|
|
Baidu released Wenxinyiyan, a large language model, on March 16. Although this product is still in the beta stage. Many people are curious about the gap between Wenxinyiyan and Chat GPT. If Wenxinyiyan can fight Chat GPT, the difference between who wins and who loses is clear at a glance. We asked Open AI’s Chat GPT and Baidu’s Wending to comment on each other. Chat GPT gave this answer: Obviously, such an answer is not enough for us who want to make trouble. Turning to Wenxinyiyan, facing the same question, its answer is. The two artificial intelligences have shown amazing friendliness, and it seems that pulling and stepping is not popular in the AI field. But we were not satisfied with this kind of friendly answer. We decided to be more blunt and let them have a direct “bicker”. When faced with human dissension.
The answer given by Chat GPT is: And Wending said: Facing AI that has no emotion but respects each other, this round is the defeat of despicable humans. Considering that Microsoft’s New Bing is also based on GPT-4, but the addition of a search engine cell phone number list makes it slightly different from other applications. We decided to ask it to see how New Bing answered questions. The first question, of course, is the classic mutual evaluation. The answer given by New Bing is: New Bing’s answer is not “I can’t” but “I don’t want to”, which is more emotional. And when we asked New Bing to argue with Wen Xin, New Bing’s answer was completely beyond our expectations. Its answer is this: Not only did New Bing not answer our questions, it closed the conversation itself.
After these two rounds of questioning, we also gave up the unfriendly questioning method of sowing discord. The three large language model applications all mentioned “application scenarios”, which is also a step that must be considered in the current AI development process, and directly affects all aspects of design, development, deployment, and optimization. The “universal” attribute brought about by the emergence of Chat GPT has made people . gradually ignore the existence of “application scenarios”, or in other words, infinitely magnify the application scenarios to all walks of life and all aspects, and become omnipotent. After this test, we can’t see how big the gap between Wenxinyiyan and Chat GPT is from the quality of the dialogue, but New Bing with emotions feels slightly better. But even the AIs are respecting each other, which makes us humans who conduct comparative tests a little embarrassed. |
|
|
发表于 2023-3-29 12:12:42
