DeepMind Veteran Raising $1B for Non-LLM AI

- David Silver was the lead researcher for AlphaGo, the program that defeated world champion Lee Sedol at Go in 2016, and later led the AlphaZero project, which taught itself to master Go, chess, and shogi without any human game data. - The $1 billion seed round, if it closes, would be the largest in European history, with Sequoia Capital reportedly leading and Nvidia, Google, and Microsoft in talks to participate. - Reinforcement learning (RL) trains an agent to make decisions by interacting with an environment and learning from rewards and punishments, a contrast to LLMs which are trained on static datasets to predict the next word. - An AI architecture based on reinforcement learning instead of transformers would have fundamentally different computing requirements, potentially creating an opportunity for new custom silicon and ASIC designs beyond the current generation of LLM-focused accelerators. - Hyperscalers are increasingly designing their own custom chips for specific AI workloads, such as Google's TPU, Amazon's Trainium, and Microsoft's Maia, to optimize performance and reduce costs. - Silver’s go-to-market strategy will likely differ from typical enterprise software, focusing first on achieving research breakthroughs to build a defensible moat before establishing a commercial model, a common path for deep-tech companies founded by technical experts. - This funding round is part of a broader "AI Mega-Round" trend where venture capitalists are making pre-product seed investments of over $100 million based on the strength and track record of the founding team. - The stated mission of Ineffable Intelligence is to build an "endlessly learning superintelligence that self-discovers the foundations of all knowledge," a goal that Silver's previous work at DeepMind has consistently pursued.