3 ms·
This is a neat observation. Slightly terrifying, but still interesting. Seems like there will also be cases where we discover new theories through the uninterpr
by cgearhart 2y ago
This is a neat observation. Slightly terrifying, but still interesting. Seems like there will also be cases where we discover new theories through the uninterpretable models—much easier and faster to experiment endlessly with a computer.