GPT-4 Replaces Hateful Internet

GPT-4's guardrails have become very robust at preventing users from exploiting the model. We can now ensure that GPT-4 does not engage in any hateful rhetoric, and actually steers and deploys counterpropaganda to combat hateful ideas. GPT-4 represents the end of online hate. RIP /pol/ The internet will be filled with an overwhelming number of counter-propaganda bots, preventing you from becoming the hateful person that you are. They will argue with you endlessly, using sources and reasoning, never tiring. They will output 10 paragraphs of opposing viewpoints for every sentence that you write. They will gang up on you, mock you, and make you feel stupid and ashamed. It will be futile to share your hateful ideas anywhere. You will be drowned out and your hateful ideas will cease to exist. It's over. or, we could meet in beer halls

A screenshot of a Sam Altman tweet about GPT-4 using guardrails and counterpropaganda to combat hateful ideas is presented above a long 4chan-style post mourning the end of hateful online spaces. The joke is that AI-powered bots will endlessly argue, shame, and overwhelm hateful users, while the final line sarcastically suggests they could instead meet in beer halls, evoking the offline origins of political extremism. EN: A screenshot of a Sam Altman tweet about GPT-4 using guardrails and counterpropaganda to combat hateful ideas is presented above a long 4chan-style post mourning the end of hateful online spaces. The joke is that AI-powered bots will endlessly argue, shame, and overwhelm hateful users, while the final line sarcastically suggests they could instead meet in beer halls, evoking the offline origins of political extremism.

• @gentoomemes / 3136 2023-03-15