Post #4012545

@crankylinuxuser@infosec.exchange @ErikJonker@mastodon.social Oh, were those Chinese models trained on data that explicitly consented to that? And once those models get a bit dated, surely we can train new ones without the need for a huge organisation, right? And that training wasn't super resource intensive, right? And those local models are good enough that it eliminates the biggest bottleneck in coding, the review, right?