In 2024, Google DeepMind scientists shared the Nobel Prize in Chemistry for a neural community, AlphaFold, which predicts the buildings of proteins. It confirmed that AI might make groundbreaking scientific discoveries, however AlphaFold might not be the most effective template for accelerating science. As a substitute, one other method could maintain the important thing: AI brokers.
AlphaFold relied on a dataset of roughly 170,000 experimentally validated protein buildings that took 53 years and roughly $21 billion value of experimental work to assemble. Comparable datasets will likely be tough or unattainable to create in lots of fields.
AI brokers can as an alternative mannequin the iterative, extremely contingent means of precise analysis. Whereas instruments like AlphaFold apply a robust method to a restricted query, brokers are inherently generalists. They don’t characterize a brand new approach to do science—as an alternative, they digitally mannequin the human means of discovery.
Learn the complete op-ed on how brokers might speed up scientific discovery.
Contained in the “censorship-industrial advanced” concept shaping US coverage
For years, the concept of a “censorship-industrial advanced” that suppressed conservative and populist speech unfold in right-wing circles on-line. However now, the idea has made its means into the Trump administration.
Over the previous 9 months, MIT Expertise Evaluation investigated its origins and traced its rise. In a digital Roundtables session on Thursday, August 13, senior reporter Eileen Guo and government editor Amy Nordrum will discover what they found, the place the idea goes, and what it might imply for the way forward for democracy and the web.
Register right here to affix the session at 19:00 GMT / 2:00 pm ET / 11:00 am PT.
The must-reads
