Analysis frame
Peer-reviewed research
Laboratory characterization tests a different claim from a digital prediction; each result needs the denominator that belongs to its own stage.
- Experimental materials scientists
- Institutions funding scientific AI systems
- Will the workflow generalize beyond the validated chemistry?
- How does end-to-end performance compare with a matched alternative?
- Funding can shift toward laboratories able to validate selected predictions.
- Reporting attempts and unsuccessful syntheses could sharpen comparisons across discovery systems.
What a fair comparison needs
Track attempted syntheses as well as successes, with comparable time, resources and human effort. A computational candidate count cannot supply the missing denominator.
Go to the source
Read the evidence behind this analysis. External links open in a new tab.
published MOFGen paper biosecurity Perspective university account earlier preprint


