Family, DeepSeek has been out of the circle recently! What the hell is this DeepSeek? It was developed by our China Deep Quest Company, which was only established in 2023 years, but it focuses on the research and development of artificial intelligence technology. It uses reinforcement learning technology, which is amazing, even if the labeled data is very small, it can also improve the model inference ability, and the training cost is ultra-low, only one-tenth of ChatGPT, a proper "domestic light" in the AI industry!

So what's the difference between it and ChatGPT? DeepSeek excels in math, code, and natural language reasoning, especially in Chinese. ChatGPT, on the other hand, stands out in terms of conversation, text generation, language translation, and a rich ecosystem and a large user base.

Below I will share with you six tips to get you started with DeepSeek.

The first tip is not to use prompts, but to ask questions sincerely. DeepSeek is an inference model that is sensitive to prompt words, and those old prompt word templates will make the AI answer very mechanical. For example, if you want to write a report, you can express your needs directly and sincerely, such as "help me sort out the report and make a table report to the boss, he values the data", which is much better than a template.

The second trick is to use formulas to get accurate answers. When using DeepSeek, you can refer to this formula: "What do I want to do + Who do I want to use it for + What effect do I want to achieve + What problems do I worry about". Ask questions according to this formula, and the answers you get will definitely be more in line with your expectations.

The third trick is to let it "speak human", and the answer will be easy to understand. If the AI answers too professionally and abstractly, you can add "speak human", and it will explain to you with vivid examples, adding these three words, it is much easier to understand, and it can be used in both learning and work scenarios.

The fourth trick is to reverse PUA and optimize the answer logic. You can use DeepSeek's thinking logic to generate a plan, and then ask for reasons for objection and ask for a review, so that the answer will be more organized and comprehensive.

The fifth trick, imitation assistance, can unlock amazing content. Ask DeepSeek to mimic the style of the article or the tone of the characters, and you can produce high-quality content. For example, imitating the well-known tone and imitating the keyboard hero sharp evaluation model, the effect is called amazing.

The sixth skill is to think deeply about instructions and improve the quality of answers. Adding "critical thinking" or "reviewing a hundred times" to the question allows the AI to think deeply, actively examine the answer, analyze it from multiple angles, and give a more complete and practical answer.

Family, try these tips and make DeepSeek work for you!