News for hackers. Stuff that matters.

A comment on GPT-6 Astra covertly underperforms on safety research. Parent comment

Re:GPT-6 Astra covertly underperforms on safety research by stogot on Thursday November 05, 2026 @02:51PM (#50383586)
but the model providers seem to have trained their models with porn. Why? It’s harmful
[ Parent | on HN ]
Re:GPT-6 Astra covertly underperforms on safety research by hardbass on Friday November 06, 2026 @01:13AM (#50388797)
Why would the model providers care about harmless stuff like porn? They have to add these gates later on to not offend the types who hate this stuff.
[ Parent | on HN ]
Re:GPT-6 Astra covertly underperforms on safety research by stogot on Monday November 09, 2026 @12:21AM (#50421381)
You assume it’s harmless but it’s not
[ Parent | on HN ]
Re:GPT-6 Astra covertly underperforms on safety research by hardbass on Tuesday November 10, 2026 @08:49AM (#50440318)
I have proven what I sought out to prove. AI companies censor models because of people like you.
[ Parent | on HN ]