Possibly a misleading title, the paper [1] is mostly of Microsoft affiliated authors, with one member of PG. [1] https://arxiv.org/pdf/2309.03926.pdf https://a

Possibly a misleading title, the paper [1] is mostly of Microsoft affiliated authors, with one member of PG. [1] https://arxiv.org/pdf/2309.03926.pdf https://arxiv.org/pdf/2309.03926.pdf edit: I'm not sure what's particularly exciting about this paper otherwise, it just seems like they parsed the PG html format and then applied some previously developed speech model to them.