3 ms·
TLDR: they fell behind because of low tolerance for hallucinations and privacy focus (meaning preference for locally-run models).
by DenisM 2y ago
TLDR: they fell behind because of low tolerance for hallucinations and privacy focus (meaning preference for locally-run models).
- jonplackett 2y agoIn a few generations locally run models will be as good as gpt4 is now. Having on device ai that good will be amazing for so much stuff