The Baker lab released RFdiffusion3 as open‑source software, expanding de novo protein design to create binders for DNA, small molecules and catalytic enzymes. RFdiffusion3 employs an all‑atom design approach with ten‑fold speed improvements over its predecessor and demonstrated experimental proof‑of‑concept for DNA‑binding proteins and catalysts. The model’s authors framed the release as a community resource to accelerate therapeutic and industrial protein engineering. RFdiffusion3’s atom‑level controls broaden design possibilities beyond prior residue‑level methods and could speed creation of synthetic transcription factors, targeted binders and enzyme catalysts. Open access to a high‑performance design engine is likely to spur startups and academic groups to translate computational outputs into lead molecules, but it also raises questions about downstream validation, biosafety controls and commercialization paths.