Google is in "internal strife" because of its own ChatGPT

Now it seems that Microsoft, like Gao Qiqiang, knows a little about the art of war. Microsoft has adopted a strategy to weaken Qin: it has already started integrating Bing and ChatGPT, and then released it quickly, forcing Google to maintain it. Search for Big Brother's status and hastily accept the challenge.

Now this plan has shown results.

Belatedly realized, and hurriedly. In this competition, Google's first battle was when its stock plummeted, and now employees within the company are beginning to question it.

According to news and posts seen by CNBC and other media, Google employees are discussing the release of ChatGPT’s competing product, Google’s next-generation conversational AI system Bard, on the popular internal forum Memegen. Expressed negative views, calling it "rushed", "poor" and "un-Google-like".

As a response to ChatGPT, the release of Google Bard was indeed rushed. On Monday, Google CEO Sundar Pichai announced Bard unexpectedly, without even waiting for the Paris presentation event on Wednesday.

Why is Pichai in such a hurry? What happened on Tuesday seems to provide an explanation. On this day, Microsoft invited reporters to demonstrate the latest version of Bing search at the company's headquarters in Redmond, Washington. Working with OpenAI, the inventor of ChatGPT, Microsoft is adding an advanced AI conversational model to its search engine to support new versions of Bing and Edge.

As you can see from Microsoft’s demo, one of the new Bing search modes displays traditional search results and AI annotations side by side, while the other mode allows users to Talk directly to the AI ​​chatbot, and you can ask it questions in the ChatGPT-like chat interface.

"This is a new page for search. A new paradigm has emerged. The arms race starts today. We will move forward at the fastest speed and add new content every day. More importantly, we will A lot of good things will happen when we reinvent the search engine, and this is an important moment," said Satya Nadella, CEO and Chairman of Microsoft. Between words, Nadella seemed to be revolutionizing Google search.

But Google’s counterattack on Wednesday was not satisfactory. On this day, Google held the "Google presents: Live from Paris" conference. Everyone was looking forward to Bard's debut, but it was a huge disappointment. There was little in the way of Bard's new presentation at the conference, and some employees didn't even know about it.

Google’s disappointing press conference

When demonstrating the augmented reality search function, the speaker also lost the mobile phone used to demonstrate the demo, which attracted People smile awkwardly but politely.

Meanwhile, a negative story about Bard exploded on Twitter. Many netizens pointed out that Bard's first demo contained a factual error. In response to the question, "What could I tell my 9-year-old child about the new discoveries made by the James Webb Space Telescope (JWST)?" Bard provided three key points, one of which noted that "the telescope has captured images of extrasolar planets." First photo." However, some astronomers on Twitter pointed out that this was incorrect and that the first photo of an exoplanet was actually taken by the European Southern Observatory's Very Large Telescope (VLT) in 2004.

#On the day the news exploded, Google’s stock plummeted by 700 billion yuan overnight.

These disappointing moves also angered many employees within Google. While these employees often humorously poke fun at the company's quirks and missteps on Memegen, their posts after Bard's announcement took a more serious tone, even directly targeting their own CEOs.

Complaints from the outside world are nothing. Real criticism depends on our own employees:

"Dear Sundar, the launch and layoffs of Bard are hasty, clumsy and short-sighted," said one "Please return to the long-term outlook," employees wrote in the post. This post received many likes from employees.

"Sundar and his leadership deserve a Perf NI," read another highly liked post. Perf NI refers to the lowest category in the company's employee performance appraisal system. "In their pursuit of 'focus', they are ridiculously short-sighted. This is not 'Google' at all."

The jokes in these criticisms have their origins. Last year, Pichai called on employees to be more focused. Leadership often implores employees to become "Googlers," which, loosely speaking, means being ambitious, working hard, and remembering respect and teamwork.

Back to Bard’s attention-grabbing demo mistake. For pre-trained language models, whether it is OpenAI's ChatGPT or Google's Bard, they are trained based on massive data. It is not uncommon to answer questions with "gibberish". The high probability of Bard's error is that the knowledge in the training data involving "newly discovered issues obtained by the James Webb Space Telescope" is not clearly stated.

How to solve this problem? This usually means that the AI ​​model needs to optimize training data, perform parameter adjustments, special annotations, etc. before going online.

Let’s see what ChatGPT itself says. Image source: https://www.zhihu.com/question/582930903

The work that should be done is not done well, and the functions have to be integrated into the search business that requires high accuracy. Google is in great trouble. It will appear soon.

“The rush to bring Bard to market in a panic confirmed the market’s concerns about us,” read one highly liked post, which featured a picture of a bird covering its face.

Google’s position in the field of artificial intelligence is well known. At an all-hands meeting in December, as ChatGPT surged into the public eye, employees asked about the company's competitive advantage in artificial intelligence. Company executives responded that the company's reputation could be damaged if it moves too quickly on AI chat technology, which is far from perfect.

In a controversial speech some time ago, Yann LeCun, Turing Award winner and Meta’s chief artificial intelligence scientist, also expressed a similar view. At the time, LeCun was asked why Meta and Google hadn’t launched a system like ChatGPT, and he responded, “Because both Google and Meta would suffer huge losses from launching a system that made things up.” OpenAI seemed to have nothing to lose.

One of the most read posts this week showed an image of a trash can on fire with Google’s “G” logo on it. The caption reads: "My feelings since last year."

Another post addressed the layoffs announced last month. At the time, the company announced it was laying off 12,000 employees, or 6% of its workforce. The post featured a Nicolas Cage emoji and read: "Firing 12,000 people will increase the stock price by 3%, and a rushed artificial intelligence demonstration will decrease the stock price by 8%."

To put it simply, it’s hard to find

So what should Google do next?

The above is the detailed content of Google is in "internal strife" because of its own ChatGPT. For more information, please follow other related articles on the PHP Chinese website!