Until recently, Google was the place AI talent flocked to. However, recent weeks show that times are changing. Over the past month, four senior researchers involved in developing the Gemini model and other projects have left the company. What makes this even more interesting is that they are heading to rival Anthropic. This shows that the balance of power in the artificial intelligence industry is beginning to shift—and very significantly.
Who has left Google
According to a Bloomberg report, Jonas Adler and Alexander Pritzel, two developers and researchers who were key figures at Google and Gemini, have decided to join Anthropic.
Adler was involved in developing AI tools for programming, while Pritzel worked on model pre-training—the stage in which a model learns from vast amounts of data and acquires basic “knowledge” about the world. The whole matter came to light at the end of June, with the information provided to the agency by an anonymous source.
However, this is not an isolated case, as these two researchers are not the first to defect. Just a few days earlier, Noam Shazeer, vice president of engineering and co-author of the landmark 2017 paper Attention Is All You Need, which laid the foundations for the technology underpinning virtually the entire field of generative AI today, announced his move. However, it is important to note that unlike the previous employees, he is not joining Anthropic, but OpenAI.
Another person leaving Google is Nobel Prize winner John Jumper, who also announced that he was moving to Anthropic. His departure had the most tangible impact of all, as he is behind the AlphaFold project, which is part of Google DeepMind. On the day he left Google, Alphabet shares fell by 6%, equivalent to nearly $245 billion in a single trading day.
Why now, and why Anthropic?
Four departures within a single week suggest that this is no coincidence, but an entirely new trend. Anthropic wants to take its AI beyond ordinary chatbots and into specialized fields such as programming, healthcare, science, and research. And this is precisely where the most highly qualified people are needed and where their expertise is invaluable.
However, money is not the only variable in the equation. Another important factor is the planned stock market listings of Anthropic and OpenAI, giving the companies added value that Google can hardly offer. The opportunity to join a company planning an IPO and thereby gain a stake in the future growth of its share price may be one of the reasons for changing employers.
Nor should we overlook access to computing power. Within Google, access to the company’s chips has become something employees have to wait for. External demand is limiting capacity for internal research, making it more difficult to secure hardware for the company’s own projects. Ironically, one of the external customers purchasing capacity on Google’s TPU chips is Anthropic itself, which is now winning over the brightest minds in AI.
What does this mean for Google and the AI industry as a whole?
The departure of the most highly qualified experts and researchers is primarily a blow to Google’s reputation. Gemini will remain a competitive AI technology, continue to rank among the leaders, and the loss of a few people will not bring it down. However, given the rapid succession of departures—and the fact that they benefit its competitors—the question is how challenging the subsequent development of AI models will be.
Historically, Google has been adept not only at filling vacancies, but also at luring talented employees back. More important, however, will be determining whether this is merely a coincidence and a handful of unrelated departures, or evidence that the far more interesting AI game is already being played on an entirely different field.
Source: TechCrunch, The Next Web, Business Insider
