22 ms·
> If science paper would be in its training That might well make it worse at science. For this among other reasons: https://www.vox.com/future-perfect/215043
by drewcoo 4y ago
> If science paper would be in its training
That might well make it worse at science.
For this among other reasons:
https://www.vox.com/future-perfect/21504366/science-replication-crisis-peer-review-statistics https://www.vox.com/future-perfect/21504366/science-replicat...
- TuringTest 4y agoOn the contrary, it might surface what good and bad scientists have been doing. A summary of everything that's going on in a scientific field may allow reviewers to find out which processes are solid, which ones are grounded on faulty premises or bad data, and which ones merit further research. Doing the same without the AI summarisations could take decades of review work.