2024 Big Ideas: Miracle Drugs, Programmable Medicine, and AI Interpretability
Dec 8, 2023
auto_awesome
Julie Yoo, a general partner at Andreessen Horowitz with expertise in miracle drugs, discusses the transformative potential of GLP-1 therapies for obesity and diabetes, emphasizing access and affordability challenges. Jorge Conde dives into programmable medicine, highlighting innovations in gene and cell therapies aimed at curing diseases. Anjney Midha focuses on AI interpretability, advocating for clarity in AI decision-making processes, especially in critical fields like healthcare. Together, they explore how these big ideas could revolutionize medicine and technology in 2024.
Miracle drugs need innovators to address the intersection of policy, bioinformatics, manufacturing, financing, and clinical operations to make them accessible and affordable without bankrupting the system.
Programmable medicines, like gene therapy, have the potential to revolutionize drug development by reusing components and lowering costs, with the FDA adopting new processes to evaluate and approve these therapies.
The shift towards interpretability in AI models, through analyzing consistent patterns of activation across multiple neurons, opens the door to better understanding and controlling AI models for increased transparency and reliability in their applications.
Deep dives
Democratizing Miracle Drugs
In 2023, the impact of therapies like GLP-1s and cell and gene therapies was profound, but the current healthcare insurance system is not equipped to handle the cost or accurately assess the value of these therapies. Innovators are needed to address the intersection of policy, bioinformatics, manufacturing, financing, and clinical operations to make these miracle drugs accessible and affordable without bankrupting the system.
Programming Medicine's Final Frontier
Traditional drug development is time-consuming, risky, and expensive. However, programmable medicines, like gene therapy, are poised to revolutionize the field. By reusing components, similar to SpaceX's reusable rockets, we can lower costs and expand the scope of these therapies. The FDA is adopting new processes to evaluate and approve these programmable medicines, following the footsteps of the FAA's adaptive approach to aviation safety. This shift towards programmable medicines will transform the way we develop and distribute treatments.
From Black Box to Clear Box: Understanding AI Models
The topic of AI interpretability is gaining importance as large language models (LLMs) become more prevalent. One breakthrough is the shift from analyzing individual neurons in LLMs to interpret features, which represent consistent patterns of activation across multiple neurons. This approach provides clearer insights into how LLMs process information and make decisions. Although research is still ongoing, this shift opens the door to better understanding and controlling AI models, paving the way for increased transparency and reliability in their applications.
Lengthy drug development process
Developing a new drug typically takes around 10 to 15 years. This lengthy process consists of three stages: drug discovery, preclinical development, and clinical development. During drug discovery, researchers identify a target in a disease and work on finding a medicine to hit that target, which can take several years. Preclinical development involves testing the developed molecule to ensure it has the desired qualities of a medicine. This phase is conducted outside of humans, in dishes, cells, and animal models, and can also take many years. Finally, clinical development consists of conducting human clinical trials, which can last around 5 to 7 years. Regulatory approval can take an additional 1 to 2 years. When these phases are combined, it explains why drug development is a lengthy process.
The potential of programmable medicine
Programmable medicine has the potential to significantly transform drug development. Unlike traditional medicines, programmable medicines have multiple components that can be redeployed for different applications. For instance, in gene therapy, a medicine can deliver a genetic payload to cells with defective genes. This concept of reusing components opens the possibility of delivering different genes to different cell types for various diseases. Programmable medicines allow for greater flexibility in targeting diseases with known mutations, potentially paving the way for a future where cures become more common. Additionally, regulatory bodies like the FDA are exploring ways to adapt their processes to accommodate the development of programmable medicines more efficiently, aiming for both rigor and speed.
Smart energy grids. Voice-first companion apps. Programmable medicines. AI tools for kids. We asked over 40 partners across a16z to preview one big idea
they believe will drive innovation in 2024.
Here in our 3-part series, you’ll hear directly from partners across all our verticals, as we dive even more deeply into these ideas. What’s the why now? Who is already building in these spaces? What opportunities and challenges are on the horizon? And how can you get involved?
View all 40+ big ideas: https://a16z.com/bigideas2024
Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
