2026-10-25 (~10:00 AM ET, MSN syndication — msn.com/en-us): Jack Dorsey's Block says it built Dessa, a full-duplex voice model, trained on human speech alone using a single Nvidia GB200 NVL72 rack, and that it runs on a MacBook Pro M4 Max or better — training and inference claims both notably small-scale if they hold. MSN-syndicated with the original outlet unidentified, so the two hardware claims and the no-synthetic-data training claim are capture-grade only. Watch: a Block first-party announcement or demo, a model card or listening samples, and whether "human audio alone" (no transcripts, no synthetic speech) survives confirmation.