San Francisco and London— Sundar Pichai, the CEO of Google, recently used X (formerly Twitter) to talk about a big step forward in using artificial intelligence (AI) in basic science. He said that a DeepMind model based on Gemma helped find a new possible cancer treatment pathway in partnership with Yale University.
Pichai said that the accomplishment was a “exciting milestone for AI in science.” He said that the company’s C2S-Scale 27B foundation model “generated a novel hypothesis about cancer cellular behavior, which scientists experimentally validated in living cells.”
The CEO was hopeful about the discovery’s future, saying, “With more preclinical and clinical tests, this discovery may show a promising new way to make therapies to fight cancer.”
The Science That Led to the Discovery
The Cell2Sentence-Scale 27B (C2S-Scale) foundation model, created by DeepMind researchers in collaboration with Yale University, is the AI model that makes this finding possible. This model is made just for looking at single cells.
DeepMind verified in a related blog post that the C2S-Scale model’s ability to generate new ideas allowed it to come up with an entirely new theory about how cancer cells act. The model’s predictions were not just theories; they were instantly backed up by “experimental validation in living cells,” which confirmed the results and may have sped up the search for novel medicines.
Social Media Cheers AI’s effect on society
Pichai’s article, which immediately got millions of views, got a lot of people excited, especially in the tech and science communities, who praised the shift from entertainment to real-world advantages.
A former Google engineer said that the change was very important: “This is exactly where AI will have the biggest impact on society: speeding up breakthroughs in basic science like oncology.” It’s really inspiring to go beyond entertainment and directly help the social good, as shown by the C2S-Scale model and the Yale collaboration.
Another user made it clear how big the model is in its field by saying, “To clarify: it’s a huge step in AI-for-science; this model is for cell and molecular biology instead of the regular AI we know.” In science, 27B params is a lot.
Most people were in support of this use of AI, with one post saying, “Hell yes!” I know it sounds strange, but combating cancer is much more vital than the junk that the other frontier laboratories are making.
The announcement also struck a chord with many who had the sickness themselves. One comment said, “This is so special.” I think there should be more AI resources put into making medicines faster because my father is fighting cancer. Thanks, Google.
This partnership shows how important advanced generative AI models are becoming in scientific research. They are moving from creative or general tasks to solving hard problems in areas like biology and medicine.