Summary of Key Developments
Recently, Google has experienced two significant changes within its AI teams: First, Demis Hassabis, the founder of DeepMind, has undergone a shift in role from CEO to Chairman and Chief Scientist of Alphabet, effectively losing his day-to-day management authority. Second, Jeff Dean, a veteran employee of Google, left the company with his core team to start their own business (Discovery Loop), with Google providing investment and cloud services. These developments, combined with previous talent losses (such as the departures of Shazeer and Jumper), delays in the Gemini model, and a surge in capital expenditures that led to a negative free cash flow for the first time, have caused a strong market reaction—Alphabet's stock price plummeted by over $175 billion in a single day. This is the fourth time in six weeks that Alphabet's stock has declined due to AI-related setbacks. These events highlight three major challenges facing Google's AI efforts: the loss of key talent, internal management integration issues, and the contradiction between the organizational inertia of large companies and the need for radical innovation.
Key Personnel Changes: Promotions and Respectful Departures
1. Hassabis' Role Shift
Hassabis, who was the driving force behind DeepMind's breakthroughs like AlphaGo and AlphaFold, now holds a higher title (Chairman and Chief Scientist), but he has lost control of the day-to-day operations of DeepMind. The development and delivery of Gemini have been handed over to Koray, the former CTO, who reports directly to Sundar Pichai. In essence, Hassabis has moved from being the person in charge of the project to an advisor, retaining strategic oversight of the AGI (Artificial General Intelligence) efforts but no longer managing the implementation.
2. Jeff Dean's Respectful Departure
Jeff Dean, Google's 30th employee and a key contributor to core technologies such as search, TensorFlow, and TPU, decided to leave the company to start his own business with three of his colleagues. Instead of trying to retain him, Google chose to invest in their venture and provide them with cloud services and computing resources. This approach benefits both parties: the startup gets the necessary resources, and Google maintains opportunities for collaboration and potential future profits while preventing talent from going to competitors. The question remains—why would such experienced individuals prefer to leave a large company? They argue that the inertia within large organizations makes it difficult to drive bold innovations.
Stock Price Plunge: Market Doubts About Google's AI Capabilities
The stock price has declined four times in six weeks, each time due to AI-related issues:
- June: The departures of two top researchers (Shazeer and Jumper) caused a $200 billion loss in the company's value.
- Mid-July: The delay of Gemini 3.5 Pro (due to insufficient code capabilities) led to a 3% drop in the stock price.
- End of July: Capital expenditures doubled to $44.9 billion, resulting in a negative free cash flow of $5.9 billion, and the stock price fell by 7%.
- August 5: The announcements about Hassabis' role change and Dean's departure caused another nearly 4% drop in the stock price (a loss of $175 billion).
The market's concern is clear: Google has invested heavily in AI (with annual expenditures expected to reach $20 billion) but has not produced a leading model (the delay of Gemini and the lack of cutting-edge models in various areas have raised doubts). Investors question whether Google lacks the ability to effectively utilize its resources.
DeepMind's Transformation: From an Independent Lab to a Corporate Department
DeepMind was once an independent entity acquired by Google, operating with its own research culture (focusing on breakthroughs like AlphaGo before considering commercialization. After merging with Google Brain in 2023, Hassabis led the integration process. However, with his recent demotion and Koray reporting directly to Pichai, DeepMind's independence has been compromised. Google's goal is now to shift from a purely academic focus to more practical product development and commercialization. For example, Koray's background in connecting research with product development means he will prioritize the timely release of Gemini and its integration with Google's search and cloud services for profitability.
The Inertia of Large Companies vs. Startups
The departure of Jeff Dean highlights a common issue in large companies: the larger the organization, the harder it is to innovate boldly. Their vision of automating the entire AI research process (from identifying problems to designing experiments, executing them, and evaluating results) may face significant bureaucratic hurdles within Google, whereas startups can experiment more freely.
This reminds me of Bell Labs, which once relied on AT&T's funding to develop revolutionary technologies like transistors and lasers. However, when its talented employees left, they founded companies like Xerox and Intel. Google is facing a similar situation where its own talents are leaving to start new businesses. While this can lead to innovation diffusion in the industry, it also results in a loss of internal creativity.
Google's Response: A Strategy of "Stop Loss" or "Helpless Situation?"
Google's approach of investing in Dean's team and maintaining a cooperative relationship can be seen as a way to minimize losses without outright conflict. However, this does not address the underlying issue: why do top talents prefer to leave? Google possesses the most comprehensive AI resources (TPU chips, data, distribution channels), but if these cannot be utilized efficiently, they can become a burden. For instance, despite achieving significant milestones (950 million Gemini users and 900 million Gemma downloads), the market is seeking the next generation of leading models. If Google cannot improve its internal management (e.g., by streamlining decision-making processes and giving research teams more autonomy), it may continue to lose top talent.
The Talent War in the AI Era
These changes represent an intensification of the "talent war" in the AI industry. Large companies have resources but lack flexibility, while startups are agile but struggle with access to resources. Google's approach of retaining some talents while letting others go may help, but whether it can retain its core staff depends on its ability to balance commercial goals with academic freedom. After all, major breakthroughs in AI often come from less profit-driven, more exploratory research—something that large companies find particularly challenging to maintain.
For us as individuals, this means that innovation in the AI field may increasingly come from startups like Discovery Loop, in addition to giants like Google and OpenAI. The balance between these two types of organizations will determine the future direction of AI development.