The speaker reflects on their experience with a PhD in computer science, contrasting it with a friend's experimental protein research. They highlight the revolutionary impact of AI, particularly DeepMind's AlphaFold, which can predict protein structures with high accuracy, saving significant research time and recently winning a Nobel Prize. This AI advancement has transformed protein research, traditionally a lengthy experimental process, into a computational task completed in seconds. The speaker argues that AI will continue to drive breakthroughs across various scientific fields due to its ability to process diverse data modalities and leverage vast computational power. They introduce the concept of AI analogs, virtual models that replicate real-world systems, allowing for scalable experimentation and new knowledge creation. The speaker focuses on drug design, a field facing increasing challenges and costs, and explains how AI can reverse these trends by modeling complex biological interactions. They describe the development of AlphaFold 3, which models biomolecular interactions with unprecedented accuracy, enabling rapid drug design. The speaker envisions a future where AI-driven agents design personalized drugs for individual patients, revolutionizing treatment for diseases like cancer. They conclude by emphasizing the potential of AI analogs to drive scientific and technological advancements across various domains, calling for collaboration from experts in AI and machine learning.