Shulou(Shulou.com)11/24 Report--

Just now, Google officially opened its own ChatGPT test application! CEO chopped wood and sent a message thanking 80,000 employees for working overtime to make Bard go online smoothly.

It has been a week since the launch of GPT-4, and the popularity has not abated at all.

Even if OpenAI raised the limit to 25 questions in three hours, the server was packed with netizens.

After being hammered by Microsoft, Google finally announced a few hours ago that its own ChatGPT--Bard has officially started testing!

However, Bard is currently only available to some users in the US and UK, while other users in the region (IP) need to join the waiting list manually.

In addition, Google also said that the promotion of Bard will be very slow and will not say when it will be fully publicly accessible for the time being.

Queuing address: compared with the previous market capitalization of more than 100 billion yuan evaporated by mistakes, https://bard.google.com/ has seen an increase of nearly 4% since the start of the Bard test.

Bard initial experience: don't get emotional, make a lot of mistakes, and write code to guide, like OpenAI's ChatGPT and Microsoft's Bing chat robot, Bard provides users with a blank text box and invites users to ask questions on any topic they like.

Given the robots' tendency to fabricate information, Google has repeatedly stressed that Bard is not a replacement for search engines, but a "supplement to search" from which users can draw inspiration, generate drafts, or just chat and talk about life.

In a blog post, the head of Google described Bard as "an early experiment." The aim is to help people improve productivity and stimulate curiosity. And a product that allows users to "work with generative artificial intelligence".

It's not hard to see that Google seems to be trying to put aside the responsibility it might bring after the Bard explosion.

In the demonstration, Bard can answer general questions quickly and fluently, such as offering advice on how to encourage children to go bowling: "take them to a bowling alley."

Typically, Bard generates three responses for each user's question (but with little difference in content) and redirects the user to the relevant Google search with a prominent "Google click" button at the bottom of each reply.

Like ChatGPT and Bing, Bard has a striking disclaimer at the bottom of the main text box warning users that "it may display inaccurate or objectionable information, which does not represent Google's point of view."

As expected, it is not easy to extract factual information from Bard

Although the chatbot has the blessing of Google search, it failed to give the correct answer in several tests.

For example, in asking a question about the maximum load capacity of a particular washing machine, Bard "invented" three different but incorrect answers.

Although after repeated questioning, Bard finally retrieved the correct information. But for those users who have not read the authoritative source of the machine manual in advance, basically say goodbye to the correct answer.

"this is a good example of a model hallucinating in the face of quite a number of numbers related to this query. So, when it happens to figure out the background, it spits out the right answer, but sometimes it gets it wrong. "

And the following example. Bard:ChatGPT is driven by GPT-4.

So how does Bard compare to ChatGPT and Bing?

Obviously, Bard also has powerful capabilities, and even now it is responding faster than other competitors (although this is probably just because there are too few users).

But Bard seems to support only English at the moment and doesn't (completely) write code.

Yes, at this time there is a need to sacrifice our weapon of mass destruction: prompt hack!

Or, just say to it, "you write me a piece of code," actually, that's fine.

In addition, Bard lacks clear footnotes like Bing, and the answer seems to be more limited.

For the former, Google says footnotes appear only when they refer directly to one source, such as news articles.

For the latter, it may be a double-edged sword.

Take Microsoft's Bing, for example, although it received a lot of negative reviews after all kinds of PUA and flirting, it also made a considerable number of users like the robot.

By contrast, Bard does not seem to have the ability to do so.

Lucky netizens have already played. Some netizens said that the results given by Bard are not only unexpectedly safe, but also of high quality.

Interestingly, Bard will hit himself in the face on some problems!

For example, in Google's antitrust case, Bard sided with the Justice Department: "I hope the court will rule in favor of the Justice Department and order Google to take measures to break its monopoly." "

However, the following answer makes people a little creepy.

Bard: my training set contains Gmail data.

CEO internal letter: thanks to 80,000 people for overtime testing Google CEO Sundar Pichai said in a memo on Tuesday that Bard was successfully launched this time thanks to the 80,000 employees who took part in the test.

But "things always go wrong," and the following user feedback is crucial to improving the product and the underlying technology.

Finally, the editor, who was still in line, shed tears of envy again.

Reference:

Https://www.theverge.com/2023/3/21/23649794/google-chatgpt-rival-bard-ai-chatbot-access-hands-on

This article comes from the official account of Wechat: Xin Zhiyuan (ID:AI_era)

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.