Google’s AI chatbot Bard, which is based on ChatGPT, has opened internal testing.

The last time Bard made a fool of himself on a live broadcast around the world, he made a serious mistake in answering an objective astronomy question, sending Google stock plummeting. After a short-term improvement, the internal test was finally opened. Google executives valued this project very much, and the CEO personally tweeted to help promote it.

Different from ChatGPT’s GPT 3.5/4 model, Bard uses Google’s own LaMNDA model. There is no difference in the interactive interface. Google Bard will generate three different answers for users to choose (what multi-ending plot):

Of course, if you are not satisfied with the three answers, you can also directly refresh and regenerate the answer. In addition, Bard also integrates a Google search button. If you are not satisfied with the AI’s answer, you can directly search for content on the Internet.

For domestic users, Bard has a big problem:Chinese is not supportedand secondly, it can’t write code… But Bard’s content is relatively new, and it can also recognize the content of the past two years.

At present, Bard is only open in the US and UK regions. The application requirements are similar to Microsoft’s Bing Chat. You only need a Google account to click and queue.

Application address: https://bard.google.com/