPT. But it's finally here, causing quite a stir and potentially upending the entire industry.
To some extent, ChatGPT may be the fastest growing application in history, with more than 100 million active users in just over two months, which shows how popular it is. welcome! Recently, Jack Ma, who returned to China, also expressed his views on ChatGPT: "Technology such as ChatGPT has already brought challenges to education, but this type of technology is only the beginning of the era of artificial intelligence. We must use artificial intelligence to solve problems, rather than being Artificial intelligence control."
Recently, ChatGPT's model has improved its actions, and its performance has been improved in all aspects, especially the text box, which has better user experience. Not only that, OpenAI has also further created a new operating system, which greatly lowers the threshold for programmers and further provides them with convenience.
Source: Open AI
It is understood that previously, ChatGPT’s model only interacted in a conversational manner. This conversational format enables ChatGPT to answer a series of questions, admit mistakes, challenge incorrect questions and reject inappropriate requests.
Only four months after ChatGPT was launched, it once again refreshed everyone’s understanding of text boxes. It proves that the future of technology isn't the Whiz-Bang interface or the Metaverse, but "typing commands into a text box on your computer." The command line is back - now it's a lot "smarter", which is great news for users.
Let’s talk specifically about what’s going on! For years, most users could only interact with computers by typing information into a blank screen (the command line was used to tell the machine what to do).
In order to improve this interaction method, the company has optimized the previous interface! It turns out that you need to enter something at the command line and consider the order in which you enter it to make the computer run. Now, you just click on the big icons on the new interface, and you can tell people what the computer can do through pictures and icons, which is much simpler. Simply put, the command line gives way to the graphical user interface, as shown in the picture above.
Users who have used the ChatGPT large model know that it can understand needs that have never been seen before, and follow them according to the requirements. Ask for tasks to be completed automatically. Because there are enough parameters, it can overcome the limitations of single domain and multiple models and achieve "intelligent emergence".
In addition, due to the superior code generation capabilities of ChatGPT and GPT-4, the coding-led software 1.0 era has quickly entered the data-driven coding software 2.0 era.
For programmers, their daily coding work has changed from using open source integration in the past to currently using AI combination. 80%-90% of the work every day can be completed using ChatGPT.
Specifically speaking, most of the tools used by programmers in the software development cycle from writing code to development, architecture module design, module design, testing, etc. will be reconstructed by ChatGPT and regenerate.
As we all know, every industry is empowered by artificial intelligence, which also allows developers in various industries to introduce plug-in Prompt through LLM/GPT, and then use Natural language conversation. This greatly lowers the threshold for programmers, expands entry points, and makes ChatGPT a new generation operating system.
Here, if you want to better embrace the era of artificial intelligence, I would like to make some suggestions for developers: first, you must learn to express and make good use of prompts, secondly, your English level must be excellent, and finally Use your imagination and get started! (Haha, you’re doing something special!)
ChatGPT is so popular that it has also attracted many imitators. The two most recent ones are They are the open source large language models ChatGLM and Google Bard. Now let’s talk about them in detail:
Recently, another large domestic language similar to ChatGPT has emerged The model ChatGLM came into our view, and fortunately it is open source!
ChatGLM-6B also supports a Chinese-English bilingual conversational language model. It is based on the General Language Model (GLM) architecture and has 6.2 billion parameters. It is not too convenient!
From a technical perspective, ChatGLM-6B uses similar technology to ChatGPT and is optimized for Chinese question and answer and dialogue. After about 1T identifiers of Chinese and English bilingual training, supplemented by supervised fine-tuning, feedback self-service, human feedback reinforcement learning and other technologies, the 6.2 billion parameter ChatGLM-6B has been able to generate answers that are quite consistent with human preferences.
Combined with model quantization technology, users can deploy it locally on consumer-grade graphics cards (a minimum of 6GB of video memory is required at the INT4 quantization level).
ChatGPT has more parameters than the Bard model, allowing it to capture more and more complex relationships in the language, but this also Brings the risk of overfitting.
Objectively speaking, Google Bard may not be as flexible as ChatGPT, but it may also be more powerful because of new language use cases. It is worth emphasizing that both Bard and ChatGPT models are deep learning neural networks based on Transformer.
So, the question is, who will benefit from ChatGPT? Who will be left behind? There are two different views.
The optimistic view is that for ordinary skilled workers, ChatGPT will prove to be a powerful tool that can improve their capabilities and expertise while promoting overall economic development.
The pessimistic view is that companies will simply use it to replace jobs that once did not seem to require automation, as well as high-paying jobs that require creative skills and logical reasoning. The end result is that a handful of high-tech companies and tech elites will become richer, which will do little to help overall economic growth.
In summary, ChatGPT and other recently released generative AI models have the potential to automate a variety of tasks. These tasks were previously thought to belong solely to the realm of human creativity and reasoning. From writing to creating graphs to summarizing and analyzing data.
Here, as a digression, from an economic point of view, the era of artificial intelligence has arrived, and an artificial intelligence gold rush has begun. Everyone from app developers to venture-backed startups to world-class companies are rushing to tap new business opportunities from generative AI models like ChatGPT.
As businesses scramble to find ways to use ChatGPT technology, the opportunity has opened to rethink how to get the most from a new generation of artificial intelligence, economists say. In the future, more "killer" applications will appear to improve the efficiency of daily production work and better benefit mankind.
https://techxplore.com/news/2023-03-chatgpt-struggles-wordle-puzzles-lot.htmlhttps://techxplore .com/news/2023-03-watermarking-chatgpt-dall-e-generative-ais.html
https://www.technologyreview.com/2023/03/25/1070275/chatgpt-revolutionize -economy-decide-what-looks-like/
https://m.ithome.com/html/682618.htm
https://openai.com/research/gpts- are-gpts
The above is the detailed content of The ChatGPT text box has been upgraded again to create a new operating system.... For more information, please follow other related articles on the PHP Chinese website!