In the video, Chris Gibson, co-founder and CEO of Recursion, discusses how the company is using advanced AI and machine learning to address the high failure rate of new pharmaceutical drugs, which is around 90% during clinical trials. By building a robust biological dataset through extensive experimentation and leveraging AI technologies, Recursion aims to accelerate drug discovery and transform the pharmaceutical industry.
In the video, Chris Gibson, co-founder and CEO of Recursion, discusses the significant challenge in the pharmaceutical industry, where approximately 90% of drugs fail during clinical trials. This high failure rate highlights the complexities of biology and chemistry, prompting Recursion to leverage advanced tools like machine learning and artificial intelligence (AI) to improve drug discovery outcomes. The company was founded in 2013 with a focus on integrating AI early in its development process, utilizing basic machine learning classifiers derived from academic research.
By 2015 and 2016, Recursion began experimenting with more sophisticated AI techniques, including convolutional neural networks. Today, the company operates the fastest supercomputer in the biopharma sector, which is used to train extensive AI models to tackle biological challenges. Gibson emphasizes that AI’s effectiveness is heavily reliant on the quality and quantity of data available for training, which led Recursion to invest significant effort in building a robust biological dataset.
To create this dataset, Recursion developed robotics capable of conducting millions of experiments weekly, utilizing real human cells, chemicals, and biological perturbations. They employed techniques like CRISPR-Cas9 to knock out every gene in the human genome across various cell types and profiled millions of chemical compounds. This extensive data collection has enabled the company to store and organize information in a manner that is accessible for machine learning applications.
When pursuing new diseases, Recursion’s scientists can now utilize large language models trained on their dataset to conduct searches for potential drug opportunities. This process accelerates the identification of viable projects, significantly reducing the time required for drug discovery. Gibson believes this represents a fundamental shift in how drug discovery will be approached in the future, driven by the capabilities of AI and the comprehensive biological data they have amassed.
Gibson concludes by urging other organizations to adopt AI solutions to address their operational challenges, emphasizing that the transition is not as daunting as it may appear. He encourages businesses to seek out talent that can identify innovative use cases for AI, as this technological revolution will transform work processes in profound ways. By staying at the forefront of AI implementation, companies can enhance their efficiency and contribute positively to humanity’s advancement.