has always been talking about foreign AI, today talk about domestic AI, as the first domestic search engine, once the Internet boss Baidu, also launched its own intelligent chatbot: Baidu search AI partner (hereinafter referred to as Baidu chat), link: https://chat.baidu.com/, or click the AI icon in the upper right corner of Baidu's home page to enter. If you can't open it or don't have permission, try logging in to Wen Xin to join the test application to try. Link: https://yiyan.baidu.com/. The Baidu chat dialogue interface is shown in the figure below.
The function is basically similar to chatGPT and other large language models, I didn't say write code at first, I let it write one, Baidu chat also wrote out, it seems to have the function of writing code
Click on the link Inspiration Hub in the lower left corner to open all apps, out of 124 apps. In fact, it is to set a number of specific roles for Baidu chat, specifying the generation of content suitable for specific scenarios, Baidu may do this may think that users do not know what Baidu chat can do, or do not know how to use Baidu chat, so give a breakdown of the use scenarios. The main ones used are AI painting, AI writing, office assistant, generating entertainment content, etc. Whether a thing is good or not, compare it to know, let's compare the use effect with the Bing chat chat program.
Try generating content specifically. Open the Xiaohongshu copywriting creation app, enter the copywriting theme you want to create, the system prompts such as "lipstick number Amway", we will enter: "lipstick number Amway", the generated result is more in line with the style of Xiaohongshu notes, and also gives emojis and tags,
It is said that this function should also be able to be realized in the Baidu chat dialog box, and the same content was entered in Baidu chat, I asked twice, the first answer was relatively brief, I looked at the reference, AI is a direct summary and copied this article of Baijia number, one is not more than one less, the order is unchanged, and the article of Zhihu has nothing to do with it. What I wanted was the copy of Little Red Book, I asked the question again, and the answer given the second time was a little richer. But the answer was the same two times, and I still referred to the same article. If you ask 1 + 1 equals a few, this kind of answer is fixed, it is normal to ask several times, but let you recommend lipstick, this and there is no fixed answer, color, season, price, brand, occasion can refer to more, you still always give the same answer, I feel that this is not very good, which shows that its knowledge base reserve is insufficient and intelligence is not high,
Try it in Bing Chat, I first ask: "lipstick number Amway", the result is the same as Baidu chat, but a few lipsticks are listed and given a sentence or two introduction. Then I asked: "Help me write a little red book copywriting, theme: lipstick number Amway", this time the answer is more detailed, first a few sentences of guiding words, and then by color classification to recommend different brands of lipstick with lipstick instructions. Finally, there are concluding remarks. I feel that this is different, very professional, not like Baidu chat to write copywriting casually to fool the past, and Bing chat has many reference articles, each time is a different article to refer to different answers to give different answers.
Let's experience Baidu chat and Bing chat article writing ability, let them write an article about the possible impact of artificial intelligence on human beings in the future, the number of words is about 3,000 words, Baidu chat wrote 638 words on the end, as shown in the following figure, I let it continue, it generated 3 more paragraphs, but still similar to the above 3 paragraphs, said medical field and industrial field, did not write new content and sentences.
Microsoft Chat also failed to write 3,000 words, plus the supplementary negative impact summary only wrote 1916 words. But the article written by Microsoft chat is quite exquisite, as shown in the figure below, to list the article title, article content classification summary, paragraph title, written in detail, it seems that the reading experience is good.
Let's experience the ability to understand Baidu chat and Bingchat. For example, ask: "He just came, and many people don't know him yet." How to understand this sentence. This is an ambiguous sentence, it means that he just came, many people still don't know him, or that he just came, many people he doesn't know. Both Baidu Chat and Bing Chat understand only one layer of meaning.
Another puzzle: there is a sheep, a wolf, a basket of vegetables and a person to cross the river, the wolf will eat the sheep, the sheep will eat the vegetables, but the boat is too small, people can only transport the same at a time, how can they cross the river safely? This is a classic intelligence problem, these AIs should have learned this knowledge, see how Baidu chat and Bing chat answer. As shown in the figure below, Bing Chat's answer is correct, Baidu's answer is more chaotic, the wolf has not yet passed to bring it back? Answered indiscriminately, I don't even know what I said... This is still a classic puzzle.
Let's do the calculation problem, a manager has three daughters, the ages of these three daughters add up to 13, and the ages of the three daughters are multiplied to equal the manager's own age. At this time, one of the manager's subordinates already knew the manager's age, but was still not sure of the age of the manager's three daughters, and the manager suddenly said that only one of his daughters had long hair (indicating that the long-haired daughter was older). What are the ages of the three daughters? Why? Look at the answers of Baidu chat and Bing chat, as shown in the figure below.
Baidu calculated that the three daughters were 3, 3 and 7 years old, and the manager was 63 years old. This manager gave birth to two daughters at the age of 60... If the manager is a woman, the risk is quite high and the difficulty is quite high. The key is that Baidu assumes that the manager is between 30 and 50 years old, why does it still give the manager's age to 63? The calculation result of Bing chat is that the age of the three daughters is 1, 3, and 9 years old, the manager's age is 27 years old, and the first daughter was born at the age of 18, which cannot be said to be impossible, but its reasoning is a little unreasonable, why is one daughter a little older, and the other two daughters cannot be the same age? The best answer to this question should be the age of three daughters aged 2, 2 and 9, and the manager age of 36.
Here's the painting experience. Let the AI draw a picture of two people playing Go, Baidu provides 1 painting result, what does it mean that two people face the same direction? Don't you have to face the board to play Go? Feel like Bing is providing better results.
Baidu chat inspiration center also a pseudo-original rewrite, it feels very interesting, Baidu is to do content, whether it is Baijia number or for search engines, Baidu is to encourage originality, you provide a pseudo-original rewrite what does it mean, a little with their own spear to attack their own shield feeling, I don't know that the pseudo-original article it generates can pass Baidu original review No, if it can pass, everyone can use this tool to carry the article, if you can't pass, what pseudo-original you are engaged in, funny.
Let's test so much first, everyone can experience it for themselves if they like it, it can't be said to be better than others, but it's better than not, domestic AI is also made, but there is still a long way to go, and the pace is fast.
Original: https://dyss.top/957