Naive question – do local models call into question the business model for AI company profitability?
From what I understand Gemma 4 is at least as capable as the best frontier model from only a few years ago. If that becomes a trend (new local-run models get released every year that are as good as the previous frontier models) does that mean a hell of…