Yet Google now finds itself in an extraordinary predicament. It is flailing under the financial and operational burdens of rapidly developing new generative artificial intelligence (AI) systems. These systems have been built with trillions of dollars worth of investments, coupled with significant energy investments, with returns that are still not fully realized. Even with these challenges, players such as OpenAI have achieved incredible breakthroughs, as seen by the rapid proliferation of ChatGPT. By some estimates, a million users have already interacted with ChatGPT—despite this tool’s propensity to hallucinate when presenting factual information and performing math calculations.
By the close of 2022, Alphabet, Google's parent company, witnessed a 39 percent decline in stock prices compared to the previous year's end. Google has begun making improvements to its own AI system, LaMDA. We can improve its output by refreshing the knowledge base. Plus, we can use that to adopt new safety measures to make it more competitive as an AI leader.
In a strategic shift, Google has reassigned some contractors who previously focused on reporting child-abuse imagery to test Bard, an AI initiative. Sundar Pichai, Google's CEO, has encouraged employees with available time to contribute to Bard's testing, signaling the company's commitment to advancing its AI capabilities. While folks inside the company and outside observers agree that Google’s big-footed, put-your-dancing-shoes-away approach to AI is dragging it down.
Significantly, one of the leading AI theorists in the world, James Manyika, an Oxford-trained roboticist and McKinsey consigliere, sounded alarm bells on Google’s reticence on AI. Conversations with Pichai prior to ChatGPT's public release echoed this sentiment, emphasizing the need for more decisive action in AI development.
All of that skepticism aside, Google continues to be the 800 pound gorilla of the tech industry. It currently provides seven services, each with more than 2 billion users as customers. These are some examples of their flagship products, Chrome, Gmail, and YouTube among them, which are quickly absorbing functionality built on Gemini. Then in mid-March, OpenAI released GPT-4, a language model well beyond LaMDA in tasks that required critical analysis and coding. Conversely, in February 2024, Google’s image generator was released as a main feature of the new Gemini app.
Kent Walker, Google's top lawyer, has been an advocate for hastening Bard's launch. This very defensive move of Google’s goes hand in hand with the company’s current shift in strategy to accelerate its own efforts on the AI front. A year after that event was described as a “code-red” moment for Google’s AI strategy, the company is back on track. Its prospects now look a whole lot brighter.
Looking back on past struggles and ahead to future prospects, Demis Hassabis remarked on the breadth of Google’s research talents.
"We have the broadest and deepest research base, I would say, of any organization by a long way." – Demis Hassabis
This attitude highlights how Google could use its almost limitless resources in making its way through the crowded AI market.
Leave a Reply