(fathy.fr)

170 points riddley | 2 comments | 05 Sep 25 00:54 UTC | HN request time: 0.409s | source

1. microtherion ◴[05 Sep 25 10:57 UTC] No.45137193[source]▶

It's surprisingly capable. One tricky problem is trying to solve Captchas with it.

replies(1): >>45137769 #

2. neuroelectron ◴[05 Sep 25 12:32 UTC] No.45137769[source]▶

Multimodal LLMs can solve captchas easily if they're allowed to.

Forking Chrome to render in a terminal (2023)