James Zou, who leads Stanford’s AI for Science Lab, built a “virtual biotech” with tens of thousands of AI agents. The agents zeroed in on a protein linked to lung cancer and designed a therapy, dramatically speeding drug development.

In another breakthrough, an AI algorithm can watch cardiac‑ultrasound video clips and diagnose heart‑related diseases in a few seconds, a task that previously required years of manual analysis by clinicians.

Anima Anandkumar, a computer scientist at Caltech and co‑leader of the AI4Science initiative, says AI is ushering a new scientific golden age, comparing its impact to the invention of the telescope and the microscope.

Restrictions on advanced AI models, such as Anthropic’s Claude, forced Finkbeiner’s group to rely on a lower‑level version. He warns that limiting access to powerful models could slow the pace of medical breakthroughs for patients waiting for cures.

If AI can turn years of research into weeks or days, new treatments could reach patients far sooner. However, overly strict regulation might hold back these life‑saving advances.