I don't know what I'm wrong about. If I did, I'd change my mind. So none of these should be considered definitive. If you have an idea that you think I could help, please reach out.
Continual learning systems
Reinforcement learning has become ridiculously better in the last 5 years. What new applications are enabled by this? I think they must have the properties:
- A static solution isn't enough, it must need to learn over time
- Existing solutions aren't enough; you must explore the space to become good.
- Humans can do it, but barely.
Any system that meets these properties is ripe for disruption, and I'd like to invest.
Laboratory science productivity
An investment thesis in tech posits that the number of developers is small, and the number of skilled developers is even smaller, so any technology that can make existing developers more productive is ridiculously valuable. This is true for many other professions. The laboratory sciences, in particular, are hard to recruit for, requiring many years of advanced education to do. If you're making tools that make it easier to run a medical lab, let me know.