Latent Space interviewed Radical AI’s Joseph Krause about self-driving labs for materials discovery. The discussion centers on why the durable advantage in the field may come from lab execution rather than model access alone.
That matters because AI can propose experiments, but materials breakthroughs still require physical testing, instrumentation, and reliable feedback loops. The lab becomes part of the intelligence system.
The conversation points to a broader theme in applied AI: the strongest products may combine models with proprietary workflows, equipment, and data generated from real-world experimentation.