Keep pulling the thread on Demis Hassabis.
Demis Hassabis estimates a 50% probability of achieving Artificial General Intelligence (AGI) by the year 2030.
Demis Hassabis conjectures that any pattern found in nature can be efficiently discovered and modeled by a classical learning algorithm.
Google DeepMind's Weather Next system is better at weather prediction than traditional fluid dynamics-based systems that run on large supercomputers.
Demis Hassabis believes AI systems will materially help discover new solar panel materials, room-temperature superconductors, and optimal batteries within the next five years.
Demis Hassabis estimates the AI revolution will have an impact 10 times greater than the industrial revolution and will occur 10 times faster, compressing a century of change into a decade.
Demis Hassabis believes the probability of human extinction due to AI (P-Doom) is non-zero and non-negligible, but the uncertainty is too high to assign a specific number.
AlphaFold 3 is the first step towards modeling dynamic biological interactions, including those between proteins, RNA, and DNA.
Demis Hassabis's long-term research goal is the 'Virtual Cell' project, which aims to create a complete, dynamic simulation of a biological cell, likely starting with a yeast cell.
Demis Hassabis predicts that within the next 5 to 10 years, AI systems will enable the creation of open-world video games that dynamically generate personalized content and narratives.
Google DeepMind is collaborating with Commonwealth Fusion to apply AI to solve plasma containment challenges in fusion reactors.
Demis Hassabis believes that passive observation is sufficient for an AI to learn intuitive physics, challenging the neuroscience theory that embodied action is required for deep perception.
Google DeepMind's Alpha Evolve system uses Large Language Models to propose potential solutions and then applies evolutionary computing to search for novel algorithms.