Keep pulling the thread on Demis Hassabis.
DeepMind was founded on the core insight that significant AI progress could be made by combining deep learning and reinforcement learning, which were previously siloed academic fields.
The original mission statement for DeepMind was a two-step plan: first, to solve intelligence by building AGI, and second, to use AGI to solve all other problems.
The victory of AlphaGo was the key milestone that demonstrated to DeepMind that its AI algorithms were general and powerful enough to be applied to major scientific problems.
DeepMind's AlphaFold system solved the 50-year-old grand challenge in biology of protein folding.
Isomorphic Labs is focused on building AI technologies to automatically design chemical compounds that bind to specific protein targets.
Demis Hassabis predicts that AI will reduce drug discovery times from an average of 10 years down to months or even weeks within the next few years.
Demis Hassabis predicts that AI-driven drug discovery will enable the creation of personalized medicines.
Google DeepMind's 'Weather Next' is the world's most accurate weather simulator and is far faster than traditional models used by meteorologists.
Google DeepMind is developing a 'virtual cell,' an AI-based simulation of a highly dynamical and emergent biological system.
The success of AlphaFold suggests that classical computing systems can effectively model complex quantum systems like protein folding, potentially reducing the need for quantum computers for such tasks.
Demis Hassabis believes the proper strategy for developing AGI is to first build it as an intelligent tool, and only then address more profound questions like agency and consciousness.
Demis Hassabis predicts AGI will be achieved around the year 2030.