For Episode 38 of the BioRevolution podcast, Andreas Horchler and Louise von Stechow discuss the benefits and challenges of AI co-scientists, which can condense the work of weeks, months, or even years of reading, debate, and hour-long lab meetings into minutes.
Researchers at Stanford University and Google DeepMind have released multi-agent AI “scientist” systems. But how good are they, really, and will they overcome current limitations of large language models like OpenAI’s GPT or Google’s Gemini—such as the lack of robust reasoning capabilities, which are crucial for scientific discovery?
