Comment on LLMDeathCount.com

<- View Parent
REDACTED@infosec.pub ⁨2⁩ ⁨days⁩ ago

I get your point, but the reality is that companies do actually put (well, started to) safeguards in place. I feel like I could get murdered on lemmy for saying this, but I was a ChatGPT subscriber for a year, up until last month. The amount of “Sorry Dave, I cannot do that” replied I recently started getting was ruining my experience. OpenAI recently implemented entire new system that transfers you to a different model if it detects something nental going on with you.

source
Sort:hotnewtop