[账号注册] How to learn ChatGPT: not just prompt writing, and not just chatting—systematic learning methods and organized thinking

ChatGPT is an AI technology that can solve complex tasks. It is an NLP-based large natural language model, so chat-style commands are how it is used, but that does not mean it is a chatbot.
ChatGPT's capability is content output and screening—output and screening based on any form of text or strings. Therefore, generating a resume or an article is merely a relatively simple prompt-engineering task. Let us re-examine ChatGPT from the perspective of intuitive tasks, and you will find that it is both powerful and simple.

[baike]Basic tasks[/baike]
You should use it on the basis of being familiar with ChatGPT, with task objectives and ideas, so that you can stimulate its capabilities to the maximum. When you want to generate an article, a sentence like "help me generate an article about XX" obviously cannot accomplish the goal. You should bring in your own thinking and make requirements about the whole and the details of the content; this depends on your understanding of the content. So your prompt engineering should be something like this:

1. Help me write an article of XXX kind, with the title XX. I hope its role is XXX, and its structure is XXX. Please help me list an outline, and the outline should pay attention to XXXX.
2. This is the content structure I want: what to write first and what to write later. When writing about methods, please list XXXX and specify which symbols to use.
3. I hope you use XXX writing style; this is the reference content XXXX.
4. Please do not use XXXX. What should your wording be like, and what should the tone and emotion be like? Please refer to XXXX.

When facing a task, especially one that needs to incorporate our thinking, the task is extremely complex. It is not like 1+1=2, which is an absolute value; complex tasks are highly random and unpredictable. Take articles as an example: different people write in different ways, especially when it involves artistic literature, which involves people's emotions and viewpoints. This is relatively abstract and not easy to describe, and it also requires complex settings and reference content. If it involves your wording style, you also need to prepare a dictionary; if it involves trending topics, you also need to provide the trending topics. Do not expect ChatGPT to understand real-time trending topics. If you understand knowledge bases, you should know that its knowledge base updates lag behind.

[baike]Complex tasks[/baike]
In an era of diversified and personalized work, our needs are bizarre and varied, innovative, and extremely complex. To improve work efficiency, using only the web version of ChatGPT cannot solve complex process problems. In particular, enterprises now all need to reduce costs and increase efficiency. You should know better how to use AI to improve productivity, use automation, use APIs to complete automated work, apply AI technology to your own job, and solve most of your work there. You should step back and act as the commander, with AI as your assistant.

Carry out complex tasks based on the ability to handle basic tasks and achieve automated office work. Therefore, OpenAI launched the API. Everyone can use ChatGPT to write scripts and connect to the API to realize personalized workflows and solve work tasks. For example, if you have a large amount of customer data and need to analyze it to identify potential customers and create targeted marketing strategies for them, then the process of using the API is similar to this:

1. Use Python to process local data, retain and extract the main content.
2. Submit the data to ChatGPT through the API.
3. Let ChatGPT analyze the data; issue specific requirements through prompt engineering.
4. Return the data.
5. Connect with customers.

You only need to spend one day polishing this process, and it will completely change your way of working and greatly improve your work efficiency. You can have more time to learn and continuously optimize the details of this process. Tasks that originally took you a day to complete now only require a few button clicks.

[baike]Specific tasks[/baike]
In addition to routine tasks, there are also more professional tasks in specialized fields. Many high-tech fields keep their data confidential, and ChatGPT does not possess this knowledge either. When facing such tasks, OpenAI launched the capability to fine-tune large models. It can help you train private data and create personalized models to serve tasks in specific fields. For example,

private-domain customer operations are now popular among enterprises, and chat conversations take up most of the time. It is very appropriate to solve this problem through AI. We can turn ChatGPT into a chatbot, train the large model on the company's winning sales scripts, and create our proprietary chat model. Then, with prompt engineering, it can work like a person according to our settings. The robot can serve the enterprise without sleep or rest, and afterward you only need to continuously optimize this model, and cost reduction and efficiency gains will follow. Do not feel pressured: it will take jobs away from people who do not learn AI, but if you learn this technology, it will bring you promotions and raises.

[baike]Learning methods[/baike]
Through the above content, you should have a new understanding of ChatGPT. These technologies and ideas are actually practical experience. We have summarized more than 2 years of practical experience and research results into a set of enterprise office-level training courses that take beginners to experts in a short period. It can let you fully master this technology within 7 days to half a month. We guarantee you will learn it, and we will teach you until you do.

Those who need it can click here to learn about the course  https://www.niuboyi.com/peixun/