Hacker News new | ask | show | jobs
by make3 1264 days ago
As a NLP researcher at a well known place & working on this, I totally 100% agree with you, this is ultra dangerous and should not be used, straight up, on anything remotely serious
2 comments
anjc 1264 days ago
Did you look at it? It summarises per-paragraph and the correctness can largely be verified by a layperson.
link
saliagato 1264 days ago
What is the best approach to building something like this? Do you work at HF?
link