Baidu's Wenxin Yiyan product launch event was somewhat disappointing
After waiting for a long time, Wenxin Yiyan was finally released today. If nothing unexpected happens, we will receive an invitation code in a few days. Recently, many people in my circle of friends have been sharing a similar article: XX Company announced that it has integrated Baidu’s Wenxin Yiyan capabilities to create XX Company’s artificial intelligence product/service.
After confirming, the company that shared the article has not yet integrated Wenxin Yiyan, they only represent themselves to have obtained the first batch of internal testing qualifications. Similarly, our company has also received an invitation, but we did not use this method to promote ourselves. It is better to be honest, we have not yet integrated it and have not experienced any benefits.
In recent days, OpenAI has made remarkable progress, and the powerful performance of GPT has made us feel that singularity is approaching. Before, when talking to chatbots, I felt like I was talking to artificial idiots. But now, ChatGPT can fully understand our questions and even answer progressive follow-up questions.
New opportunities are coming.
Due to certain reasons, using ChatGPT’s API in China still has risks. Therefore, everyone is exceptionally looking forward to Wenxin Yiyan and its superior Chinese language abilities. Robin on the stage at the product launch event was still humble and a little nervous. His opening remarks lowered everyone’s expectations. Actually, it is known without seeing it that the gap between Baidu and OpenAI is certainly huge in large models.
However, we still want to know what level Baidu can achieve. After all, Baidu has been focusing on investing in artificial intelligence in recent years and has been referred to as the Huangpu Military Academy in China’s artificial intelligence field.
An hour after the start of the product launch event, with an extremely excited mood, I want to say that the videos before and after the launch event were great, and the adjectives and nouns mentioned in the videos were excellent. As a Chinese person, I sincerely hope that Baidu can succeed. In the rare era of large models, we entrepreneurs who are waiting to develop applications need good domestic companies to lead the way in large models.
The large screen on the scene of the product launch event was also very professional, and Robin was still as handsome as ever, with a white shirt and a straight figure. However, in the product demonstration part, the PPT used videos instead of a live demo. Okay, videos are also acceptable, but they ended quickly…
I have obtained the following information about Wenxin Yiyan:
Wenxin Yiyan can answer some basic questions and can also perform mathematical and logical reasoning. However, in the video, there was no demonstration of its ability to answer follow-up questions, which is obviously more important.
It can understand English, but its understanding of English is not as good as Chinese.
It supports multi-modal generation, including text, images, audio, and video. The audio part also incorporates Chinese dialects.
It can generate text based on commands. But it is unknown whether it can further optimize text based on instructions.
Baidu believes that in the era of large models, entrepreneurs have great potential in application development directions. For example, text, image, audio and video generation, digital humans, 3D generation, etc.
Baidu basically has the technical ability of large models, but the training is not sufficient at present, so the product is not yet perfect. It needs everyone to give it some time.
There is still some disappointment. When launching a product, it is really unnecessary to talk so much theory. Talk is cheap, show me the product. Recently, I watched Sam Altman, the leader of ChatGPT, speak in a public entrepreneurship course at Stanford University. He repeatedly emphasized the importance of creating a great product as the driving force for success in entrepreneurship.
Some entrepreneurs come from a sales background and believe that building a sales team is the key to entrepreneurial success, so they rush to expand their sales team before their team is mature. Others come from an operations background and excel in various growth hacks, enjoying planning various activities to acquire users.
Sam’s view is that only by developing excellent products can everything else have meaning. As long as an excellent product is developed, other problems that need to be solved, such as fundraising, increasing visibility, recruiting employees, and business expansion, will be resolved easily. Therefore, the focus of entrepreneurs should always be on communicating with users and iterating products. It seems that ChatGPT operates based on this logic, having spent ten years honing its product before becoming famous.
His argument is not necessarily absolutely correct, but at least it can be seen that this person has a passion for product quality. This is what we often say, you reap what you sow.
The launch event is over.
It’s time to get back to work. As fellow entrepreneurs, I feel that the location of the product launch event, the size of the event screen, and the attractiveness of the PPT are not that important. If one lacks a name, they should focus on practicing swordsmanship.
This article is not intended to ridicule or criticize anyone. I am still patiently waiting for an invitation code for Wenxin Yiyan, hoping that it can make great strides like OpenAI. After all, at present, if Baidu cannot do well in large models, it may be difficult for other companies to succeed. I hope Baidu can succeed and that we can use Wenxin Yiyan’s API soon to do our own thing.
