Interview, Fireside Chat
On Machine Learning in Medicine and More
- Software, machine learning, and compute are expected to drive impacts across biology, creating a new market of "software as a drug" similar to the early biologics revolution.
- Digital therapeutics and programs for behavioral issues like type 2 diabetes, anxiety, and PTSD will scale from tens to tens of millions of users via mobile technology and efficient coaching.
- Clinical testing and early cancer detection using machine learning and genomics are projected to achieve cure and success rates between 80% and 100%, potentially eliminating cancer as a condition within 10 years.
- The cloud biology sector is described as very early, with companies expected to either build internally or provide services, while regulatory agencies offer opportunities to define best practices and build barriers to entry.
- Investment returns in this sector are anticipated to resemble software multiples due to high margins and managed risks, contrasting with the historical 10x returns of traditional life science investing.
- A new class of individual founders capable of bridging biology and computer science is emerging to lead companies with unique cultures, with investors expecting these entities to grow revenues and become acquisition targets.
- Strategic acquisition scenarios include major life science companies like Pfizer buying digital therapeutics firms like Omada, tech giants like Google acquiring similar entities, or diagnostic firms like Freenome being purchased by therapeutics or sequencing companies.
- Success in this space depends on specific conditions, including accurate machine learning tests, inexpensive genomic advances, and the right teams, with risks involving the need for these advancements to materialize.
- Traditional healthcare systems are expected to see more integration of new technologies rather than a complete reset, though challenges remain regarding the infrastructure required for widespread deployment.