Can New Materials Really Be Designed from Thought Alone?

I partly agree with this vision. Computational methods can greatly accelerate materials discovery by predicting promising compositions and properties before conducting experiments, which can save significant time and resources. However, experiments will still be necessary to verify synthesizability, stability, processing behaviour, and real-world performance.

I believe the idea is realistic as a combination of computation and targeted experimentation, rather than computation completely replacing experiments. With continued progress in DFT, materials databases, machine learning, and computing power, this approach may become increasingly practical over the next few decades.