A model’s capability to tackle a task it’s never been explicitly trained on—no labeled examples, no step-by-step demos. It draws on its vast pretraining—scientific papers, codebases, books, internet-scale knowledge—and generalizes instantly to new problems.
OR
Not a chance. Even Schrödinger wouldn’t hedge that bet.
We’re a team that includes people who worked on research and engineering at OpenAI from the launch of GPT-3 through GPT-4 and the ChatGPT era.
We started Zero Shot Fund because we realized that knowledge was being generated at a faster pace than our ability to take action on it. A potential medical treatment, revolutionary new material or breakthrough robotics capability still needs technology and platforms to help deploy these discoveries into the world.
We’re optimistic about AI and as we head towards an agentic and post-AGI world we’re excited to work with founders creating the companies we’ll need to make the jump from knowledge to action.