By firing this employee, Google made a huge mistake that cost it $2.7 billion

News JVTech By firing this employee, Google made a huge mistake that cost it $2.7 billion

A breakup and regrets

In 2021, Google made an error which, very recently, cost $2.7 billion. Engineer Noam Shazeer, then one of the pillars of the development of artificial intelligence within the company, left the ship after a disagreement with management. At the time, Shazeer designed Meena, a revolutionary chatbot capable of conducting complex conversations on a wide variety of topics. However, Google refused to launch this innovationfearing security and ethical concerns.

Frustrated by this decision, Shazeer, alongside his colleague Daniel De Freitas, decided to leave Google to create their own startup, Character.AI. This departure marked a turning point, not only for the company, but also for the artificial intelligence industry since the startup established itself as one of the most promising companies in Silicon Valley. With new features and a growing number of users – more than 20 million monthly active users in 2023 – the startup quickly reached a billion-dollar valuation.

Costly reintegration

Aware of the huge missed opportunity with Shazeer's departure, Google had to resign itself to making an agreement in 2024 to bring the AI ​​genius back into its ranks. This deal did not take the form of a simple recruitment, however, as Google had to pay $2.7 billion to license Character.AI's technology while persuading Shazeer and his team to return to the companyparticularly within its AI division, DeepMind.

It was not a classic acquisition: by choosing not to fully acquire Character.AI, Google sought to avoid lengthy regulatory processes that would have delayed the integration of the technology. But the price paid for a simple license and the return of some talent clearly illustrates the extent to which Shazeer's departure was seen as a major strategic error.

Gemini: the redemption project

Today, Noam Shazeer is therefore back at Google, but at a much higher cost than before. Now leading development of the next version of Gemini, Google's artificial intelligence model intended to compete with giants like OpenAI and its ChatGPT, Shazeer is tasked with returning Google to the top of the AI ​​race.

If Gemini is seen as a key element of Google's strategy for the future, its development has been fraught with pitfalls. Earlier this year, Google even suspended an image feature generated by Gemini following embarrassing errors. These setbacks show that the reinstatement of Shazeer and his team is just a first step in a broader battle to make up lost ground in the field of artificial intelligence.