Tim DraperFounder and Managing Partne, Draper Associate · I back the ideas that look crazy first

If there are 20,000 ways to catch cancer… why have we only approved 1,000 drugs? 🤯

19 Feb 2026 · 1 min

Watch on YouTube

0:59

In this brief dialogue, Speaker A highlights the staggering diversity of cancer—estimated at 20,000 distinct pathways—contrasted with the modest tally of roughly 1,000 approved drug molecules, underscoring a massive treatment gap. Speaker B raises a critical question about how artificial intelligence can determine safety, prompting Speaker A to discuss the challenges of ensuring that AI‑designed compounds do not produce off‑target effects or harmful drug‑drug interactions. The conversation then shifts to the founding motivation of their company: to make AI‑driven drug‑discovery technology usable by both technical and non‑technical chemists. Consequently, they are investing heavily in an intuitive interface that democratizes access to sophisticated design tools while maintaining rigorous safety considerations.

Each and every one of us here in this room will have 20,000 ways in which you can catch cancer that all need to be treated by different treatments. However, over the last century, humanity has only approved around 1,000 drug molecules for treatment, and we need to do much more.
I guess you need the human there too, but— Yes, very important. But the— how does the AI know if something is safe?
You can always solve for one thing, such as kills bugs really effectively, but while you're doing that, you might lose out on other things and it doesn't treat anything. You also need to make sure it doesn't hit any other targets. It might kill cancer, but it also might lead to drugs starting to interact with each other. There's all kinds of off-effects that might happen. And that was also where the kind of the idea for the whole company came from, which was all about making technology accessible to people that actually make the decisions in the lab. So we are putting a lot of effort into the design of the interface to make sure that technical chemists, non-technical chemists alike can use the interface.