Google doesn't need the LLM crown

Reed asserts that Google's state-of-the-art AI models are lagging behind by six months in coding capability. Multiple key figures, including DeepMind CEO Demis Hassabis and Chief Scientist Jeff Dean, have recently departed from the company. Despite these departures, Google's position in the AI race is not as dire as one might think.

The focus of the AI competition has shifted from being the top-ranked entity to utilizing AI for practical tasks. Users can currently interact with Google's Gemini through natural language searching in their emails and YouTube, as well as receive AI feedback on written content within Google Docs. Google's vision for Gemini includes expanding its capabilities to incorporate more tasks, integrating various Google products and external services, and leveraging data from billions of users.

The Gemini Spark serves as an initial step towards this vision. As AI infrastructure continues to develop and grow, Google's extensive product portfolio—ranging from mobile operating systems to browsers, email, maps, and beyond—will become more advanced, seamlessly interacting with users' daily lives, including smart appliances and vehicles, and potentially robotics.

However, Google will not employ the top models for these features due to the prohibitive cost and insufficient compute power. While there may be a desire to construct extensive product frameworks, individuals like Hassabis and Dean are opting to tackle other challenges more intriguingly. It is understandable that those who have made significant contributions and benefited financially would prefer to pursue new endeavors.

Sergey Brin, Google's co-founder without an official role, now wields considerable influence through his supervoting shares and status, according to multiple sources within the company. The most critical asset Google needs is smart and technically skilled personnel to address its current model development, architecture, and efficiency challenges.

The company has historically been able to attract and retain numerous individuals, even losing some prominent talent recently.