AI company has sufficient training capacity despite inference outsourcing

Original: @MikeIsaac they still have a lot of compute (more than enough) available to train their models

Source: x.com ↗