With $300 million in seed capital secured, a team of former researchers from OpenAI and DeepMind is starting a company dedicated to automating scientific research. Periodic Labs hopes to establish AI laboratories that can autonomously design and execute experiments without immediate human supervision.
Andreessen Horowitz and Accel led the round, which also attracted interest from some well-known names in tech, including Nvidia, Elad Gil, Jeff Dean, Eric Schmidt, and, of course, Jeff Bezos Google Brain
Cubuk had AI systems discover more than 2 million new crystals. Before that, former OpenAI Research VP William Fedus
Machine-learning algorithms are expected to speed up the experimental process tremendously and churn out massive datasets that can be used to train further artificial intelligence applications when tied to automated lab platforms. According to the experts, the project stands as a metaphor for the wider trend of AI moving from pure algorithmic functioning to physical experiment. Periodic Labs
OpenAI Insiders: OpenAI Taps Thinking Machines as First APAC Services Partner
The company plans to change the way research is done with its introduction of automation interfaced with state-of-the-art machine learning, and thus herald an era where it’s the artificial intelligence itself, not just another tool, doing the discovery. The scientific world will be keeping a close watch on all the attainments emerging from this ambitious approach, as the money begins to flow in and preliminary AI-driven studies start.
