According to exclusive news from the media Semafor, Google will appoint DeepMind Chief Technology Officer Koray Kavukcuoglu as Google's chief AI architect, responsible for coordinating future AI product development and reporting directly to CEO Sundar Pichai.

It is reported that Kavukcuoglu was a disciple of Meta Chief Scientist Yann LeCun when he was studying for a doctorate in computer science, and later joined DeepMind and joined Google with the acquisition and merger. Pichai's memo said that Kavukcuoglu will "accelerate the integration of world-leading models and products to achieve more seamless integration, faster iteration and higher efficiency", which shows that Google has high hopes for this appointment.

Meta poached Jack Rae, chief researcher at Google DeepMind, and Google promoted Yann LeCun's disciple, DeepMind's chief technology officer. This wave of operations is simply dramatic.

▲ Koray Kavukcuoglu

1. Kavukcuoglu was appointed to promote Google's AI product development

Kavukcuoglu was an aerospace engineer and received a doctorate in computer science. He studied under Yann LeCun, an AI pioneer, professor at New York University, and current chief scientist of Meta AI. He joined DeepMind as a researcher in 2012 and was promoted to DeepMind VP in 2019. During his tenure, he led many major AI breakthroughs. Since 2020, Kavukcuoglu has won the AI ​​2000 Most Influential Scholar Award and the Machine Learning Honor Award for five consecutive years.

▲ Yann LeCun

Under the close collaboration and leadership of Kavukcuoglu and DeepMind co-founder Demis Hassabis, Google has come from behind in the field of AI and launched the model Gemini to take the lead in key AI benchmarks. But Google is now under new pressure to turn the raw power of the Gemini series of models into successful products.

In recent months, Kavukcuoglu has worked with Google's product leaders to explore how to use AI models to drive Google's new product development. This appointment formalizes the work, and Kavukcuoglu will move from London to Mountain View in Silicon Valley.

2. Product launch lags, Google personnel changes are inevitable

Since ChatGPT disrupted the technology industry, Google has undergone a comprehensive transformation. The company merged DeepMind, which it acquired in 2010, with Google Brain and appointed Hassabis to lead the AI ​​business.

To accelerate product development, Google has promoted the streamlining of team work processes across the company, including buyout plans, layoffs and leadership adjustments. "We are entering a new phase of AI platform transformation, which requires the company to shift gears and speed up to ensure that product evolution keeps pace with model development," Pichai wrote in the memo.

In April, Google transferred Josh Woodward, head of Google Labs, to the head of the Gemini project. Woodward led the creation of the popular AI product NotebookLM, which can convert text into podcast-style programs. Google hopes to replicate more similar success stories, but the current innovation speed of AI models has far exceeded the progress of product development.

To some extent, this is an inevitable result, because new products must be based on new capabilities. To this end, Kavukcuoglu was appointed Chief AI Architect, who will continue to inform the entire company of the direction of technology development and help product teams make forward-looking plans.

Conclusion: Can the new appointment help Google AI win the "speed race" of AI products?

In the past, the technology industry relied on predictable incremental changes, but AI has not yet formed a stable development trajectory similar to Moore's Law. Its breakthroughs are explosive and progress is amazing. The launch of AI products has become a "speed race." If Google fails to create new AI applications, competitors may seize the lead. The rise of ChatGPT is a typical example: it takes advantage of AI architecture breakthroughs pioneered by Google researchers, who were encouraged to make their research public at the time.

Today, AI is changing the way users search the Internet, and some early adopters have used ChatGPT or Perplexity instead of traditional search engines. On Tuesday, The Verge reported that Google is promoting a buyout plan in its core revenue-generating department, "Search Business".

Gemini's new features, such as "Project Mariner" and "Astra", have shown the potential of new products. It is the mission of Pichai's new direct subordinates to ensure that these technologies ultimately empower Google products.