MLS Football Jersey

Since 2023, Google has been releasing a succession of faulty AI products, leading to questions about the reliability of its AI technology and products, and triggering a negative impact on the share price of its parent company, Alphabet.

In February 2023, Google’s large-scale language model Bard, released in Paris, suffered a factual error during a demonstration, causing Alphabet’s market value to shrink by $100 billion.

In December 2023, Gemini, another large language model released by Google, is accused of excessive editing and exaggerated effects in a demonstration video, raising questions about its true capabilities.

In February 2024, shortly after the release of Gemini 1.5, it was upstaged by Sora, an avatar owned by Vincennes Video.

That same month, Alphabet shares fell 4.5 per cent after Gemini-generated images of historical figures were accused of being overly diverse and discriminating against white people.

Google’s AI product releases have frequently flopped, not only exposing its shortcomings in technology development and product management, but also casting a shadow over its future development.

Google, the face of Silicon Valley, the giant of AI, why does something unexpected always happen?

Recently, semi-retired Google co-founder Sergey Brin, who attended the San Francisco Hackathon, admitted to the problems with Gemini, giving a middle-of-the-road reason: incomplete testing.

After Google’s acquisition of DeepMind, DeepMind and Google Brain had different organisational cultures and longstanding rivalries, before hastily merging into Google DeepMind in April last year in order to fight OpenAI.

After the epidemic’s brief bonus period, Google realised the bad side of too many people, and announced in January 2023 that it was laying off 12,000 employees, accounting for 6% of its total global workforce. Layoffs of this magnitude had not been seen in the last 20 years.

Google’s AI Stumbles: A Sign of Things to Come or a Temporary Setback

The layoffs destroyed the trust of the employees, and when survival becomes the number one priority, innovation and self-worth are difficult to achieve. Every employee is trying to find a way to join the team least likely to be laid off, to keep their job first.

More than a decade ago, Google was not like this.

At that time, Google was an engineer’s paradise, advocating a bottom-up “grassroots culture” and “20% of the time” outside of their own work, employees found a problem, and then look for like-minded colleagues to roll up their sleeves and work hard.

Now, with the thriving geek atmosphere gone, the company’s scale is bound to expand, and Wall Street and investors are bound to make good connections, so the unfulfilled ambition of the talent has made its own choices.

The Transformer paper for ChatGPT came from Google, but as of July last year, the eight authors had left Google, some of them started their own businesses, and some of them joined OpenAI to go to the starry sea of AGI.

When it comes to the reasons for their departure, they coincidentally expressed their dissatisfaction with Google’s bureaucracy – not allowing risk-taking, not allowing the rapid introduction of new products.

It’s hard to maintain a kingly position, but it’s easier than you might think to be broken from the outside. Google’s past is, in a way, also a shackle, compared to a startup that was built from the ground up.

Google’s main revenue is still search advertising, although also through Google One to provide AI products subscription model, but this part of the revenue is still minimal, and OpenAI, Perplexity and other companies to compete for market share.

Google should understand better than anyone what will happen if it lets AI eat the search cake and lose its core cash cow. Consultancy Gartner predicts that traditional search volumes could fall by 25 per cent by 2026 because of AI alternatives to search such as Perplexity.

Still, Google is one of the most AI-savvy companies in the world.

Google is also bringing more to the table than search, with native multimodal Gemini 1.5, the open source model Gemma, and the base world model Genie……

The development of generative AI is like a child’s imagination, full of infinite possibilities, but also full of unknowns.

Last December, Google released Gemini, which showed the potential of generative AI. Although the demo video was said to be exaggerated, the realistic human-computer interactions demonstrated by Gemini were still fascinating.

After more than 20 years, the way the Internet is accessed and used is once again being disrupted, and the rise of startups such as OpenAI is a reminder of what Google’s co-founders were all about.

AI has rekindled big companies’ passion for innovation, and they are stepping down from their customary judges’ chairs to emerge as challengers. 49-year-old Microsoft is making a strong comeback, and 25-year-old Google still has a chance to win again.

However, the future development of generative AI also faces some challenges, such as technical ethics, data privacy and other issues need to be properly resolved.

How will generative AI develop in the future? Let’s wait and see.

Leave a Reply

Your email address will not be published. Required fields are marked *

MLS Football Jersey