Google’s AI Masterstroke: How $2.7 Billion Lured Back Noam Shazeer
Noam Shazeer is a name that resonates deeply in the world of AI. Known for his pioneering work at Google, he has been instrumental in shaping the technologies that power today’s most advanced AI systems. In 2024, Shazeer made headlines when Google reportedly spent $2.7 billion to rehire him, following his successful venture with Character.AI. But to understand why Google made such a substantial investment, we need to look back at his achievements, frustrations, and eventual departure from the tech giant.
Early Contributions: Transforming AI at Google
Shazeer’s career at Google began in 2000, where he quickly rose through the ranks as one of the company’s most brilliant engineers. One of his key early accomplishments was the development of Meena, a conversational AI system designed to surpass the performance of existing chatbots. However, Shazeer’s most significant contribution came in 2017 with his role in the "Attention Is All You Need" project.
Tekedia Mini-MBA edition 15 (Sept 9 – Dec 7, 2024) has started registrations; register today for early bird discounts.
Tekedia AI in Business Masterclass opens registrations here.
Join Tekedia Capital Syndicate and invest in Africa’s finest startups here.
"Attention Is All You Need" is a landmark 2017 research paper in machine learning authored by eight scientists at Google, including Shazeer. The paper introduced the Transformer architecture, a novel deep learning model based on the attention mechanism. It marked a paradigm shift in how machines process and generate human language. The Transformer architecture has since become the foundation of large language models like GPT, BERT, and others that power modern AI systems.
This architecture has revolutionized AI by enabling models to handle complex tasks such as machine translation, question answering, and what is now recognized as multimodal Generative AI. The other authors of the paper include Ashish Vaswani, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan Gomez, Łukasz Kaiser, and Illia Polosukhin. By 2023, all the authors had left Google to establish their own AI startups, except Łukasz Kaiser, who joined OpenAI.
The success of the Transformer underscored Shazeer’s exceptional ability to solve complex problems and push the boundaries of AI. It remains the dominant architecture for natural language processing (NLP) and generative AI tasks today.
Some Transformer Models:
| Model | Architecture | Product |
| GPT-4 (OpenAI) | Transformer (Autoregressive) | ChatGPT, Microsoft Copilot |
| Claude 2 (Anthropic) | Transformer (Focused on AI safety) | Claude AI for enterprise use |
| Gemini 1 (Google DeepMind) | Transformer (Hybrid) | Google Bard, Google Cloud AI |
| PaLM 2 (Google) | Transformer | Bard, Vertex AI (Google Cloud) |
| LLaMA 2 (Meta) | Transformer | Research applications, Meta’s AI tools |
| T5 (Google) | Text-to-Text Transformer | Google NLP tasks, research models |
| BERT (Google) | Transformer (Bidirectional) | Google Search, NLP applications |
| RoBERTa (Meta) | Transformer (Optimized BERT) | Research, NLP benchmarking tools |
| Turing-NLG (Microsoft) | Transformer | Microsoft Azure AI, Natural Language Generation |
| Megatron-LM (NVIDIA) | Transformer (Large-Scale) | AI research, NLP and text generation tasks |
| OPT (Meta) | Transformer | Research, Meta’s AI tools |
Frustration and Departure: Leaving Google to Found Character.AI
Despite his numerous successes, Shazeer grew increasingly frustrated with Google’s corporate structure. The company’s reluctance to release his chatbot Meena, due to concerns over safety and fairness, stifled his vision of what AI could achieve. This clash with Google's leadership led to Shazeer leaving in 2021 to co-found Character.AI, alongside his colleague Daniel De Freitas
At Character.AI, Shazeer found the creative freedom he had been seeking. His new company focused on developing personalized AI systems that could engage in deep, meaningful conversations with users. Within a short time, Character.AI became a major player in the AI space, achieving a $1 billion valuation and gaining widespread attention.
Google’s $2.7 Billion Move: Why Rehire Shazeer?
By 2024, the AI landscape had shifted dramatically, with companies like OpenAI and Microsoft challenging Google’s dominance. Recognizing the need for fresh leadership and groundbreaking innovation, Google reportedly spent $2.7 billion to bring Shazeer back into its fold.
Just last month, Google and Character.AI revealed that Shazeer, De Freitas, and certain members of Character.AI’s research team would be joining Google’s AI unit, DeepMind.
The payment wasn’t just financial; the deal included a licensing agreement for Character.AI’s intellectual property, allowing Google to leverage the company’s advanced technology while reintegrating Shazeer into its AI division. This licensing deal granted Google instant access to Character.AI’s technology, bypassing regulatory approval. According to reports, Google's employees see Shazeer’s return as the key motivator behind the acquisition of Character.AI, reinforcing his pivotal role in shaping the company’s future in AI.
Shazeer is now positioned to lead Gemini, Google’s ambitious project aimed at surpassing competitors like OpenAI’s GPT models. His return is seen as pivotal for Google, especially given his unmatched expertise in LLMs and the Transformer architecture that continues to power much of modern AI.
Former Google CEO Eric Schmidt was reportedly so impressed with Shazeer that it paved the way for Shazeer to develop an AI model capable of human-level intelligence. "If there’s anyone in the world who can achieve it, it’s him," Schmidt said of Shazeer in 2015. Two years later, in 2017, Shazeer went on to create Meena, a chatbot capable of engaging humans across a variety of topics
The Future of AI with Shazeer Back at Google
Shazeer’s return signals a renewed commitment from Google to lead the next wave of AI innovation and competition. With his expertise, the company hopes to regain its competitive edge and push the boundaries of what AI can achieve. His experience at Character.AI, combined with his intimate knowledge of Google’s AI infrastructure, makes him uniquely qualified to guide the company’s future efforts.
In conclusion, Noam Shazeer’s journey—from creating the Transformer, to leaving Google in frustration, to leading a billion-dollar AI startup, and finally being rehired to head Google’s AI revolution—illustrates his unparalleled vision and technical brilliance. As he takes charge of Gemini, the world watches to see how this AI genius will again shape the future of technology.
Uploaded files: