In this podcast, Peter Gostev, the Head of AI at Moonpig, shares his insights on using large language models (LLMs) in e-commerce and banking. He stresses the value of hands-on experimentation to build a genuine understanding of AI, rather than just relying on theoretical strategies. While he recognizes the hurdles of deploying AI in large, regulated companies like NatWest—such as slow implementation and coordination challenges—he points out the agility and speed with which smaller, impactful projects can be executed at Moonpig. Gostev also notes that the current excitement around AI varies widely based on expectations, yet he believes the true potential of LLMs, especially with innovations like "test-time compute," is still largely overlooked and underutilized.
Sign in to continue reading, translating and more.
Continue